Re: Our first inbound email via IPv6
Livingood, Jason jason_living...@cable.comcast.com writes: In preparation for the World IPv6 Launch, inbound (SMTP) email to the comcast.net domain was IPv6-enabled today, June 5, 2012, at 9:34 UTC. Roughly one minute later, at 9:35:30 UTC we received our first inbound email over IPv6 from 2001:4ba0:fff4:1c::2. That first bit of mail was spam, and was caught by our Cloudmark messaging anti-abuse platform (the sender attempted a range of standard spam tactics in subsequent connections). ... rim shot: i suggest that the e-mail industry consider a two-level approach to rejecting ipv6 spam based on source address. for more information see: http://www.circleid.com/posts/20110607_two_stage_filtering_for_ipv6_electronic_mail/ paul
Re: Our first inbound email via IPv6
Randy Bush ra...@psg.com writes: ... i have assiduously avoided gaining serious anti-spam fu. but it seems to me that ipv6 does not create/enable significantly more spam-bots. the malware will generally have complete control over the bottom 64 bits of an ipv6 address. there's no reason to expect to ever receive more than one spam message from any single ipv6 source. so, we'll all be blackholing /64's. moreover, there are going to be more native endpoints in ipv6 than there were in ipv4, since the NAT incentives are very different in the larger address pool. so, we'll all need network operators to whitelist the parts of their address spaces that they plan to send e-mail from, so that we can avoid having to blackhole things one /64 at a time. as before: for more information see: http://www.circleid.com/posts/20110607_two_stage_filtering_for_ipv6_electronic_mail/ paul
Re: ROVER routing security - its not enumeration
Doug Montgomery dougm.tl...@gmail.com writes: ... I think we debate the superficial here, and without sufficient imagination. The enumerations vs query issue is a NOOP as far as I am concerned.With a little imagination, one could envision building a box that takes a feed of prefixes observed, builds an aged cache of prefixes of interest, queries for their SRO records, re queries for those records before their TTLs expire, and maintains a white list of SRO valid prefix/origin pairs that it downloads to the router. this sounds like a steady state system. how would you initially populate it, given for example a newly installed core router having no routing table yet? if the answer is, rsync from somewhere, then i propose, rsync from RPKI. if the answer is, turn off security during bootup, then i claim, bad idea. ... Point being, with a little imagination I think one could build components with either approach with similar black box behavior. i don't think so. and i'm still waiting for a network operator to say what they think the merits of ROVER might be in comparison to the RPKI approach. (noting, arguments from non-operators should and do carry less weight.) -- Paul Vixie KI6YSY
rate limiting (Re: Open DNS Resolver reflection attack Mitigation)
Joe Maimon jmai...@ttec.com writes: Is there any publicly available rate limiting for BIND? How about host-based IDS that can be used to trigger rtbh or iptables? Google and Level3 manage to run open resolvers, why cant I? rate limiting on recursive servers is complicated by the lack of caching in most stub resolvers and applications. this makes it hard to tell by pure automation when a request flow is a spoof-source attack and when not. for most of us this isn't a problem since we'll put access control lists on our recursive name servers, only allowing queries from on-campus or on-net. for intentionally open resolvers, i expect there's a lot of monitoring and hand tuning, and that many deliberately low-grade attacks get by. noting that there are at least 15 million open recursive servers (most in low-quality CPE boxes front-ending cable or DSL links), an attacker has a long menu of places to send a small number of queries (to each) so that any rate limiting done by any one of the open recursive servers would not defend any victims against spoofed-source. spoofed-source is becoming wildly more popular. that's probably where to fix this. also the 15 million open recursives would be good to see fixed. at the moment most attacks are using authority servers, where it's far easier to automatically tell attack flows from non-attack flows. -- Paul Vixie KI6YSY
Re: isc - a good business
On 2012-05-30 12:53 AM, Nabil Sharma wrote: Paul: Where can we read details about the services ISC provided to the FBI, and how they were compensated? it's in the AP News article published a few weeks ago. for an example: http://www.foxnews.com/scitech/2012/04/23/hundreds-thousands-may-lose-internet-in-july/ As Mahatma Gandhi says: it is difficult, but not impossible, to conduct strictly honest business. Sincerely, Nabil the FBI's business dealings are transparent in this case. judge for yourself whether it's also strictly honest. paul
Re: rpki vs. secure dns?
On 5/29/2012 10:27 AM, Stephane Bortzmeyer wrote: On Mon, May 28, 2012 at 10:01:59PM +, paul vixie vi...@isc.org wrote a message of 37 lines which said: i can tell more than that. rover is a system that only works at all when everything everywhere is working well, and when changes always come in perfect time-order, Exactly like DNSSEC. no. dnssec for a response only needs that response's delegation and signing path to work, not everything everywhere. So, DNSSEC is doomed :-) i hope not. if we had to start over on something that can protect the cache against trivial pollution and also enable new applications like DANE, we'd be ten years from first prototype instead of ten years from ubiquity. paul
Re: rpki vs. secure dns?
On 2012-05-29 5:37 PM, Richard Barnes wrote: I agree with the person higher up the thread that ROVER seems like just another distribution mechanism for what is essentially RPKI data. noting, that up-thread person also said i havn't studied this in detail so i'm probably wrong. But does that distribution method easily allow you to get the full set of available data? From what little I know, it seems to me that ROVER is optimized for point queries, rather than bulk data access. Which is the opposite of making it easy to get full data :) that's close to the problem but it is not the problem. RPKI is a catalogue. it's possible to fetch all of the data you could need, before starting what's basically the batch job of computing the filters you will use at BGP-reception-time to either accept or ignore an incoming route. if your fetch and recompute steps don't work, then you'll have to continue filtering using stale data. if that data becomes too stale you're likely to have to turn off the filtering until you can resynchronize. ROVER is not a catalogue. it's impossible to know what data you could need to precompute any route filters, and it's impossible to know what 'all possible rover data' is -- in fact that would be a nonsequitur. you could i suppose query for every possible netblock (of every possible size) but that's an awful lot of queries and you'd have to do it every day in order to see new stuff or to know when to forget old stuff. the problem is in time domain bounding of data validity and data reachability. ROVER expects you to be able to query for the information about a route at the time you receive that route. that's point-in-time validity and reachability, which you might not have depending on where the DNS servers are and what order you're receiving routes in. RPKI+ROA expects you to have periodic but complete access to a catalogue, and then your future use of the data you fetched has only the risk of staleness or invalidity, but never reachability. as others have stated, there is no reference collection of bad ideas. otherwise we would have written this one up in 1996 when a couple of dns people looked at the routing system and said 'hey what about something like [ROVER]?' and the routing people explained in detail why it wouldn't work. paul
Re: rpki vs. secure dns?
ah, the force is strong in this one. On 2012-05-30 3:52 AM, Shane Amante wrote: On May 29, 2012, at 9:23 AM, Alex Band wrote: ... As far as I know, ROVER doesn't work like that. You can make a positive statement about a Prefix+AS combination, but that doesn't mark the origination from another AS 'unauthorized' or 'invalid', there merely isn't a statement for it. (Someone please confirm. I may be wrong.) Actually, I believe it does. Specifically, there are two types of DNS RR's: a) RLOCK: Route Lock b) SRO: Secure Route Origin Please refer to the following URL for definitions of each: http://tools.ietf.org/html/draft-gersch-grow-revdns-bgp-00#section-3. as dns is an unreliable protocol, and is not atomic across multiple questions, what is the effect of seeing one of those rrsets but not the other? (here again we see the disadvantage of starting from incomplete information.) On 2012-05-30 4:24 AM, Shane Amante wrote: On May 29, 2012, at 8:44 PM, Paul Vixie wrote: ... the problem is in time domain bounding of data validity and data reachability. ROVER expects you to be able to query for the information about a route at the time you receive that route. that's point-in-time validity and reachability, which you might not have depending on where the DNS servers are and what order you're receiving routes in. RPKI+ROA expects you to have periodic but complete access to a catalogue, and then your future use of the data you fetched has only the risk of staleness or invalidity, but never reachability. as others have stated, there is no reference collection of bad ideas. otherwise we would have written this one up in 1996 when a couple of dns people looked at the routing system and said 'hey what about something like [ROVER]?' and the routing people explained in detail why it wouldn't work. Just one correction to the above. As pointed out in Section 4 of draft-gersch-grow-revdns-bgp-00 near-real-time route origin verification is merely one instantiation of the ROVER concept. Please refer to that section for other potential uses of such published data. my comment on that draft was (on the dnsop mailing list, march 10, 2012): your draft-gersch-dnsop-revdns-cidr-01 is very clean and simple; the draft and the design are of admirable quality. as a co-author of RFC 2317 i agree that it does not suit the needs of bgp security since it seeks only to provide a method of fully naming hosts, not networks. importantly, i see no reference to RFC 1101 in your draft. RFC 1101 describes a way to name networks, and while at first it did not seem to be compatible with CIDR, implementation (in netstat -r back in BSD/OS 3.1) showed that RFC 1101 was in fact not as classful as it appeared. ... you may find that some of your work has already been done for you, or, you may find that this is related work that should be referenced in your draft along with the reasons why your proposed method is necessary. joe and dan (authors of this draft) responded: thanks for your comments and support. We will definitely reference RFC draft 1101 in our next version. and indeed this was done, but weakly: Changes from version 01 to 02 ... Expanded the related work discussion to include RFC 1101. looking at draft -02 we see: 4.1. Naming via RFC 1101 The problem of associating records with network names dates back to at least [RFC1101]. This work coincides with some of the early development of DNS and discusses issues regarding hosts.txt files. The RFC observation makes a key observation that one should provide mappings for subnets, even when nested. The approach taken here clearly states how to map an IPv4 prefix that is on an octet boundary. The RFC maps 10.IN-ADDR.ARPA for class A net 10, 2.128.IN-ADDR.ARPA for class B net 128.2, etc. This is essentially the same as the approach proposed here, although we append an m label (discussed later in this document). [RFC1101] also mentions more specific subnets, but the details are limited. We believe the approach proposed here builds on the best ideas from this RFC and expands on the details of how the naming convention would work. in other words no explaination for why the existing RFC 1101 encoding will not serve, even though RFC 1101 encoding been used for network naming in the CIDR era, without limitation. this reader is left wondering, what's the real impetus here? I would also ask people to expand their minds beyond the it must have a (near-)real-time mechanism directly coupled to the Control Plane for a variety of reasons. Such a tight coupling of /any/ two systems inevitably, and unfortunately, will only fail at scale in ways that likely would never have been predicted a priori[1] -- i think you're paying insufficient attention to this discussion, if you think that failure predictions have not already been well made with respect to the rover
isc - a good business
greetings. i didn't notice this before, and i want to complete the record. i'm paying more attention to the quoting this time, too. On Wed, May 23, 2012 at 04:33:28PM -0400, Christopher Morrow wrote: On Wed, May 23, 2012 at 1:40 AM, bmanning at vacation.karoshi.com wrote: Paul will be there to turn things off when they no longer make money for his company. is the dns changer thingy making money for isc? no. yes. depends on what you mean. we're under contract to the department of justice to run the servers. the amount is intended to be cost-recovery level. (doing stuff like this for nothing does not scale, unless the business is successful enough elsewhere, and even then there'll be limits.) From: bmanning at vacation.karoshi.com (bmanning at vacation.karoshi.com) Date: Wed, 23 May 2012 21:00:27 + Message-ID: 20120523210027.ga26...@vacation.karoshi.com. pretty sure. a contract w/ the Feds, outsouring contracts w/ affected ISPs when the Fed deal runs out, development funding to code these kinds of fixes into future versions of software, any number of second and third order fallout. i don't know of any outsourcing contracts that will come into force when the fed deal runs out. there is no development funding because there are no code fixes for this kind of problem. i am certainly hoping for second and third order fallout; we pay the people who write BIND using money collected from other people who buy support for BIND, or consulting, or training, or custom feature development. that's how we keep BIND free, and that's how we use a BSD-like license rather than a GPL-like license -- noone who derives their appliance or product from BIND has any obligation to anybody. (noting, nlnetlabs for unbound and nsd also use a BSD-like license, so ISC is not alone here.) No telling how effective constent self-promotion is. One thing is clear, Paul is able to tell a great story. thanks for your kind words. i've been trying to uplevel my storytelling capability since i've long since lost my coding skills. but its all speculation from here. ISC is well positioned to extract value from both ends of the spectrum. They have a great business model. The optics look pretty odd from here, at lesat to me however - I am very glad for: )open source )other vendors of DNS SW. the women and men of ISC work their asses off to keep the world safer and saner. as you puzzle over our business model please keep in mind that there is no exit possibility here -- no IPO, no buyout. salaries at ISC are fair and reasonable, but nobody's getting rich doing this work. so while i am likewise glad for open source and for other vendors of open source DNS software, i'm struggling to grasp the intended suggestion. should ISC stop giving away software? should we stop adding the features to our software that make it relevant and desireable? should we stop selling the support and training and consulting that make it possible to give away this software? if you have a specific accusation of evil-doing, or a specific suggestion for how ISC can become more morally pure, then please say exactly what you mean, and we can discuss that. for more information, see: http://www.isc.org/community/blog/201001/why-isc-not-profit. the short version is: ISC is a good business, we do good things, mostly for free, but also for our customers. and we're totally unapologetic about what we do and who we are. paul
Re: isc - a good business
On 5/28/2012 11:52 AM, Randy Bush wrote: ... maybe a bit too much layer ten for my taste. ... on that, we're trying to improve. for example, we used to forego features that some of us found repugnant, such as nxdomain remapping / ad insertion. since the result was that our software was less relevant but that there was no reduction in nxdomain remapping as a result of BIND not providing it. so we dropped some of the layer ten stuff and moved more in the direction of providing features the community of interest found relevant. some say we sold out. maybe that's what manning is on about, i can't tell. the software is free, and isc cherishes our relevance. if you catch us doing wierd layer ten stuff that bugs you, give a shout. maybe we don't really mean it. ... and i run and appreciate the software. that's why we're here. paul
Re: rpki vs. secure dns?
more threads from the crypt as i catch up to 6000 missed nanog posts. Dobbins, Roland rdobb...@arbor.net writes: On Apr 28, 2012, at 5:17 PM, Saku Ytti wrote: People might scared to rely on DNS on accepting routes, but is this really an issue? Yes, recursive dependencies are an issue. I'm really surprised that folks are even seriously considering something like this, but OTOH, this sort of thing keeps cropping up in various contexts from time to time, sigh. so, first, i think you mean circular dependencies not recursive dependencies. second, i'd agree that that's probably bad engineering. third, rsync's dependencies on routing (as in the RPKI+ROA case) are not circular (which i think was david conrad's point but i'll drag it to here.) my reason for not taking ROVER seriously is that route filter preparation is an essentially offline activity -- you do it from a cron job not live. and to do this you have to know in advance what policy data is available which may or may not have the same coverage as the routes you will receive between one cron job and the next. we could in other words use DNS to store route policy data if we wanted to use a recursive zone transfer of all policy zones, as a replacement for rsync. (but why would we do this? we have rsync, which worked for IRR data for many years.) ROVER expects that we will query for policy at the instant of need. that's nuts for a lot of reasons, one of which is its potentially and unmanageably circular dependency on the acceptance of a route you don't know how to accept or reject yet. my take-away from this thread is: very few people take RPKI seriously, but even fewer take ROVER seriously. -- Paul Vixie KI6YSY
Re: isc - a good business
(all caught up after this.) Jay Ashworth j...@baylink.com writes: - Original Message - From: paul vixie vi...@isc.org On 5/28/2012 11:52 AM, Randy Bush wrote: ... maybe a bit too much layer ten for my taste. ... on that, we're trying to improve. for example, we used to forego features that some of us found repugnant, such as nxdomain remapping / ad insertion. since the result was that our software was less relevant but that there was no reduction in nxdomain remapping as a result of BIND not providing it. To clarify that a bit... let's keep trying. You're saying you used to decline to include in BIND the capability to break the Internet by returning things other than NXDOMAIN for names which do not exist... no, that's not what i'm saying. but now you're *ok* with breaking the internet, and BIND now does that? no, that's also not what i'm saying. If that's what you mean, I'll explain to you why that's a bad layer 10 call. it's not, but i'm listening. *Now*, you see, we no longer have a canonical Good Engineering Example to which we can point when yelling at people (and software vendors) which *do* permit that, to say see? You shouldn't be doing that; it's bad. The Web Is Not The Internet. i see what you mean, and i'm sad that this arrow is no longer in your quiver. perhaps you can still refer to nlnetlabs unbound for this purpose. if i thought there was even one isp anywhere who wanted to use nxdomain remapping but didn't because bind didn't have that feature, i'd be ready to argue the point. but all isc did by not supporting this feature was force some isp's to not use bind, and: isc is not in the sour grapes business. meanwhile isc continues to push for ubiquitous dnssec, through to the stub, to take this issue off the table for all people and all time. (that's the real fix for nxdomain remapping.) -- Paul Vixie KI6YSY
Re: rpki vs. secure dns?
On 5/28/2012 9:42 PM, David Conrad wrote: On May 28, 2012, at 1:59 PM, Paul Vixie wrote: third, rsync's dependencies on routing (as in the RPKI+ROA case) are not circular (which i think was david conrad's point but i'll drag it to here.) Nope. My point was that anything that uses the Internet to fetch the data (including rsync) has a circular dependency on routing. It's just a question of timing. when you say it's a question of timing, i agree, but then i think you won't agree again. in batch mode i can express a policy that's true from that moment forward until removed. in real time mode i have to be able to express a policy which is both valid and reachable at the moment of validation. that's a question of timing but the word just as in just a question of timing trivializes a galaxy-sized problem space. ROVER expects that we will query for policy at the instant of need. Might want to review https://ripe64.ripe.net/presentations/57-ROVER_RIPE_Apr_2012.pdf, particularly the slide entitled Avoid a Cyclic Dependency. i read it. this is also what draft-gersch-grow-revdns-bpg says. this makes for a fail open design -- the only thing the system can reliably tell you is reject. this precludes islands of, or a mode of, fail closed. while i don't consider a mode of fail closed to be practically likely, i'd like to preserve the fig leaf of theoretical possibility. (and the more likely possibility that islands of fail closed could be sewn in.) As far as I can tell, ROVER is simply Yet Another RPKI Access Method like rsync and bittorrent with its own positives and negatives. i can tell more than that. rover is a system that only works at all when everything everywhere is working well, and when changes always come in perfect time-order, where that order is nowhere defined, and is in any event uncontrollable. rsync's punt of the ordering problem is batch mode with policy start times rather than policy valid instants. to my mind anyone who doesn't punt the ordering problem has to instead solve it. rover does neither. paul
vixie, father of multitudes
thanks to several folks who let me know this was going on. i hadn't even noticed that i wasn't getting nanog@. thanks to seclists.org for hosting an archive i could use. --- From: bmanning () vacation karoshi com Date: Wed, 23 May 2012 05:40:16 + On Tue, May 22, 2012 at 10:07:52PM -0700, Michael J Wise wrote: On May 22, 2012, at 9:10 PM, bmanning () vacation karoshi com wrote: On Tue, May 22, 2012 at 08:52:52PM -0700, Michael J Wise wrote: On May 22, 2012, at 8:35 PM, Randy Bush wrote: father of bind? that's news. http://boingboing.net/2012/03/29/paul-vixies-firsthand-accoun.html He was there, and Put The Fix In, to down the network. Certainly news to Phil Almquist and the entire BIND development team at UCB. Paul was at DECWRL and cut his teeth on pre-existing code. While he (and ISC) have since revised, gutted, tossed all the orginal code, rebuilt it twice - and others have done similar for their DNS software, based on the BIND code base, implementation assumptions, and with little or no ISC code, and they call it BIND as well, it would be a HUGE leap of faith to call Paul Vixie the father of BIND - The Berkeley Internet Naming Daemon. Methinks we're talking at cross purposes. maybe... :) my comment was refering to the father of bind statement. i don't describe myself that way. i inherited bind at 4.8.3 and fixed stuff. i rewrote a lot of it for 4.9. we (mostly me but with huge work by robert halley and mark andrews) rewrote most of it for bind 8.1. (there was no 8.0.) other people (not me) wrote bind 9.x. other people (mostly not the same people) are writing bind 10. if my wikipedia entry is wrong in this regard i invite folks to fix it. last i heard it's disallowed for people to edit their own entries, so i have not tried. i am not the father of anything, except four healthy kids. i do sometimes call myself the wierd uncle of the internet but father of bind is not what i mean. As for being there and Put The Fix In... Makes for great PR but in actual fact, its a bandaid that is not going to stem the tide. An actual fix would really need to change the nature of the creaky 1980's implementation artifacts that this community loves so well. I don't think we're talking about the same thing at all. Paul was there to shut down the DNS changer system and replace it with something that restored functionality to the infected machines. And I gather Paul will be one of the people who will turn the lights out on it. yes, and yes. He didn't shut down DNS Changer, he put up an equivalent system to hijack DNS traffic and direct it to the right place... SO folks didn't see any problem and the DNS Changer infection grew and got worse. When he is legally required to take his bandaide out of service, then the problem will resolve by folks who will have to clean their systems. it's true, the fbi team who powered all that stuff off and loaded it into a u-haul truck are the ones who shut down dns changer. or perhaps it was the police in estonia who arrested all those people. i'm not the shutter-downer. As for turning the lights out - that will only happen when the value of DNS hijacking drops. As it is now, ISC has placed DNS hijacking code into their mainstream code base... because DNS hijacking is so valuable to folks. In a modestly favorable light, ISC looks like an arms dealer (DNS redirection) to the bad guys -AND- (via DNSSEC) the good guys. Either way, they make money. well, no. but that seems off-topic. start a new thread if you care. (and, cc me!) And yes, I think I agree with you. Paul will be there to turn things off when they no longer make money for his company. well, no. when the court order runs out we will have to shut things down. but the money FBI is paying us for this is just to cover costs. and, it's not my company. isc is a 501(c)(3), basically a ward of the state of delaware, having no shares and therefore no shareholders. Your other comments are non-sequitur to the main issue. Perhaps I am not a member of the Paul Vixie cult of personality. so sad. When those servers are turned off, Customer Support folks at many ISPs will prolly want to take their accrued vacation. Amen. And there will be thousands more of them when the court order expires than existed when the Feds called him in. um. no. hundreds of thousands less than before the feds called ISC in. see dcwg.org. it's lovely to have so many fans. keep those cards and letters coming. (but, cc me!) paul
rpki vs. secure dns?
http://tech.slashdot.org/story/12/04/27/2039237/engineers-ponder-easier-fix-to-internet-problem The problem: Border Gateway Protocol (BGP) enables routers to communicate about the best path to other networks, but routers don't verify the route 'announcements.' When routing problems erupt, 'it's very difficult to tell if this is fat fingering on a router or malicious http://www.itworld.com/security/272320/engineers-ponder-easier-fix-dangerous-internet-problem,' said Joe Gersch, chief operating officer for Secure64, a company that makes Domain Name System (DNS) server software. In a well-known incident, Pakistan Telecom made an error with BGP after Pakistan's government ordered in 2008 that ISPs block YouTube, which ended up knocking Google's service offline http://slashdot.org/story/08/02/25/1322252/pakistan-youtube-block-breaks-the-world. A solution exists, but it's complex, and deployment has been slow. Now experts have found an easier way. this seems late, compared to the various commitments made to rpki in recent years. is anybody taking it seriously?
Re: wet-behind-the-ears whippersnapper seeking advice on building a nationwide network
Benson Schliesser bens...@queuefull.net writes: Hi, Paul. sorry for the delay. i'll include the entirety of this short thread. For what it's worth, I agree that ARIN has a pretty good governance structure. (With the exception of NomCom this year, which is shamefully unbalanced.) ... as the chairman of the 2011 ARIN NomCom, i hope you'll explain further, either publically here, or privately, as you prefer. My understanding is that the NomCom consists of 7 people. Of those, 2 come from the board and 2 come from the AC. Together, those 4 members of the existing establishment choose the remaining 3 NomCom members. In the past, there was at least the appearance of random selection for some of the NomCom members. But in any case, due to its composition, the NomCom has the appearance of a body biased in favor of the existing establishment. Please correct any misunderstanding that I might have. Otherwise, I encourage an update to the structure of future NomComs. can you explain what it was about prior nomcoms that gave the appearance of random selection? to the best of my knowledge, including knowledge i gained as chair of the 2008 ARIN NomCom, we've been doing it the same way for quite a while now. so i do not understand your reference to at least the appearance of random selection in the past. since ARIN members-in-good-standing elect the board and advisory council, and also make up three of the four seats of the nominations committee, i do not share your view on bias as expressed above. i think it shows that ARIN is clearly governed by its members -- which is as it should be. by your two references to the existing establishment do you intend to imply that ARIN's members don't currently have the establishment that they want, or that they could not change this establishment if they wanted to, or that ARIN's members are themselves part of the existing establishment in some way that's bad? ARIN's bylaws firmly place control of ARIN into the hands of its members. if you think that's the wrong approach, i'm curious to hear your reasoning and your proposed alternative. -- Paul Vixie KI6YSY
Re: wet-behind-the-ears whippersnapper seeking advice on building a nationwide network
On Thu, 22 Sep 2011 21:05:51 -0500 Benson Schliesser bens...@queuefull.net wrote: Earlier this year I received the following from ARIN member services: This year the NomCom charter was changed by the Board. In the past the 3 Member volunteers were selected at random. This year the 3 volunteers will be chosen by the 4 current members of the NomCom (2 from the Board 2 from the AC) yow. i should have remembered this, you'd think. The above quote was sent to me in response to a query I made, inquiring how the NomCom would be chosen in 2011. It is consistent with what I was told in 2010, when I was chosen to be part of the 2010 NomCom. At that time I was told that Member volunteers were chosen randomly. During my NomCom tenure, however, it was suggested to me privately that there was very little randomness involved in the selection process; I was told that individuals were specifically chosen for NomCom. I don't know what to make of this disparity, honestly, which is why I referenced the appearance of random selection. suggested to you privately by arin staff? The NomCom acts as a filter, of sorts. It chooses the candidates that the membership will see. The fact that the NomCom is so closely coupled with the existing leadership has an unfortunate appearance that suggests a bias. I'm unable to say whether the bias exists, is recognized, and/or is reflected in the slate of candidates. But it seems like an easy enough thing to avoid. you seem to mean that the appearance of bias would be easy to avoid, then. As for my use of existing establishment: I'm of the impression that a relatively small group of individuals drive ARIN, that most ARIN members don't actively participate. I have my own opinions on why this is, but they aren't worth elaborating at this time - in fact, I suspect many ARIN members here on NANOG can speak for themselves if they wanted to. In any case, this is just my impression. If you would rather share some statistics on member participation, election fairness, etc, then such facts might be more useful. i think our participation level in elections is quite high and i'll ask for details and see them published here. ARIN's bylaws firmly place control of ARIN into the hands of its members. if you think that's the wrong approach, i'm curious to hear your reasoning and your proposed alternative. One of ARIN's governance strengths is the availability of petition at many steps, including for candidates rejected by the NomCom. Likewise, as you noted, leaders are elected by the membership. For these reasons I previously noted that ARIN has a pretty good governance structure and I continue to think so. It could be improved by increased member involvement, as well as broader involvement from the community. (For instance, policy petitions should include responses from the entire affected community, not just PPML.) But my criticisms should be interpreted as constructive, and are not an indictment of the whole approach. thanks for saying so. -- Paul Vixie
Re: wet-behind-the-ears whippersnapper seeking advice on building a nationwide network
Benson Schliesser bens...@queuefull.net writes: For what it's worth, I agree that ARIN has a pretty good governance structure. (With the exception of NomCom this year, which is shamefully unbalanced.) ... as the chairman of the 2011 ARIN NomCom, i hope you'll explain further, either publically here, or privately, as you prefer. -- Paul Vixie KI6YSY
not operational -- call for nominations for ARIN council board
gentlefolk, ARIN is the community's self-generated steward for internet numbering resources (ip addresses and autonomous system numbers) and it is governed by volunteers from the community who serve on its advisory council and executive board. every year ARIN holds an election to fill or renew several expiring terms. candidates need not be ARIN members. please see https://www.arin.net/announcements/2011/20110725_elec.html and think about whether who you can nominate or whether you can self- nominate. paul vixie chairman, 2011 arin nomcom
Re: ICANN to allow commercial gTLDs
David Conrad d...@virtualized.org writes: I believe the root server operators have stated (the equivalent of) that it is not their job to make editorial decisions on what the root zone contains. They distribute what the ICANN/NTIA/Verisign gestalt publishes. yes. for one example, see: http://www.icann.org/en/announcements/announcement-04jan08.htm other rootops who have spoken about this have said similar/compatible things. -- Paul Vixie KI6YSY
Re: unqualified domains, was ICANN to allow commercial gTLDs
Adam Atkinson gh...@mistral.co.uk writes: It was a very long time ago, but I seem to recall being shown http://dk, the home page of Denmark, some time in the mid 90s. Must I be recalling incorrectly? no you need not must be. it would work as long as no dk.this or dk.that would be found first in a search list containing 'this' and 'that', where the default search list is normally the parent domain name of your own hostname (so for me on six.vix.com the search list would be vix.com and so as long as dk.vix.com did not exist then http://dk/ would reach dk.) -- Paul Vixie KI6YSY
Re: ICANN to allow commercial gTLDs
Jay Ashworth j...@baylink.com writes: ... and that the root wouldn't be affected by the sort of things that previously-2LD now TLD operators might want to do with their monocomponent names... someone asked me privately a related question which is, if there's a .SONY and someone's web browser looks up http://sony/ and a root name server gets a query for SONY./IN/ then what will happen? the answer is happily that the result will be a delegation (no RR in the answer section even if the root name server knows one for some reason). the answer section will be empty, the authority section will have a SONY/IN/NS RRset in it, and the additional section will have the nec'y IN/ and IN/A RRsets for those NSs. this is sometimes called the BIND9 behaviour in contrast to BIND8/BIND4 which would have answered the question if they knew the answer, even if they also knew that the qname had been delegated. BIND9 changed this, and NSD does it the same way. RFC 1034/1035 is pretty clear about this, so to be this should not be called the BIND9 behaviour but rather simply correct. which as someone pointed out, a 3-digit RFC forbids for security reasons anyway. three digit? i was thinking of http://www.ietf.org/rfc/rfc1535.txt which was written as air cover for me when i added the ndots:NNN behaviour to BIND4's stub resolver. and, looking at firefox on my workstation just now: [58] 2011-06-19 23:27:49.906040 [#4 em1 0] \ [24.104.150.12].24003 [24.104.150.2].53 \ dns QUERY,NOERROR,57397,rd \ 1 sony.vix.com,IN,A 0 0 0 [58] 2011-06-19 23:27:49.909895 [#5 em1 0] \ [24.104.150.12].26356 [24.104.150.2].53 \ dns QUERY,NOERROR,57398,rd \ 1 sony.vix.com,IN, 0 0 0 [50] 2011-06-19 23:27:49.910489 [#6 em1 0] \ [24.104.150.12].23228 [24.104.150.2].53 \ dns QUERY,NOERROR,57399,rd \ 1 sony,IN,A 0 0 0 [50] 2011-06-19 23:27:49.930022 [#7 em1 0] \ [24.104.150.12].37238 [24.104.150.2].53 \ dns QUERY,NOERROR,57400,rd \ 1 sony,IN, 0 0 0 [58] 2011-06-19 23:27:49.931059 [#8 em1 0] \ [24.104.150.12].17401 [24.104.150.2].53 \ dns QUERY,NOERROR,33742,rd \ 1 www.sony.com,IN,A 0 0 0 [124] 2011-06-19 23:27:50.112451 [#9 em1 0] \ [24.104.150.2].53 [24.104.150.12].17401 \ dns QUERY,NOERROR,33742,qr|rd|ra \ 1 www.sony.com,IN,A \ 1 www.sony.com,IN,A,600,72.52.6.10 \ 2 sony.com,IN,NS,172800,pdns1.cscdns.net \ sony.com,IN,NS,172800,pdns2.cscdns.net 0 ...i see that the browser's stub is indeed looking at the search list first when there are no dots in the name. that's correct behaviour by the RFC and also anecdotally since if i had an internal web server here called sony.vix.com i would want my web browser to assume that that was the one i wanted when i typed http://sony/;. having it go outside my network and hit a TLD first would be a dangerous information leak. (this also shows DNS's lack of a formal presentation layer as clearly as anything ever could.) inevitably there will be folks who register .FOOBAR and advertise it as http://foobar/; on a billboard and then get burned by all of the local foobar.this.tld and foobar.that.tld names that will get reached instead of their TLD. i say inevitable; i don't know a way to avoid it since there will be a lot of money and a lot of people involved. -- Paul Vixie KI6YSY
Re: unqualified domains, was ICANN to allow commercial gTLDs
Date: Sun, 19 Jun 2011 19:30:58 -0500 From: Jeremy jba...@gmail.com DK may not be hierarchical, but DK. is. If you try to resolve DK on it's own, many (most? all?) DNS clients will attach the search string/domain name of the local system in order to make it a FQDN. The same happens when you try and resolve a non-existent domain. Such as alskdiufwfeiuwdr3948dx.com, in wireshark I see the initial request followed by alskdiufwfeiuwdr3948dx.com.gateway.2wire.net. However if I qualify it with the trailing dot, it stops after the first lookup. DK. is a valid FQDN and should be considered hierarchical due to the dot being the root and anything before that is a branch off of the root. see RFC1034 i think he's seen RFC 1034 :-). anyway, i don't see the difference between http://sony/ and http://sony./ and if a technology person tried to explain to a marketing person that single-token TLD names *can* be used as long as there's a trailing dot, the result would hopefully be that glazed look of nonunderstanding but would far more likely be an interpretation of oh, so it's OK after all, we'll use it that way, thanks! furthermore, the internet has more in it than just the web, and i know that foo@sony. will not have its RHS (sony.) treated as a hierarchical name. i think we have to just discourage lookups of single-token names, universally.
Re: unqualified domains, was ICANN to allow commercial gTLDs
From: David Conrad d...@virtualized.org Date: Sun, 19 Jun 2011 16:04:09 -1000 On Jun 19, 2011, at 3:24 PM, Paul Vixie wrote: i think we have to just discourage lookups of single-token names, universally. How? that's a good question. marka mentioned writing an RFC, but i expect that ICANN could also have an impact on this by having applicants sign something that says i know that my single-label top level domain name will not be directly usable the way normal domain names are and i intend to use it only to register subdomain names which will work normally.
Re: unqualified domains, was ICANN to allow commercial gTLDs
Date: Sun, 19 Jun 2011 19:22:46 -0700 From: Michael Thomas m...@mtcc.com that's a good question. marka mentioned writing an RFC, but i expect that ICANN could also have an impact on this by having applicants sign something that says i know that my single-label top level domain name will not be directly usable the way normal domain names are and i intend to use it only to register subdomain names which will work normally. Isn't this problem self regulating? If sufficient things break with a single label, people will stop making themselves effectively unreachable, right? alas, no. if someone adds something to the internet that doesn't work right but they ignore this and press onward until they have market share, then the final disposition will be based on market size not on first mover advantage. if you live in the san francisco bay area you probably know about the sound walls along the US101 corridor. the freeway was originally built a long way from where the houses were, but then a few generations of people built their houses closer and closer to the freeway. then their descendants or the folks who bought these houses third or fourth hand complained about the road noise and so we have sound walls. no harm exactly, and no foul, except, noone likes the result much. here's this quote again: Distant hands in foreign lands are turning hidden wheels, causing things to come about which no one seems to feel. All invisible from where we stand, the connections come to pass and though too strange to comprehend, they affect us nonetheless, yes. James Taylor, _Migrations_ good stewardship and good governance means trying to avoid such outcomes.
Re: unqualified domains, was ICANN to allow commercial gTLDs
Date: Sun, 19 Jun 2011 22:32:59 -0700 From: Doug Barton do...@dougbarton.us ... the highly risk-averse folks who won't unconditionally enable IPv6 on their web sites because it will cause problems for 1/2000 of their customers. let me just say that if i was making millions of dollars a day and i had the choice of reducing that by 1/2000th or not i would not choose to reduce it. as much as i love the free interchange of ideas i will point out that commerce is what's paid the internet's bills all these years.
Re: v6 proof of life
Jima na...@jima.tk writes: 44 2001:db8::230:48ff:fef2:f340 44 2001:db8::230:48ff:fef0:1de How can 2001:db8::/32 reach your machines ? Lack of ingress filtering on Mr. Vixie's part, ... indeed. i had no idea. and lack of egress filtering on whoever-owns-those-Supermicro-board's part. That's not to say there's a route back, by any means. i'll bet i'm not alone in seeing traffic from this prefix. as a rootop i can tell you that we see plenty of queries from ipv4 rfc1918 as well. -- Paul Vixie KI6YSY
v6 proof of life
it's been a while since i looked at the query stream still hitting {rbl,dul}.maps.vix.com. this was the world's first RBL but it was renamed from maps.vix.com to mail-abuse.org back in Y2K or so. i have not sent anything but NXDOMAIN in response to one of these queries for at least ten years, yet the queries just keep coming. here's a histogram of source-ip. feel free to remove yourself :-). it's just a half hour sample. i'll put this up on a web page soon. importantly and happily, there's a great deal of IPv6 happening here. re: 524 200.59.134.164 492 2001:288:8201:1::14 455 209.177.144.50 418 193.124.83.73 392 140.144.32.205 360 143.54.204.171 355 208.76.170.121 282 2403:2c00:1::5 262 2001:1a80:103:5::2 225 2001:288:8201:1::2 195 2604:3500::ad:1 186 2001:288:8201:1::10 174 209.157.254.10 167 2001:b30:1:100::100 158 12.47.198.68 142 2001:288:0:2::60 125 2002:d58e:8901::d58e:8901 118 77.72.136.2 115 140.118.31.99 113 66.240.198.100 102 2001:9b0:1:601:230:48ff:fe8a:c7f4 100 2001:888:0:24:194:109:20:107 99 212.73.100.132 92 2405:d000:0:100::214 86 202.177.26.107 83 2405:d000:0:100::228 82 195.101.253.253 77 210.175.244.210 77 2001:558:1014:f:69:252:96:24 76 64.168.228.129 76 2001:2c0:11:1::c02 71 2001:b30:1::190 68 2001:558:1014:f:68:87:76:185 68 2001:4dd0:100:1020:53:2:0:1 67 2001:558:1014:f:69:252:96:22 67 2001:558:1014:f:68:87:76:189 63 2001:558:1014:f:69:252:96:25 57 2607:f758:6000:13::e4 56 2001:558:1014:f:68:87:76:181 55 212.234.229.242 52 2607:f758:e000:13::e4 51 2607:fdb8:0:b:2::2 51 208.188.98.249 51 2001:558:1014:f:68:87:76:190 49 66.192.109.211 45 2001:558:1014:f:69:252:96:23 44 2001:db8::230:48ff:fef2:f340 44 2001:db8::230:48ff:fef0:1de 42 213.171.61.117 41 201.116.43.232 40 2607:fdb8:0:b:1::2 40 2001:470:1f0a:c1b::2 40 190.82.65.243 38 190.169.30.2 36 2605:d400:0:27::3 35 2001:d10:2:3::1:2 31 2605:d400:0:27::7 30 220.110.24.250 28 84.55.220.139 28 2a00:1db0:16:2::347:2 23 2607:f2e0::1:3:2 23 2001:15c8:8::2:1 22 218.44.167.26 22 2001:6c8:2:100::53 22 2001:6b0:1::201 21 193.169.45.5 20 2607:f470:1003::3:3 19 2a01:8c00:ff60:2:230:48ff:fe85:d47e 19 221.133.36.229 19 213.178.66.2 19 202.169.240.10 19 2001:c28:1:1:dad3:85ff:fee1:30ec 19 163.29.248.1 18 195.58.224.34 17 2a02:460:4:0:250:56ff:fea7:23d 17 221.245.76.99 17 2001:c28:1:1:dad3:85ff:fee0:4f68 16 2001:630:200:8120::d:1 15 2a01:1b8::1 15 218.44.236.98 15 2001:ad0:: 15 2001:41d8:1:8028::54 14 2a00:1b50::2:2 14 192.149.202.9 13 200.74.222.140 12 98.130.2.250 12 2a00:1eb8:0:1::1 12 2600:c00:0:1::301 12 2001:c28:1:1:dad3:85ff:fee0:3f20 12 2001:380:124::4 12 2001:218:4c0:1:2e0:81ff:fe55:a018 12 2001:12d0::3 12 195.228.156.150 12 194.42.134.132 11 66.111.66.240 11 2607:f010:3fe:100:0:ff:fe00:1 11 211.25.195.236 11 210.162.229.210 11 2001:c28:1:1:dad3:85ff:fee0:be80 11 2001:a10:1:::2:c918
Re: Yahoo and IPv6
Date: Tue, 17 May 2011 11:07:17 +0200 From: Mans Nilsson mansa...@besserwisser.org ... It's not like you can even reach anything at home now, let alone reach it by name. that must and will change. let's be the generation who makes it possible. I'd like to respond to this by stating that I support this fully, but I'm busy making sure I can reach my machines at home from the IPv6 Internet. By name. ;-) :-). to be clear, the old pre-web T1 era internet did not have much content but what content there was, was not lopsided. other than slip and ppp there weren't a lot of networks one would call access and a smaller number of networks one would call content. i am not wishing for that, i like the web, i like content, i know there will be specialized networks for access and content. but i also think (as jim gettys does) that we ought to be able to get useful work done without being able to reach the whole internet all the time. that's going to mean being able to reach other mostly-access networks in our same neighborhoods and multitenant buildings and towns and cities, directly, and by name. it does not mean being able to start facebook 2.0 out of somebody's basement, but it does mean being able to run a personal smtp or web server in one's basement and have it mostly work for the whole internet and work best for accessors who are close by and still work even when the upstream path for the neighborhood is down.
Re: Yahoo and IPv6
Date: Tue, 17 May 2011 11:49:47 -0400 From: Steve Clark scl...@netwolves.com This is all very confusing to me. How are meaningful names going to assigned automatically? It'll probably be a lot like Apple's and Xerox's various multicast naming systems if we want it to work in non-globally connected networks. Right now I see something like ool-6038bdcc.static.optonline.net for one of our servers, how does this mean anything to anyone else? It wouldn't of course. I'm sorry if my earlier words on this were useless. Dave Taht gave a wonderful talk a few weeks ago (Finishing the Internet, see http://amw.org/prog11.pdf) during which he had us start an rsync from his wireless laptop to as many of ours as could run rsync, and then had the conference organizer turn off the upstream link. He noted that those of us using the local resource (a giant file, either an ISO or a MPEG or similar) were still getting work done whereas those of us trying to use the internet were dead in the water. Then, referring to his time in Nicaragua he said that he has a lot of days like this and he'd like more work to be possible when only local connectivity was available. Compelling stuff. Pity there's no global market for localized services or we'd already have it. Nevertheless this must and will get fixed, and we should be the generation who does it.
Re: Yahoo and IPv6
Date: Mon, 16 May 2011 14:37:46 -0400 From: Jim Gettys j...@freedesktop.org perhaps i'm too close to the problem because that solution looks quite viable to me. dns providers who don't keep up with the market (which means ipv6+dnssec in this context) will lose business to those who do. I don't believe it is currently viable for any but the hackers out there, given my experience during the Comcast IPv6 trial. Typing V6 addresses (much less remembering them) is a PITA. You are asking people who don't even know DNS exists, to bother to establish another business relationship (or maybe DNS services might someday be provided by their ISP). actually, i'm asking the opposite. only hackers run their own dns mostly; the vast majority of users who don't know what ipv6 or dnssec are, are already outsourcing to ultradns/neustar, or verisign, or dyn.com, etc, or for recursive they're using opendns, google dns, etc. these companies can either add the new services and do outreach to their customer bases, or they can allow their competitors to do so. of those who still run their own dns, the vast majority actually do know the dnssec and ipv6 issues facing them. If you get past that hurdle they get to type long IPv6 addresses into a web page they won't remember where it was the year before when they did this the last time to add a machine to their DNS. i've been using ipv6 dual stack for ten years at ISC and for one year at home (i was comcast's first north american dual stack native customer) and the only time i type long ipv6 addresses is when editing dns zone files or configuring routers and hosts. i think your experiences may have been worse than mine and i'll be interested in knowing whether they're common. The way this ought to work for clueless home users (or cluefull users too, for that matter) is that, when a new machine appears on a network, it just works, by which I mean that a globally routeable IPv6 address appears in DNS without fussing around using the name that was given to the machine when it was first booted, and that a home user's names are accessible via secondaries even if they are off line. this is why ISC DHCP and ISC BIND can communicate using RFC 2136 DNS dynamic updates, secured with RFC 2845 transaction signatures. once you get this running then you don't have to type ipv6 addresses anywhere. and i know that infoblox and other BIND Inside appliance vendors have the same capability, and that Cisco and other DNS/DHCP vendors can also participate in these open standards pretty much out of the box. this is what i worked on when i first found out about IETF back in 1995 or so. it's all done now you just have to learn it and deploy it. (and if you don't think end users ought to have to learn how to configure their DHCP to talk to their DNS, i will point them at a half dozen appliance and outsourcing vendors who can take the ones and zeroes out of this for them.) And NXDOMAIN should work the way it was intended, for all the reasons you know better than I. while i agree, i don't think the people who are substituting positive responses for NXDOMAIN care at all what you think or what i think, so i'm going to focus on what can be done which is advancing robust solutions. This is entirely possible ;-). Just go ask Evan Hunt what he's been up to with Dave Taht recently more appliance vendors including open source are definitely welcome. the pool is large enough for everybody to swim in it.
Re: Yahoo and IPv6
From: Owen DeLong o...@delong.com Date: Mon, 16 May 2011 16:12:27 -0700 ... It's not like you can even reach anything at home now, let alone reach it by name. that must and will change. let's be the generation who makes it possible.
Re: Yahoo and IPv6
Matthew Kaufman matt...@matthew.at writes: My Desktop is not able to make any IPv4 socket connections anymore. I get Protocol not supported. So there are IPv6-only users, already bitten by no . So that's -1 from me. Sounds to me like you're not on The Internet any more. in http://www.merit.edu/mail.archives/nanog/2001-04/msg00294.html we see: (*2)Q: But what IS the Internet? A: It's the largest equivalence class in the reflexive, transitive, symmetric, closure of the relationship 'can be reached by an IP packet from'. Seth Breidbart by which definition, matthew's observation would be correct. folks who want to run V6 only and still be on the internet will need proxies for a long while. folks who want to run V6 only *today* and not have any proxies *today* are sort of on their own -- the industry will not cater to market non-forces. -- Paul Vixie KI6YSY
Re: Yahoo and IPv6
From: Marshall Eubanks t...@americafree.tv Date: Sat, 14 May 2011 13:02:16 -0400 I think that the real question is, when will people who are running IPv4 only not be on the Internet by this definition ? is there an online betting mechanism we could use, that we all think will still be in business decades from now when the truth is known? if we're going to start picking the month and year when IPv4 is the new PDP-11 compatibility mode (that's a VAX reference), where the winner is whoever comes closest without going over, my pick is July 2021, and i'm in for $50.
Re: Yahoo and IPv6
Jim Gettys j...@freedesktop.org writes: ... we have to get naming squared away. Typing IPv6 addresses is for the birds, and having everyone have to go fuss with a DNS provider isn't a viable solution. perhaps i'm too close to the problem because that solution looks quite viable to me. dns providers who don't keep up with the market (which means ipv6 and dnssec in this context) will lose business to those who do. -- Paul Vixie KI6YSY
Re: NTT as a service provider in the US
powerzo...@gmail.com powerzo...@gmail.com writes: Anyone have any thoughts on NTT as a service provider in the US ? Anyone currently or previously using them please chime in. can't do it. i have thoughts but i won't answer a freemail address. i'm taking the time to say so because your post looks like trolling to me. if you ask again with a real domain name and a real meatspace signature, i'll be happy to say what i think about ntt as a service provider in the US. -- Paul Vixie KI6YSY
Re: Mac OS X 10.7, still no DHCPv6
there are two replies here. --- Christopher Morrow morrowc.li...@gmail.com writes: ..., what's the harm in dhcpv6? (different strokes and all that) only the egos and reputations of those who said that stateless autoconf was all ipv6 needed. (which is a small price to pay, according to me.) --- Dobbins, Roland rdobb...@arbor.net writes: On Feb 28, 2011, at 10:47 AM, Steven Bellovin wrote: Also don't forget privacy-enhanced addresses. Yes, which have extremely negative opsec connotations in terms of complicating traceback. /64 csma subnets with low order 64 bits controlled by infectable pc's means we'll be blackholing by /64 when we blackhole in ipv6. it's no big deal. -- Paul Vixie KI6YSY
Re: Leasing of space via non-connectivity providers
Date: Thu, 10 Feb 2011 01:13:49 -0600 From: Jimmy Hess mysi...@gmail.com With them not requiring a /8 in the first place (after CIDR); one begins to wonder how much of their /8 allocations they actually touched in any meaningful way. i expect that after final depletion there will be some paid transfers from some of the large legacy blocks. i have no personal knowledge of HP's situation or indeed any /8 holder's situation, but if the market value of these transfers ever meaningfully exceeds the renumbering penalty then their beancounters will find a way to get it done. that's the way of the world. i can imagine this NOT happening. most businesses are looking for long term strategic investments not quick-fix short-term band-aids. a buddy loaned me a macbook after my thinkpad was stolen, and i loved it, and i went down to the apple store to buy one of my own just like my buddy's loaner and they said we only sell that with the chicklet keyboard now and i tried it and hated it. i could buy my buddy's laptop but without applecare and without the ability to replace it if it's lost/stolen i'm not willing to make that investment. so for me it's another thinkpad. so if a company who traditionally needs a lot of IPv4 to grow their network knows that they can get one last quarter's worth of it from some legacy /8 holder, they might do some kind of paid transfer, or they might just hum some dire straits and keep moving with their ipv6 plans. Now it's past last call for alcohol Past recall has been here and gone The landlord finally paid us all The satin jazzmen have put away their horns And we're standing outside of this wonderland Looking so bereaved and so bereft Like a Bowery bum when he finally understands The bottle's empty and there's nothing left (Your Latest Trick) for some IPv4 based businesses a /8 would be more than a lifetime supply, but there's a value ceiling imposed by the space other people can get. (when everybody else has made other arrangements, the relative value of one's own hoard decreases.) Perhaps the RIRs should personally and directly ask each /8 legacy holder to provide account of their utilization (which portions of the allocation is used, how many hosts), and ASK for each unused /22 [or shorter] to be returned. The legacy holders might (or might not) refuse. They might (or might not) tell the RIRs Hell no In any case, the registry should ASK and publish an indication for each legacy /8 at least. So the community will know which (if any) legacy /8 holders are likely to be returning the community's IPv4 addresses that they obtained but don't have need for. The community should also know which /8 legacy holders say Hell no, we're keeping all our /8s, and not telling you how much of the community's IPv4 resources we're actually using. this gets into the controversial topic of an RIR's standing with respect to legacy space, and i'll leave that to the lawyers to talk about. but as owen and others have said, if a legacy holder were approached in this way knowing that their answer was going to be on the public record in the way, they probably would see no incentive at all to answer the question.
Re: Leasing of space via non-connectivity providers
David Conrad d...@virtualized.org writes: I'm curious: when HP acquired the assets of Compaq (or when Compaq acquired the assets of Digital), is it your position that HP (or Compaq) met the same criteria as if they were requesting an IP address directly from the IR. for 16.0.0.0/8? since i was the guy to do the initial carving on 16.0.0.0/8 i pondered this at the time of the CPQ and HP acquisitions. my research revealed that the network that DEC had numbered using 16.0.0.0/8 was still in existence and had been part of the acquisition process. there's an interesting question as to whether the acquirer should have had to renumber, since the acquirer had their own /8 and probably had the ability to contain both the old and new networks in the same /8. there's another interesting question as to whether either DEC or HP could have qualified for a /8 under current rules, since the basis for these (pre-RIR) allocations was that they needed more than a /16 and these were the days before CIDR. (at the time i received the /8 allocation at DEC, we had a half dozen /16's several dozen /24's that we wanted to stop using because we worried about the size of the global routing table... what whacky kids we all were. hint: i had hair back then.) -- Paul Vixie KI6YSY
Re: Verizon acquiring Terremark
Date: Wed, 2 Feb 2011 03:22:39 -0500 From: Jeffrey Lyon jeffrey.l...@blacklotus.net I'm sure everything will be fine in practice as others have indicated, I was merely making a point of the inherent conflict of interest. ah. if you mean it's unusual or it's difficult rather than it cannot be then i have no arguments. the reason PAIX got traction at all, coming late to the game (1995-ish) as we did, was because MFS was then able to charge circuit prices for many forms of cross connect down at MAE West. and i did face continuous pressure from MFN to go after a share of PAIX's carrier's circuit revenue. (which i never did and which none of my successors have done either.) noting, the game as moved on. if verizon behaves badly as terremark's owner then the presence of equinix in the market will act as a relief valve. i think the neutral and commercial model is very well established and that verizon will not want to be the only carrier in those facilities nor have their circuit-holders be the only customers for the real estate. it's an awful lot of space to use just as colo, and it's both over- and underbuilt for colo (vs. an IX). re: On Wed, Feb 2, 2011 at 1:38 AM, Paul Vixie vi...@isc.org wrote: Jeffrey Lyon jeffrey.l...@blacklotus.net writes: One cannot be owned by a carrier and remain carrier neutral. My two cents, my experience running PAIX when it was owned by MFN was not like you're saying.
Re: Verizon acquiring Terremark
Jeffrey Lyon jeffrey.l...@blacklotus.net writes: One cannot be owned by a carrier and remain carrier neutral. My two cents, my experience running PAIX when it was owned by MFN was not like you're saying. -- Paul Vixie KI6YSY
Re: [arin-announce] ARIN Resource Certification Update
From: Alex Band al...@ripe.net Date: Sun, 30 Jan 2011 11:39:36 +0100 I think my question is very pertinent. Of course the number of signed prefixes directly influences the number of validators. Do you think the RIPE NCC Validator tool would have been downloaded over 100 times in the last month if there were only 5 certified prefixes? i think we may be talking past each other. the number of production validators will be unrelated to any difference between prefixes signed because signing is easy and prefixes signed because operators are willing to do something hard. the operators who will sign even if it's hard (for example, deploying up/down) and also the operators who will only do it if it's easy (for example, hosted at an RIR) will each not care how many production validators there are at that moment -- their decision will be made on some other basis. Practically, in the real world, why would anyone invest time and effort in altering their current BGP decision making process to accommodate for resource certification if the technology is on nobody's radar, it's hard to get your feet wet and there are just a handful of certified prefixes out there. Wouldn't it be good if network operators think: Because it helps increase global routing security, it's easy to get started and lots of people are already involved, perhaps I should have a look at (both sides of) resource certification too. the reasoning you're describing is what we had in mind when we built DLV as an early deployment aid for DNSSEC. we had to break stiction where if there were no validators there would be incentive to sign, and if there were no signatures there would be no incentive to validate. are you likewise proposing the hosted solution only as an early deployment aid? i'm really quite curious as to whether you'll continue operating an RPKI hosting capability even if it becomes unnec'y (as proved some day if many operators of all sizes demonstrate capability for up/down). if so, can you share the reasoning behind that business decision? i know it sounds like i'm arguing against a hosted solution, but i'm not. i'm just saying that network operators are going to make business decisions (comparing cost and risk to benefit) as to whether to sign and whether to validate, and RIR's are going to make business decisions (comparing cost and risk to benefit) as to what provisioning mode to offer, and i don't plan to try to tell any network operators to sign or validate based on my own criteria, nor do i plan to try to tell any RIR that they should host or do up/down based on my own criteria. it's their own criteria that matters. let's just get the best starting conditions we can get, and then expect that everybody will make the best decision they can make based on those conditions. at ISC i have been extremely interested in participating in RPKI development and i think that sra and randy (and the whole RPKI team inside IETF and among the RIRs) have done great work improving the starting conditions for anyone who has to make a business decision of whether to deploy and if so what mode to deploy in. on the ARIN BoT i have likewise been very interested in and supportive of RPKI and i'm happy to repeat john curran's words which were, ARIN is looking at the risks and benefits of various RPKI deployment scenarios, and we expect to do more public and member outreach and consultation at our upcoming meeting in san juan PR. Paul Vixie Chairman and Chief Scientist, ISC Member, ARIN BoT re: i don't agree that that question is pertinent. in deployment scenario planning i've come up with three alternatives and this question is not relevant to any of them. perhaps you know a fourth alternative. here are mine. 1. people who receive routes will prefer signed vs. unsigned, and other people who can sign routes will sign them if it's easy (for example, hosted) but not if it's too hard (for example, up/down). 2. same as #1 except people who really care about their routes (like banks or asp's) will sign them even if it is hard (for example, up/down). 3. people who receive routes will ignore any unsigned routes they hear, and everyone who can sign routes will sign them no matter how hard it is. i do not expect to live long enough to see #3. the difference between #1 and #2 depends on the number of validators not the number of signed routes (since it's an incentive question). therefore small differences in the size of the set of signed routes does not matter very much in 2011, and the risk:benefit profile of hosted vs. up/down still matters far more. ...
Re: [arin-announce] ARIN Resource Certification Update
From: Alex Band al...@ripe.net Date: Sat, 29 Jan 2011 16:26:55 +0100 ... So the question is, if the RIPE NCC would have required everyone to run their own certification setup using the open source tool-sets Randy mentions, would there be this much certified address space now? i don't agree that that question is pertinent. in deployment scenario planning i've come up with three alternatives and this question is not relevant to any of them. perhaps you know a fourth alternative. here are mine. 1. people who receive routes will prefer signed vs. unsigned, and other people who can sign routes will sign them if it's easy (for example, hosted) but not if it's too hard (for example, up/down). 2. same as #1 except people who really care about their routes (like banks or asp's) will sign them even if it is hard (for example, up/down). 3. people who receive routes will ignore any unsigned routes they hear, and everyone who can sign routes will sign them no matter how hard it is. i do not expect to live long enough to see #3. the difference between #1 and #2 depends on the number of validators not the number of signed routes (since it's an incentive question). therefore small differences in the size of the set of signed routes does not matter very much in 2011, and the risk:benefit profile of hosted vs. up/down still matters far more. Looking at the depletion of IPv4 address space, it's going to be crucially important to have validatable proof who is the legitimate holder of Internet resources. I fear that by not offering a hosted certification solution, real world adoption rates will rival those of IPv6 and DNSSEC. Can the Internet community afford that? while i am expecting a rise in address piracy following depletion, i am not expecting #3 (see above) and i think most of the piracy will be of fallow or idle address space that will therefore have no competing route (signed or otherwise). this will become more pronounced as address space holders who care about this and worry about this sign their routes -- the pirates will go after easier prey. so again we see no material difference between hosted and up/down on the deployment side or if there is a difference it is much smaller than the risk:benefit profile difference on the provisioning side. in summary, i am excited about RPKI and i've been pushing hard for in both my day job and inside the ARIN BoT, but... let's not overstate the case for it or kneejerk our way into provisioning models whose business sense has not been closely evaluated. as john curran said, ARIN will look to the community for the guideance he needs on this question. i hope to see many of you at the upcoming ARIN public policy meeting in san juan PR where this is sure to be discussed both at the podium and in the hallways and bar rooms. Paul Vixie Chairman and Chief Scientist, ISC Member, ARIN BoT
Re: AltDB?
Date: Sat, 08 Jan 2011 15:47:51 +0900 From: Randy Bush ra...@psg.com ... more recent rumors, and john's posting here, seem to indicate that ... even to the extent that i know what's really happened or happening, i'd be loathe to comment on rumours. i have high confidence in arin's board and staff, and i believe that the right things are happening, even with the delays. right things as in what's best for the community and for the internet industry in the arin service region. as a strong proponent of rpki and of all things like rpki that will strengthen infrastructure, i remain delay-tolerant if review is the cost of getting it right. first, it would really help if the arin bot and management were much more open about these issues and decisions. at the detailed level. we are all not fools out here, present company excepted :). for a radical example, considering that arin is managing a public resource for the community, why are bot meetings not streamed a la cspan? can you cite some examples of nonprofit companies whose boards operate at the level of transparency you're asking me to consider in this example? the process of rolling out something like rpki involves some checks and balances, it's no longer just a simple matter of the technical people doing the right thing even though i remember older times when that was the way most things on the internet worked. i do not see how you are going to get rid of the liability. you have it now in whois/irr if i use it for routing (except they are so widely known to be bad data that the world knows i would be a fool to bet on them). whether the source of a roa is a user whacking on an arin web page or by other means, you still attested to the rights to that address space. my own belief here (not speaking for ARIN or for the ARIN BoT) is that the folks who use IRR/whois data to build route filters have a confidence level much lower than those who will use RPKI to do the same will have. i know that if i still had enable on anything other than my home router, that's how i'd feel. also, liability isn't just got rid of it's also documented and risk-managed, and doing that may require some kind of internal review. but all this is based on inference and rumor. can you please be more open and direct about this? thanks. i don't know. john (speaking for ARIN) gave an excellent and complete answer that i completely agree with. you're repeating some rumours which i won't comment on one way or the other. if you have specific questions which were not answered by john's response or which were raised by john's response you should ask them. saying i heard a rumour, would anyone care to refute it? is not going to move the conversational line of scrimmage at all. paul
Re: AltDB?
From: David Conrad d...@virtualized.org Date: Fri, 7 Jan 2011 21:01:52 -1000 do you have a specific proposal? i've noted in the past that arin tries hard to stick to its knitting, which is allocation and allocation policy. Yes. This is a positive (IMHO), however it seems that occasionally, ARIN's knitting tangles up folks who don't necessarily involve themselves with ARIN's existing interaction mechanisms (at least directly). the price of changing what ARIN does is, at a minimum: participation. it seems to me that if some in the community wanted arin to run SIGs or WGs on things like routing policy arin could do it but that a lot of folks would say that's mission creep and that it would be arin poaching on nanog lands. The issue I see is that there are non-address allocation{, policy} topics that can deeply affect network operations in which ARIN has a direct role, yet network operators (outside of the normal ARIN participants) have no obvious mechanism in which to comment/discuss/etc. Examples would include reverse DNS operations, whois database-related issues (operations, schema, access methods, etc.), (potentially?) RPKI, etc. It doesn't seem appropriate to me for these to be discussed in relation to addressing policy nor are the issues associated with those examples necessarily related to address allocation, hence I wouldn't think they'd be fodder for ppml. they are, though. i understand the subtlety of the question, is that a policy matter? but discussions on ppml@ have led to determinations of what is lameness? and when is a nameserver so lame that it's better to remove it from in-addr than to leave it in? i hear in what you're saying a desire to have a way to impact ARIN's behaviour outside of NRPM edits and perhaps ARIN does need to address this with some new online forum for things which aren't allocation policy but which should still be decided using community input. (as i recall my first act as a new ARIN trustee was to sign onto a policy proposal that would have changed the way e-mail templates worked, and at the end of the process the ARIN BoT shot it down because it wasn't a policy, and i understood that decision. strange, eh?) ... So, in other words, no, I don't really have a specific proposal. perhaps others will chime in. i will continue to think about it also.
Re: AltDB?
From: David Conrad d...@virtualized.org Date: Fri, 7 Jan 2011 23:11:32 -1000 On Jan 7, 2011, at 10:24 PM, Paul Vixie wrote: the price of changing what ARIN does is, at a minimum: participation. Another view is that ARIN's whole and sole reason for being is to provide services to the network operators in the ARIN region. yes. As such, it would be ill-advised for ARIN to change those services without consulting the community that ARIN serves and getting their buy-in. that's very much what i mean by participation. arin could never exist without a community to serve. if there are better ways to serve the community or better ways for the community to participate in steering arin's services, then i'm very interested in discovering them. Hopefully, there's a middle ground. this *is* the middle ground. we're beyond the span of decades when a couple of smart engineers could bang out a working solution that the rest of the community would just adopt out of opportunity and inertia. and let's not just blame-the-lawyers for that. the stakeholders in the infrastructure of the information economy now number in the 'many' and their views and needs have to be represented in the decisions that get made by places like ICANN, IETF, the RIRs, and similar. i hear in what you're saying a desire to have a way to impact ARIN's behaviour outside of NRPM edits and perhaps ARIN does need to address this with some new online forum for things which aren't allocation policy but which should still be decided using community input. Yep. Not sure it should be an ARIN-operated thing (nor am I sure that it shouldn't be), but something a bit more focused on the operation of services ARIN provides than ppml might be helpful. count me as 'intrigued' and expect me to be thinking more about this.
Re: AltDB?
Date: Sat, 08 Jan 2011 18:17:55 +0900 From: Randy Bush ra...@psg.com let me be a bit more clear on this thanks. o you affect the operational community, you talk with (not to) the operational community where the operational community talks i think arin does this today. certainly that is the intent. on the other fork of this thread, drc has noted some ways that this engagement area can be further improved, and i have counted myself as intrigued. also, i neglected to mention in my earlier notes on this thread that in addition to public policy meetings and the public policy mailing list which are open to the entire community not just arin members and which allow for remote participation not just those who can travel, arin has a consultation and suggestion process (URL below). i urge all operators and interested parties of the operational community to consider sharing their perspectives and their wisdom with arin to guide it going forward. ARIN Consultation and Suggestion Process: https://www.arin.net/participate/acsp/index.html ARIN Public Policy Mailing List: http://lists.arin.net/mailman/listinfo/arin-ppml Meetings: https://www.arin.net/participate/meetings/index.html https://www.arin.net/participate/meetings/reports/ARIN_XXVI/index.html https://www.arin.net/participate/meetings/ARIN-XXVI/remote.html https://www.arin.net/participate/meetings/ARIN-XXVII/index.html https://www.arin.net/participate/meetings/ARIN-XXVIII/index.html Fellowships: https://www.arin.net/participate/meetings/fellowship.html Scholarships: https://www.arin.net/participate/meetings/scholarships.html
Re: AltDB?
note that while i am also an ARIN trustee, i am speaking here as what randy calls just another bozo on this bus. for further background, ISC has done some rpki work and everybody at ISC including me likes rpki just fine. when the ARIN board was first considering funding ISC to do some early rpki work, went out into the hallway until the discussion was over (ending positively.) On Jan 5, 2011, at 12:32 PM, Randy Bush wrote: i have a rumor that arin is delaying and possibly not doing rpki that seems to have been announced on the ppml list (to which i do not subscribe). john curran has explained that arin is doing its due diligence on some concerns that were brought up during a review of the rpki rollout. there is no sense in which arin has said that it is not doing rpki although the current review does technically qualify as delaying rpki. i'm treating the above rumour as false. David Conrad d...@virtualized.org writes: I heard about the delay, but not about ARIN possibly not doing RPKI. That would be ... surprising. [...] it would be very much surprising to me as well. [bush] as it has impact on routing, not address policy, across north america and, in fact the globe, one would think it would be announced and discussed a bit more openly and widely. even if i thought that the operational impact could be felt in these early days when rpki remains an almost completely nonproduction service, and i don't think this by the way, i would still say that an internal review of a new service is not really something the whole community cares about. [conrad] The definition of what comes under the public policy mailing list umbrella has always been a bit confusing to me. Too bad something like the APNIC SIGs and RIPE Working Groups don't really exist in the ARIN region. do you have a specific proposal? i've noted in the past that arin tries hard to stick to its knitting, which is allocation and allocation policy. it seems to me that if some in the community wanted arin to run SIGs or WGs on things like routing policy arin could do it but that a lot of folks would say that's mission creep and that it would be arin poaching on nanog lands. -- Paul Vixie Chairman and Chief Scientist, ISC Trustee, ARIN
Re: Comcast enables 6to4 relays
John Jason Brzozowski john_brzozow...@cable.comcast.com writes: This does not alter our plans for our native dual stack trials, in fact, I hope to have more news on this front soon. comcast native dual stack is working fine at my house. traceroute6 -q1 mol.redbarn.org shows details.
Re: [Bruce Hoffman] Thank-you for your recent participation.
Rich Kulawiec r...@gsp.org writes: Amusingly, this was sent to me *after* I replied to ab...@internap complaining about getting spammed. This spam came from the icontact spammers-for-hire: they're absolute filth who have been abusing individuals and mailing lists for years. I recommend blacklisting them permanently. domains and/or cidrs, plz? -- Paul Vixie KI6YSY
Re: Nato warns of strike against cyber attackers
d...@bungi.com (Dave Rand) writes: ... With more than 100,000,000 compromised computers out there, it's really time for us to step up to the plate, and make this happen. +1. -- Paul Vixie KI6YSY
Re: getting the hint
Larry Sheldon larryshel...@cox.net writes: The only response that works -- and even this is not guaranteed -- is shunning. Drop the message. Do not respond. Ever. And for the love of Pete, when somebody (as I have) makes a mistake and does his bidding, tell the miscreant VIA PRIVATE EMAIL or a note tied to a brick, but do not prate incessantly about it on the list. +1. -- Paul Vixie KI6YSY
Re: legacy /8
From: David Conrad d...@virtualized.org Date: Sun, 11 Apr 2010 13:52:24 -1000 On Apr 11, 2010, at 10:57 AM, Paul Vixie wrote: ... i'd like to pick the easiest problem and for that reason i'm urging dual-stack ipv4/ipv6 for all networks new or old. Is anyone arguing against this? yes. plenty of people have accused ipv6 of being a solution in search of a problem. on this very mailing list within the last 72 hours i've seen another person assert that ipv6 isn't needed. while i tend to agree with tony li who of ipv6 famously said it was too little and too soon we have been Overtaken By Events and we now have to deploy it or else. the only way we're going to do that is with widescale dual-stack, either native dual-stack (which is generally easy since ipv6 address space is cheap and plentiful) or dual-stack-lite (which is ipv4-NAT ipv6-native with aggregated encap/decap at the POP or edge) or with any other method (or trick) that comes to mind or seems attractive. what we can't do is presume that any form of ipv4 steady state forever or wait for something better than ipv6 before abandoning ipv4 is practical, or that these would be less expensive (in both direct cost, indirect cost, and network/market stability) than dual-stack now, ipv6-mostly soon, and ipv6-only eventually. The problem is what happens when there isn't sufficient IPv4 to do dual stack. that problem has many low hanging solutions, some of which mark andrews gave in his response to your note. one popular address allocation policy proposal is reserving the last IPv4 /8 for use in IPv6 deployment, for example as public-facing dual-stack-lite gateways. which brings me to the subject of address allocation policies, and meetings that happen periodically to discuss same. one such address allocator is ARIN (American Registry for Internet Numbers) and one such public policy meeting is next week in toronto. details of this meeting can be found at: https://www.arin.net/participate/meetings/ARIN-XXV/ and anyone, not just from the ARIN service region and not just ARIN members, can attend. there are also remote participation options, see above web page. -- Paul Vixie Chairman, ARIN BoT
Re: legacy /8
William Warren hescomins...@emmanuelcomputerconsulting.com writes: We've been dealing with the IPV4 myth now for over 7 years that i have followed it. It's about as valid as the exaflood myth. Part fo the reason folks aren't rushing to the V6 bandwagon is it's not needed. Stop doing the chicken little dance folks. V6 is nice and gives us tons of more addresses but I can tell you V4 is more than two years form dying just by seeing all the arm flailing going around. anyone claiming that the then-existing ipv4 will stop working when the free pool runs out is indeed doing the chicken little dance. however, for many networks, growth is life, and for them, free pool depletion is a problem. -- Paul Vixie Chairman, ARIN BoT
Re: legacy /8
David Conrad d...@virtualized.org writes: Growth in IPv4 accessible hosts will stop or become significantly more expensive or both in about 2.5 years (+/- 6 months). Growth stopping is extremely unlikely. Growth becoming significantly more expensive is guaranteed. ... more expensive for whom, though? if someone has to find existing address space and transfer it at some payment to the current holder in order to grow their own network, that's a direct expense. if this became common and resulted in significant deaggregation, then everybody else attached in some way to the global routing table would also have to pay some costs, which would be indirect expenses. unless a market in routing slots appears, there's no way for the direct beneficiaries of deaggregation to underwrite the indirect costs of same. at a systemic level, i'd characterize the cost of that kind of growth as instability rather merely expense. Address utilization efficiency will increase as people see the value in public IPv4 addresses. ISPs interested in continuing to grow will do what it takes to obtain IPv4 addresses and folks with allocated- but- unused addresses will be happy to oblige (particularly when they accept that they only need a couple of public IP addresses for their entire network). At some point, it may be that the cost of obtaining IPv4 will outstrip the cost of migrating to IPv6. If we're lucky. the cost:benefit of using ipv6 depends on what other people have deployed. that is, when most of the people that an operator and their customers want to talk to are already running ipv6, then the cost:benefit will be compelling compared to any form of continued use of ipv4. arguments about the nature and location of that tipping point amount to reading tea leaves. nevertheless if everybody who can deploy dual-stack does so, we'll reach that tipping point sooner and it'll be less spectacular. -- Paul Vixie Chairman, ARIN BoT
Re: legacy /8
From: David Conrad d...@virtualized.org Date: Sun, 11 Apr 2010 10:30:05 -1000 unless a market in routing slots appears, there's no way for the direct beneficiaries of deaggregation to underwrite the indirect costs of same. And that's different from how it's always been in what way? when 64MB was all anybody had, deaggregation was rendered ineffective by route filtering. what i've seen more recently is gradual monotonic increase in the size of the full table. if the systemic cost of using all of ipv4 includes a 10X per year step function in routing table size then it will manifest as instability (in both the network and the market). as you have pointed out many times, ipv6 offers the same number of /32's as ipv4. however, a /32 worth of ipv6 is enough for a lifetime even for most multinationals, whereas for ipv4 it's one NAT or ALG box. so, i'm thinking that making ipv4 growth happen beyond pool exhaustion would be a piecemeal affair and that the routing system wouldn't accomodate it painlessly. the rate of expansion of other people's routers seems to fit the growth curve we've seen, but will it fit massive deaggregation? My tea leaf reading is that history will repeat itself. As it was in the mid-90's, as soon as routers fall over ISPs will deploy prefix length (or other) filters to protect their own infrastructure as everybody scrambles to come up with some hack that won't be a solution, but will allow folks to limp along. Over time, router vendors will improve their kit, ISPs will rotate out routers that can't deal with the size/flux of the bigger routing table (passing the cost on to their customers, of course), and commercial pressures will force the removal of filters. Until the next go around since IPv6 doesn't solve the routing scalability problem. instability like we had in the mid-1990's would be far more costly today, given that the internet is now used by the general population and serves a global economy. if the rate of endpoint growth does not continue beyond ipv4 pool exhaustion we'll have a problem. if it does, we'll also have a problem but a different problem. i'd like to pick the easiest problem and for that reason i'm urging dual-stack ipv4/ipv6 for all networks new or old. -- Paul Vixie Chairman, ARIN BoT
Re: DNS server software
Claudio Lapidus clapi...@gmail.com writes: We are a mid-sized carrier (1.2M broadband subscribers) and we are looking for an upgrade in our public DNS resolver infrastructure, so we are interested in getting to know what are you guys using in your networks. Mainly what kind/brand of software and which architecture did you use to deploy it, and how did you do the sizing, all of it would be most helpful information. Unsurprisingly, we (AS1280, AS3557) run BIND 9. see http://www.isc.org/. We have at least two recursives in each AS1280 site, and one in each AS3557 location (f-root). Stubs (either /etc/resolv.conf or DHCP) each use all local plus some non-local, for a minimum of three total. Recursive DNS servers do not use forwarding or other cache-sharing techniques, each is fully independent. Most have DNSSEC validation enabled, and of those, all are subscribed to ISC DLV, see http://dlv.isc.org/. Most server hosts here run FreeBSD on AMD64/EM64T or else i386. -- Paul Vixie KI6YSY
Re: Spamhaus...
Rich Kulawiec r...@gsp.org writes: On Fri, Feb 19, 2010 at 08:20:36PM -0500, William Herrin wrote: Whine all you want about backscatter but until you propose a comprehensive solution that's still reasonably compatible with RFC 2821's section 3.7 you're just talking trash. We're well past that. Every minimally-competent postmaster on this planet knows that clause became operationally obsolete years ago [1], and has configured their mail systems to always reject, never bounce. [2] for smtp, i agree. yet, uucp and other non-smtp last miles are not dead. [2] Yes, there are occasionally some edge cases of limited scope and duration that can be tough to handle. ... The key points here are limited scope and limited duration. There is never any reason or need in any mail environment to permit these problems to grow beyond those boundaries. so, a uucp-only site should have upgraded to real smtp by now, and by not doing it they and their internet gateway are a joint menace to society? that seems overly harsh. there was a time (1986 or so?) when most of the MX RR's in DNS were smtp gateways for uucp-connected (or decnet-connected, etc) nodes. it was never possible to reject nonexist...@uucpconnected at their gateway since the gateway didn't know what existed or not. i'm not ready to declare that era dead. william herrin had a pretty good list of suggested tests to avoid sending useless bounce messages: No bounce if the message claimed to be from a mailing list. No bounce if the spam scored higher than 8 in spamassassin No bounce if the server which you received the spam from doesn't match my domain's published SPF records evaluated as if ~all and ?all are -all i think if RFC 2821 is to be updated to address the backscatter problem, it ought to be along those lines, rather than everything must be synchronous. -- Paul Vixie KI6YSY
Re: DNS queries for . IN A return rcode 2 SERVFAIL from windows DNS recursing resolvers
Joe Maimon jmai...@ttec.com writes: Hey all, This must be old news for everyone else. While looking at a dns monitor on a load balancer that defaulted to . A queries to check liveliness on DNS resolvers, it became quite clear that windows 2000/2003 DNS server appears to return rcode=2 for queries looking for an A record for the root. The resolvers appear to work properly in all other regards. well, there is no A RR for the root domain. RCODE=2 is still an error, you should receive RCODE=0 ANCOUNT=0 for an unused RR type. but many resolvers get confused when the root domain is the QNAME, so let's assume that you're using one of those. So the monitors were switched to localhost. A (Is this a bad idea?) probably. there is no localhost in the root zone. this name is a TCP/IP stack convention, not a standard. for health monitoring purposes you should probably choose one of your own local names, since there's almost certainly no local intelligence in your resolver about them. that means to look up one of your own names the resolver probably has to iterate downward from the root zone to the top level and all the way down to your authority nameservers. (the problem here is, you may be testing more than you intend, and a failure in your own authority server or in the delegation path to it would look the same as an IP path failure or a resolver problem.) A little testing later and the results for . A are: Windows NT 4, ancount=0, authority=1, rcode=0 Windows 2000, rcode=2 Windows 2003, rcode=2 bind, ancount=0, authority=1, rcode=0 To my (inexpert) eyes that doesnt seem quite right. probably resolver bugs, either in those TCP/IP stacks or in the recursive nameserver they are using. (is the same recursive nameserver used in all four tests?) I cant seem to find any online information regarding this difference of behavior. Enlightenment appreciated. i suggest re-asking this over on dns-operati...@lists.dns-oarc.net, since it a bit deep in the DNS bits for a general purpose list like NANOG. -- Paul Vixie KI6YSY
EDNS (Re: Are the Servers of Spamhaus.rg and blackholes.us down?)
Jason Bertoch ja...@i6ix.com writes: Dec 31 10:12:37 linux-1ij2 named[14306]: too many timeouts resolving 'XXX.YYY.ZZZ/A' (in 'YYY.ZZZ'?): disabling EDNS Do you have a firewall in front of this server that limits DNS packets to 512 bytes? statistically speaking, yes, most people have that. which is damnfoolery, but well supported by the vendors, who think either that udp/53 datagrams larger than 512 octets are amplification attacks, or that udp packets having no port numbers because they are fragments lacking any udp port information, are evil and dangerous. sadly, noone has yet been fired for buying devices that implement this kind of overspecification. hopefully that will change after the DNS root zone is signed and udp/53 responses start to generally include DNSSEC signatures, pushing most of them way over the 512 octet limit. it's going to be another game of chicken -- will the people who build and/or deploy such crapware lose their jobs, or will ICANN back down from DNSSEC? -- Paul Vixie KI6YSY
Re: EDNS (Re: Are the Servers of Spamhaus.rg and blackholes.us down?)
Date: Fri, 1 Jan 2010 22:16:31 + From: bmann...@vacation.karoshi.com It would help if the BIND EDNS0 negotiation would not fall back to the 512 byte limit - perhaps you could talk with the ISC developers about that. i don't agree that your proposed change would help with this problem at all. but in any case nanog isn't the place to ask ISC to change BIND, nor is it the place to discuss protocol implementation or interpretation. i suggest bind-users@, bind-workers@, dns-operations@, dnsop@, and/or namedroppers@, depending on what aspect of your above-described concerns you focus on.
Re: Article on spammers and their infrastructure
Randy Bush ra...@psg.com writes: If ARIN and/or RIPE and/or ICANN and/or anyone else were truly interested in making a dent in the problem, then they would have already paid attention to our collective work product. the rirs, the ietf, the icann, ... each think they are the top of the mountain. we are supposed to come to them and pray. more likely that the itu will come to them and prey. ARIN (an RIR) does not think in terms of mountains. the staff and company does what members and the elected board and elected advisory council ask. ARIN is a 501(c)(6) and sticks to its knitting, which thus far means no distinguished role in spammers and their infrastructure but that could change if someone writes a policy proposal which is adopted after the normal policy development process. please do consider whether ARIN could help with spammers and their infrastructure and if so, write a policy draft to that effect. ARIN is responsive to community input, and has well established and well publicized mechanisms for receiving and processing community input. nobody has to come and pray, but likewise, nobody should expect ARIN to look for mission creep opportunities. ARIN will go on doing what the community asks, no less, no more. ARIN has no mechanism, as a company, for [paying] attention to [your] collective work product. our members, and the public at large who participates in ARIN's policy development process, do that. -- Paul Vixie Chairman, ARIN BoT KI6YSY
Re: DNS question, null MX records
Douglas Otis do...@mail-abuse.org writes: If MX TEST-NET became common, legitimate email handlers unable to validate messages prior to acceptance might find their server resource constrained when bouncing a large amount of spam as well. none of this will block spam. spammers do not follow RFC 974 today (since i see a lot of them come to my A RR rather than an MX RR, or in the wrong order). any well known pattern that says don't try to deliver e-mail here will only be honoured by friend people who don't want us to get e-mail we don't want to get. -- Paul Vixie KI6YSY
Re: DNS question, null MX records
Douglas Otis do...@mail-abuse.org writes: Agreed. But it will impact providers generating a large amount of bounce traffic, and some portion of spam sources that often start at lower priority MX records in an attempt to find backup servers without valid recipient information. In either case, this will not cause extraneous traffic to hit roots or ARPA. if you're just trying to stop blowback from forged-source spam, and not trying to stop the spam itself, then some mechanism like an unreachable MX does seem called for. note that those approaches will cause queuing on the blowerbackers, rather than outright reject/die. other approaches that could cause outright reject/die would likely direct the blowback to the blowback postmasters, who are as innocent as the spam victims. i'm not sure there's a right way to do this in current SMTP. i used to think we could offer to verify that a piece of e-mail had come from us using some kind of semi-opaque H(message-id) scheme, but in studying it i found that as usual with spam the economic incentives are all backwards. -- Paul Vixie KI6YSY
Re: Breaking the internet (hotels, guestnet style)
Steven Bellovin s...@cs.columbia.edu writes: It's why I run an ssh server on 443 somewhere -- and as needed, I ssh-tunnel http to a squid proxy, smtp, and as many IMAP/SSL connections as I really need... me too, more or less. but steve, if we were only trying to build digital infrastructure for people who know how to do that, then we'd all still be using Usenet over modems. we're trying to build digital infrastructure for all of humanity, and that means stuff like the above has to be unnecessary. -- Paul Vixie KI6YSY
Re: What DNS Is Not
From: David Conrad d...@virtualized.org Date: Thu, 26 Nov 2009 07:42:15 -0800 As you know, as long as people rely on their ISPs for resolution services, DNSSEC isn't going to help. Where things get really offensive if when the ISPs _require_ customers (through port 53 blocking, T-Mobile Hotspot, I'm looking at you) to use the ISP's resolution services. the endgame for provider-in-the-middle attacks is enduser validators, which is unfortunate since this use case is not well supported by current DNSSEC and so there's some more protocol work in our future (n!!). i also expect to see DNS carried via HTTPS, which providers tend to leave alone since they don't want to hear from the lawyers at 1-800-flowers.com. (so, get ready for https://ns.vix.com/dns/query/www.vix.com/in/ard=1ad=1).
Re: What DNS Is Not
From: David Conrad d...@virtualized.org Date: Thu, 26 Nov 2009 13:25:39 -0800 At some point, we may as well bite the bullet and redefine http{,s} as IPv7. since products and services designed to look inside encrypted streams and inspect, modify, or redirect them are illegal in most parts of the world: yes, inevitably.
Re: What DNS Is Not
Jorge Amodio jmamo...@gmail.com writes: What needs to be done to have ISPs and other service providers stop tampering with DNS ? we have to fix DNS so that provider-in-the-middle attacks no longer work. (this is why in spite of its technical excellence i am not a DNSCURVE fan, and also why in spite of its technical suckitude i'm working on DNSSEC.) http://queue.acm.org/detail.cfm?id=1647302 lays out this case. -- Paul Vixie KI6YSY
Re: What DNS Is Not
Kevin Oberman ober...@es.net writes: I find it mildly amusing that my first contact with Paul was about 25 years ago when he was at DEC and I objected to his use of a wildcard for dec.com. I was only an egg. The situations are not parallel and the Internet was a very different animal in those days (and DEC was mostly DECnet), but still I managed to maintain a full set of MX records for all of our DECnet systems. Based partly on my conversation with you, I ended up pulling over the list of DECnet nodes and generating MX's for each one, just to remove that wildcard. You were right, and I listened. Probably I forgot to thank you until now. Thanks. -- Paul Vixie KI6YSY
Re: What DNS Is Not
i loved the henry ford analogy -- but i think henry ford would have said that the automatic transmission was a huge step forward since he wanted everybody to have a car. i can't think of anything that's happened in the automobile market that henry ford wouldn't've wished he'd thought of. i knew that the incoherent DNS market would rise up on its hind legs and say all kinds of things in its defense against the ACM Queue article, and i'm not going to engage with every such speaker. there three more-specific replies below. Dave Temkin dav...@gmail.com writes: Alex Balashov wrote: For example, perhaps in the case of CDNs geographic optimisation should be in the province of routing (e.g. anycast) and not DNS? In most cases it already is. He completely fails to address the concept of Anycast DNS and assumes people are using statically mapped resolvers. anycast DNS appears to mean different things to different people. i didn't mention it because to me anycast dns is a bgp level construct whereby the same (coherent) answer is available from many servers having the same IP address but not actually being the same server. see for example how several root name servers are distributed. http://www.root-servers.org/. if you are using anycast DNS to mean carefully crafted (noncoherent) responses from a similarly distributed/advertised set of servers, then i did address your topic in the ACM Queue article. David Andersen d...@cs.cmu.edu writes: This myth ... was debunked years ago: DNS Performance and the Effectiveness of Caching Jaeyeon Jung, Emil Sit, Hari Balakrishnan, and Robert Morris http://pdos.csail.mit.edu/papers/dns:ton.pdf my reason for completely dismissing that paper at the time it came out was that it tried to predict the system level impact of DNS caching while only looking at the resolver side and only from one client population having a small and uniform user base. show me a trace driven simulation of the whole system, that takes into account significant authority servers (which would include root, tld, and amazon and google) as well as significant caching servers (which would not include MIT's or any university's but which would definitely include comcast's and cox's and att's), and i'll read it with high hopes. note that ISC SIE (see http://sie.isc.org/ may yet grow into a possible data source for this kind of study, which is one of the reasons we created it.) Simon Lyall si...@darkmere.gen.nz writes: I heard some anti-spam people use DNS to distribute big databases of information. I bet Vixie would have nasty things to say to the guy who first thought that up. someone made this same comment in the slashdot thread. my response there and here is: the MAPS RBL has always delivered coherent responses where the answer is an expressed fact, not kerned in any way based on the identity of the querier. perhaps my language in the ACM Queue article was imprecise (delivering facts rather than policy) and i should have stuck with the longer formulation (incoherent responses crafted based on the identity of the querier rather than on the authoritative data). -- Paul Vixie KI6YSY
Re: Gmail Down?
cabenth cabe...@gmail.com writes: Gmail is definitely been having a hard time this morning. Stand by for the comments from the peanut gallery about allowing google to scan your e-mail, etc like last time Gmail had an issue. i recently explored webmail for my family and found prayer, which is a pure C application (no php, no perl) built on the uw-imap c-client library. it's blindingly fast even for thousands of huge mailboxes stored in MH format. anyone who was using the cloud because they couldn't stand the poor performance of the apache-based webmail systems should take a look. http://www-uxsup.csx.cam.ac.uk/~dpc22/prayer/ is the home page. though i found it in freebsd /usr/ports/mail/prayer. -- Paul Vixie KI6YSY
Re: DNS hardening, was Re: Dan Kaminsky
Christopher Morrow morrowc.li...@gmail.com writes: how does SCTP ensure against spoofed or reflected attacks? there is no server side protocol control block required in SCTP. someone sends you a create association request, you send back a ok, here's your cookie and you're done until/unless they come back and say ok, here's my cookie, and here's my DNS request. so a spoofer doesn't get a cookie and a reflector doesn't burden a server any more than a ddos would do. because of the extra round trips nec'y to create an SCTP association (for which you can think, lightweight TCP-like session-like), it's going to be nec'y to leave associations in place between iterative caches and authority servers, and in place between stubs and iterative caches. however, because the state is mostly on the client side, a server with associations open to millions of clients at the same time is actually no big deal. -- Paul Vixie KI6YSY
Re: DNS hardening, was Re: Dan Kaminsky
note, i went off-topic in my previous note, and i'll be answering florian on namedroppers@ since it's not operational. chris's note was operational: Date: Thu, 6 Aug 2009 10:18:11 -0400 From: Christopher Morrow morrowc.li...@gmail.com awesome, how does that work with devices in the f-root-anycast design? (both local hosts in the rack and if I flip from rack to rack) If I send along a request to a host which I do not have an association created do I get a failure and then re-setup? (inducing further latency) yes. so, association setup cost will occur once per route-change event. note that the f-root-anycast design already hashes by flow within a rack to keep TCP from failing, so the only route-change events of interest to this point are in wide area BGP. ...: Do loadbalancers, or loadbalanced deployments, deal with this properly? (loadbalancers like F5, citrix, radware, cisco, etc...) as far as i know, no loadbalancer understands SCTP today. if they can be made to pass SCTP through unmodified and only do their enhanced L4 on UDP and TCP as they do now, all will be well. if not then a loadbalancer upgrade or removal will be nec'y for anyone who wants to deploy SCTP. it's interesting to me that existing deployments of L4-aware packet level devices can form a barrier to new kinds of L4. it's as if the internet is really just the web, and our networks are TCP/UDP networks not IP networks.
Re: Dan Kaminsky
Curtis Maurand cmaur...@xyonet.com writes: What does this have to do with Nanog, the guy found a critical security bug on DNS last year. He didn't find it. He only publicized it. the guy who wrote djbdns fount it years ago. first blood on both the DNS TXID attack, and on what we now call the Kashpureff attack, goes to chris schuba who published in 1993: http://ftp.cerias.purdue.edu/pub/papers/christoph-schuba/schuba-DNS-msthesis.pdf i didn't pay any special heed to it since there was no way to get enough bites at the apple due to negative caching. when i saw djb's announcement (i think in 1999 or 2000, so, seven years after schuba's paper came out) i said, geez, that's a lot of code complexity and kernel overhead for a problem that can occur at most once per DNS TTL. and sure enough when we did finally put source port randomization into BIND it crashed a bunch of kernels and firewalls and NATs, and is still paying painful dividends for large ISP's who are now forced to implement it. why forced? what was it about kaminsky's announcement that changed this from a once-per-TTL problem that didn't deserve this complex/costly solution into a once-per-packet problem that made the world sit up and care? if you don't know the answer off the top of your head, then maybe do some reading or ask somebody privately, rather than continuing to announce in public that bernstein's problem statement was the same as kaminsky's problem statement. and, always give credit to chris schuba, who got there first. Powerdns was patched for the flaw a year and a half before Kaminsky published his article. nevertheless bert was told about the problem and was given a lengthy window in which to test or improve his solutions for it. and i think openbsd may have had source port randomization first, since they do it in their kernel when you try to bind(2) to port 0. most kernels are still very predictable when they're assigning a UDP port to an outbound socket. -- Paul Vixie KI6YSY
Re: Fwd: Dan Kaminsky
William Allen Simpson william.allen.simp...@gmail.com writes: Are we paying enough attention to securing our systems? almost certainly not. skimming RFC 2196 again just now i find three things. 1. it's out of date and needs a refresh -- yo barb! 2. i'm not doing about half of what it recommends 3. my users complain bitterly about the other half in terms of cost:benefit, it's more and more the case that outsourcing looks cheaper than doing the job correctly in-house. not because outsourcing *is* more secure but because it gives the user somebody to sue rather than fire, where a lawsuit could recover some losses and firing someone usually won't. digital security is getting a lot of investor attention right now. i wonder if this will ever consolidate or if pandora's box is just broken for all time. -- Paul Vixie KI6YSY
Re: White House net security paper
Randy Bush ra...@psg.com writes: ... a few battalions of B's and C's, if wisely deployed, could bridge that gap. there is a reason Bs and Cs have spare round-tuits. fred brooks was no fool. os/360 taught some of us some lessons. batallions work in the infantry, or so i am told. this is rocket science. to me wisely means backfilling 80% of what the Good Guys do that isn't rocket science. (most A's are not doing only what only A's can do.) -- Paul Vixie KI6YSY
Re: White House net security paper
Randy Bush ra...@psg.com writes: As hire As. Bs hire Cs. Lots of Cs. this problem needs neurons, not battalions. this problem needs round-tuits, which Good Guys are consistently short of, but which Bad Guys always have as many of as they can find use for. a few battalions of B's and C's, if wisely deployed, could bridge that gap. the key to all this is therefore not really neurons but rather wiselyness. i promise to, um, mention this, or maybe more, in my nanog-philly keynote. -- Paul Vixie KI6YSY
Re: White House net security paper
Sean Donelan s...@donelan.com writes: How many ISPs have too many network security people? network security is a loss center. not just a cost center, a *loss* center. non-bankrupt ISP's whose investors will make good multiples only staff their *profit* centers. the Good Guys and Bad Guys all know this -- the difference is that the Good Guys try not to think about this whereas the Bad Guys think about it all the time. -- Paul Vixie KI6YSY
Re: Why choose 120 volts?
Leo Bicknell bickn...@ufp.org writes: ... http://www.apcmedia.com/salestools/NRAN-6CN8PK_R0_EN.pdf ... But what you'll find in the paper is that the change allows you to re-architect the power plant in a way that saves you money on PDU's, transformers, and other stuff. Thus this makes the most sense to consider in a green field deployment. noting also that architect is a noun, i find that on large plants the cost of copper wire and circuit breakers add up, where sizes (and prices) are based on ampherage not wattage. in the old days when a rack needed 6kW, that was 208V 30A (10 gauge wire) or it was two of 120V 30A (also 10 gauge wire). somewhere near the first hundred or so racks, the price of the wire and breakers starts to seem high, and very much worth halving. once in a while some crashcart CRT monitor won't run on anything but 120V but for $50 NRC it can be replaced with an LCD. everything else that's still worth plugging in (that is, having a power/heat cost per performance better than that of a blow dryer) doesn't care what voltage it lives on. -- Paul Vixie KI6YSY
Re: Colo on the West Coast
Pshem Kowalczyk pshe...@gmail.com writes: (answers can be off-list) See http://www.vix.com/personalcolo/. (updates still welcomed, btw.) -- Paul Vixie KI6YSY
Re: integrated KVMoIP and serial console terminal server
Owen DeLong o...@delong.com writes: My favorite front end for serial console management is conserver. It's really great software and the price is right. http://www.conserver.com see also /usr/ports/sysutils/rtty on freebsd, which pulls from MASTER_SITES= ftp://ftp.isc.org/isc/rtty/ \ ftp://gatekeeper.research.compaq.com/pub/misc/vixie/ since the ftp server mentioned here in 1996 http://www.merit.edu/mail.archives/nanog/1996-08/msg00223.html is dead. -- Paul Vixie KI6YSY
Re: IXP
Bill Woodcock wo...@pch.net writes: ... Nobody's arguing against VLANs. Paul's argument was that VLANs rendered shared subnets obsolete, and everybody else has been rebutting that. Not saying that VLANs shouldn't be used. i think i saw several folks, not just stephen, say virtual wire was how they'd do an IXP today if they had to start from scratch. i know that for many here, starting from scratch isn't a reachable worldview, and so i've tagged most of the defenses of shared subnets with that caveat. the question i was answering was from someone starting from scratch, and when starting an IXP from scratch, a shared subnet would be just crazy talk. -- Paul Vixie
Re: IXP
From: Paul Vixie vi...@isc.org Date: Sat, 18 Apr 2009 00:08:04 + ... i should answer something said earlier: yes there's only 14 bits of tag and yes 2**14 is 4096. in the sparsest and most wasteful allocation scheme, tags would be assigned 7:7 so there'd be a max of 64 peers. i meant of course 12 bits, that 2**12 is 4096, and 6:6. apologies for slop.
Re: IXP
stephen, any idea why this hasn't hit the nanog mailing list yet? it's been hours, and things that others have sent on this thread has appeared. is it stuck in a mail queue? --paul re: To: Deepak Jain dee...@ai.net cc: Matthew Moyle-Croft m...@internode.com.au, Arnold Nipper arn...@nipper.de, Paul Vixie vi...@isc.org, na...@merit.edu na...@merit.edu Subject: Re: IXP Date: Sat, 18 Apr 2009 05:30:41 + From: Stephen Stuart stu...@tech.org Not sure how switches handle HOL blocking with QinQ traffic across trunks, but hey... what's the fun of running an IXP without testing some limits? Indeed. Those with longer memories will remember that I used to regularly apologize at NANOG meetings for the DEC Gigaswitch/FDDI head-of-line blocking that all Gigaswitch-based IXPs experienced when some critical mass of OC3 backbone circuits was reached and the 100 MB/s fabric rolled over and died, offered here (again) as a cautionary tale for those who want to test those particular limits (again). At PAIX, when we upgraded to the Gigaswitch/FDDI (from a DELNI; we loved the DELNI), I actually used a feature of the switch that you could black out certain sections of the crossbar to prevent packets arriving on one port from exiting certain others at the request of some networks to align L2 connectivity with their peering agreements. It was fortunate that the scaling meltdown occurred when it did, otherwise I would have spent more software development resources trying to turn that capability into something that was operationally sustainable for networks to configure the visibility of their port to only those networks with which they had peering agreements. That software would probably have been thrown away with the Gigaswitches had it actually been developed, and rewritten to use something horrendous like MAC-based filtering, and if I recall correctly the options didn't look feasible at the time - and who wants to have to talk to a portal when doing a 2am emergency replacement of a linecard to change registered MAC addresses, anyway?. The port-based stuff had a chance of being operationally feasible. The notion of a partial pseudo-wire mesh, with a self-service portal to request/accept connections like the MAEs had for their ATM-based fabrics, follows pretty well from that and everything that's been learned by anyone about advancing the state of the art, and extends well to allow an IXP to have a distributed fabric benefit from scalable L2.5/L3 traffic management features while looking as much like wires to the networks using the IXP. If the gear currently deployed in IXP interconnection fabrics actually supports the necessary features, maybe someone will be brave enough to commit the software development resources necessary to try to make it an operational reality. If it requires capital investment, though, I suspect it'll be a while. The real lesson from the last fifteen or so years, though, is that bear skins and stone knives clearly have a long operational lifetime. Stephen
Re: IXP
Date: Sat, 18 Apr 2009 10:09:00 + From: bmann...@vacation.karoshi.com ... well... while there is a certain childlike obession with the byzantine, rube-goldburg, lots of bells, knobs, whistles type machines... for solid, predictable performance, simple clean machines work best. like you i long for the days when a DELNI could do this job. nobody makes hubs anymore though. but the above text juxtaposes poorly against the below text: Date: Sat, 18 Apr 2009 16:35:51 +0100 From: Nick Hilliard n...@foobar.org ... These days, we have switches which do multicast and broadcast storm control, unicast flood control, mac address counting, l2 and l3 acls, dynamic arp inspection, and they can all be configured to ignore bpdus in a variety of imaginative ways. We have arp sponges and broadcast monitors. ... in terms of solid and predictable i would take per-peering VLANs with IP addresses assigned by the peers themselves, over switches that do unicast flood control or which are configured to ignore bpdu's in imaginative ways. but either way it's not a DELNI any more. what i see is inevitable complexity and various different ways of layering that complexity in. the choice of per-peering VLANs represents a minimal response to the problems of shared IXP fabrics, with maximal impedance matching to the PNI's that inevitably follow successful shared-port peerings.
Re: IXP
Date: Sat, 18 Apr 2009 16:35:51 +0100 From: Nick Hilliard n...@foobar.org ... i just don't care if people use L2 connectivity to get to an exchange from a router somewhere else on their LAN. They have one mac address to play around with, and if they start leaking mac addresses towards the exchange fabric, all they're going to do is hose their own connectivity. yeah we did that at PAIX. if today's extremenetworks device has an option to learn one MAC address per port and no more, it's because we had a terrible time getting people to register their new MAC address when they'd change out interface cards or routers. hilarious levels of fingerpointing and downtime later, our switch vendor added a knob for us. but we still saw typo's in IP address configurations whereby someone could answer ARPs for somebody else's IP. when i left PAIX (the day MFN entered bankruptcy) we were negotiating for more switch knobs to prevent accidental and/or malicious ARP poisoning. (and note, this was on top of a no-L2-devices rule which included draconian auditing rights for L2/L3 capable hardware.) As you've noted, there is a natural progression for services providers here from shared access to pni, which advances according to the business and financial requirements of the parties involved. If exchange users decide to move from shared access peering to PNI, good for them - it means their business is doing well. But this doesn't mean that IXPs don't offer an important level of service to their constituents. Because of them, the isp industry has convenient access to dense interconnection at a pretty decent price. yes, that's the progression of success. and my way of designing for success is to start people off with VNI's (two-port VLANs containing one peering) so that when they move from shared-access to dedicated they're just moving from a virtual wire to a physical wire without losing any of the side-benefits they may have got from a shared-access peering fabric. Q in Q is not how i'd build this... cisco and juniper both have hardware tunnelling capabilities that support this stuff... it just means as the IXP fabric grows it has to become router-based. Hey, I have an idea: you could take this plan and build a tunnel-based or even a native IP access IXP platform like this, extend it to multiple locations and then buy transit from a bunch of companies which would give you a native L3 based IXP with either client prefixes only or else an option for full DFZ connectivity over the exchange fabric. You could even build a global IXP on this basis! It's a brilliant idea, and I just can't imagine why no-one thought of it before. :-). i've been known to extend IXP fabrics to cover a metro, but never beyond.
Re: IXP
Date: Sat, 18 Apr 2009 13:17:11 -0400 From: Steven M. Bellovin s...@cs.columbia.edu On Sat, 18 Apr 2009 16:58:24 + bmann...@vacation.karoshi.com wrote: i make the claim that simple, clean design and execution is best. even the security goofs will agree. Even? *Especially* -- or they're not competent at doing security. wouldn't a security person also know about http://en.wikipedia.org/wiki/ARP_spoofing and know that many colo facilities now use one customer per vlan due to this concern? (i remember florian weimer being surprised that we didn't have such a policy on the ISC guest network.) if we maximize for simplicity we get a DELNI. oops that's not fast enough we need a switch not a hub and it has to go 10Gbit/sec/port. looks like we traded away some simplicity in order to reach our goals.
Re: IXP
Large IXP have 300 customers. You would need up to 45k vlan tags, wouldn't you? the 300-peer IXP's i've been associated with weren't quite full mesh in terms of who actually wanted to peer with whom, so, no.
www.vix.com/personalcolo (Re: US west coast personal colo)
i just restored http://www.vix.com/personalcolo/ from backup. last update 2007. i guess this calls for another round of send me your updates, folks. re: Sean Donelan s...@donelan.com writes: Is anyone still doing personal colo on the west coast? I'm looking for a new home for my personal server on the west coast, and it seems like the economy has taken out most of the old personal colo offers. Even the old web page on www.vix.com/personalcolo is gone. -- Paul Vixie
Re: [OT] Re: Fiber cut in SF area
Christopher Morrow morrowc.li...@gmail.com writes: and I also would ask.. what's the cost/risk here? 'We' lost at best ~1day for some folks in the outage, nothing global and nothing earth-shattering... This has happened (this sort of thing) 1 time in how many years? Expending $$ and time and people to go 'put padlocks on manhole covers' seems like spending in the wrong place... as long as the west's ideological opponents want terror rather than panic, and also to inflict long term losses rather than short term losses, that's true. in this light you can hopefully understand why bollards to protect internet exchanges against truck bombs are not only penny wise pound foolish (since the manholes a half mile away won't be hardened or monitored or even locked) but also completely wrongheaded (since terrorists need publicity which means they need their victims to be fully able to communicate.) -- Paul Vixie
Re: ISC DLV
Paul Ferguson fergdawgs...@gmail.com writes: On Sat, Apr 4, 2009 at 9:55 PM, Marcelo Gardini do Amaral mgard...@gmail.com wrote: Guys, are you having problems to validate DNSEC using ISC DLV? No idea, but I did see another reference to this over on the OARC dns-ops list: https://lists.dns-oarc.net/pipermail/dns-operations/2009-April/003726.html note, this isn't a ddos, so it's probably not related to the other dns ddos events that have been discussed here recently. see also geoff's reply on that thread: Date: Sat, 04 Apr 2009 23:15:55 -0700 From: Geoffrey Sisson ge...@geoff.co.uk To: dns-operati...@lists.dns-oarc.net Subject: Re: [dns-operations] ISC DLV broken? Sender: dns-operations-boun...@lists.dns-oarc.net m...@ucla.edu (Michael Van Norman) wrote: Starting a bit after 18:00, my home machines starting failing DNSSEC validation using the ISC DLV. ... Are other people seeing this? Yes, starting at around the same time (PDT). peter_los...@isc.org (Peter Losher) wrote: ISC is aware that there is a issue with lookups against dlv.isc.org and are investigating the cause behind it. You may want to disable DNSSEC validation against dlv.isc.org at this time. It appears as if the RRSIG RRset returned by the DLV nameservers for dlv.isc.org is missing the RRSIG for the KSK, so validation for dlv.isc.org is failing. It _does_ contain the RRSIG for the ZSK (key id 64263). As a test I tried changing the trusted key to the ZSK, and DLV validation appeared to work correctly. This is, of course, not a recommended work-around. Geoff ___ dns-operations mailing list dns-operati...@lists.dns-oarc.net https://lists.dns-oarc.net/mailman/listinfo/dns-operations
Re: ISC DLV
David Conrad d...@virtualized.org writes: ... I'm sure the folks at ISC will attempt to minimize reoccurrence. yes. though with two outages in the last month, some early DLV adopters might be getting a bit nervous. as with DNSSEC itself when folks first started turning it on a few years ago, the failure codepaths for DLV are inevitably not as well oiled as the success codepaths. (we're on it.) -- Paul Vixie
Re: Global Blackhole Service
where you lose me is where the attacker must always win. Do you have a miraculous way to stop DDOS? Is there now a way to quickly and efficiently track down forged packets? Is there a remedy to shutting down the *known* botnets, not to mention the unknown ones? there are no silver bullets. anyone who says otherwise is selling something. The attacker will always win if he has a large enough attack platform/... While all this is worked out, we have one solution we know works. we had to destroy the village in order to save it. If we null route the victim IP, the traffic stops at the null route. Since most attackers don't care to DOS the ISP, but just to take care of that end point, they usually don't start shifting targets to try and keep the ISP itself out. if you null route the victim IP, the victim is off the air, so the DDoS is a success even though it mostly does not reach its target. you're proposing that we lower an attacker's costs. in a war of economics that's bad juju, and all wars are about economics. there are no silver bullets. isp's who permit random source addresses on packets leaving their networks are creating a global hazard, and since they are defending their practices on the basis of thin profit margins it's right to call this the chemical polluter business model. as long as the rest of us continue to peer with these chemical polluters, then anyone on the internet can be the victim of a devastating DDoS at any time and at low cost. that's not a silver bullet however. if most ISP's controlled their source addresses there would still be DDoS's and then the new problem would be lack of real-time cooperation along the lines of hi i'm in the XYZ NOC and we're tracking a DDoS against one of our customers and 14% of it is coming from your address space, here's the summary of timestamp-ip-volume and here's a pointer to your share of the netflows, can you remediate? the answer will start out just like today's BCP38 answer, no we can't afford the staff or technology to do that, and then lawyers would worry about liability, and we'd all have to worry about monopolies, censorship, social engineering, and so on. in all of these cases the problem is the margins themselves. just as the full cost of a fast food cheeseburger is probably about $20 if you count all the costs that the corporations are shifting onto society, so it is that the full cost of a 3MBit/sec DSL line is probably $300/month if you count all the costs that ISPs shift onto digital society. the usual argument goes (and i'm just putting it out here to save time, though i'm betting several respondants will not read closely and so will just spew this out as though it's their original idea and as though i had not dismissed it many times over the decades): we cannot build a digital economy without cost shifting since noone would pay what it really costs during the rampup. i don't dignify that with a reply, either here in effigy, or if anyone happens to trot it out again.
Re: Global Blackhole Service
a minor editorial comment: Jens Ott - PlusServer AG j@plusserver.de writes: Jack Bates schrieb: Paul Vixie wrote: Do you have a miraculous way to stop DDOS? Is there now a way to quickly and efficiently track down forged packets? Is there a remedy to shutting down the *known* botnets, not to mention the unknown ones? the quoted text was written by jack bates, not paul vixie. -- Paul Vixie
Re: Global Blackhole Service
blackholing victims is an interesting economics proposition. you're saying the attacker must always win but that they must not be allowed to affect the infrastructure. and you're saying victims will request this, since they know they can't withstand the attack and don't want to be held responsible for damage to the infrastructure. where you lose me is where the attacker must always win.
Re: v6 DSL / Cable modems
Ricky Beam jfb...@gmail.com writes: ... In the mid-80's, /8's were handed out like candy because there were lots of address space and we'll never use it all. ... ahem. allow me to introduce myself. i was alive and actively using the internet in the mid-80's, and when we got our /8 it was justified very differently than what you said. we had field offices in 100 countries and we had 130,000 employees and our internal network spanned five continents. (we thought long and hard about netmasks before we started rolling it out.) it was not true in Digital Equipment Corporation's (DEC's) case that a /8 was handed out like candy or that the justification was anything like lots of address space or we'll never use it all. IPv6 was designed to not need DHCP. DHCPv6 has come about since people need more than just an address from autoconfiguration. IPv6 promised a lot of things, like no-forklift insertion of IPv6 into the existing IPv4 network, and some hosts, such as printers, might need never be upgraded. a lot of those promises were trash, just stuff that folks had to say to get through whatever they were getting through. as much as i'd like a time machine to go back and whisper yo, dude, that's *so* not gonna happen in some ears, what matters to us now is not what IPv6 was promised to be or even what it could have been but instead: what it could now become. I can recall many posts over the years from the IPng WG telling people they didn't need DHCP. some people drink their own cool-aid. advice: get better at ignoring them. i dislike the compromises and mistakes other people will make when faced with NAT, and i don't want to live in a world dominated by products and services containing those compromises or those mistakes. i want end-to-end so i can stop budgeting half a day for each VoIP phone i send home with an employee. i don't want to remap addresses mid-path because i just know that the best programmers are the lazy ones and they WILL encode endpoint IP addrs in their sessions no matter what we tell them or how much it hurts us all. IPv6 coulda been and shoulda been lots of better things than we're getting, but due to circumstances beyond our present control, it's what we've got to work with, and it could still avoid a lot of problems whose alternative costs could be higher (NAT, double NAT, triple NAT, IPv4 markets, IPv4 black markets, IPv4 route piracy, explosive deaggregation, to name some). the most fundamental re-think required to wrap a brain around IPv6 compared to IPv4 is that we will never run out of addresses again unless someone (ignorantly) assigns a /125 to a LAN and needs more than 7 hosts thereon, or something similar. that part of IPv4's dark past will not follow us to IPv6 and we can stop thinking all related or derivative thoughts, for IPv6. but, and this matters so please pay attention, IPv6 does nothing to solve the routing table problem that IPv4 has had since 1995 or so, and IPv6 can amplify this part of IPv4's dark past and make it much worse since there can be so many more attached devices. the fundamental implication is, forget about address space, it's paperwork now, it's off the table as a negotiating item or any kind of constraint. but the size of the routing table is still a bogeyman, and IPv6 arms that bogeyman with nukes. -- Paul Vixie
Re: DNS Amplification attack?
Mark Andrews mark_andr...@isc.org writes: Authoritative servers need a cache. Authoritative servers need to ask queries. The DNS protocol has evolved since RFC 1034 and RFC 1035 and authoritative servers need to translate named to addresses for their own use. See RFC 1996, A Mechanism for Prompt Notification of Zone Changes (DNS NOTIFY). if i had RFC 1996 to do over again i would either limit outbound notifies to in-zone servernames, or recommend that primary server operators configure stealth slaves for servername-containing zones, or (most likely) i would point out that the need to look up secondary servernames requires that an authority-only nameserver be able to act as a stub resolver and that such a server much have access to an independent recursive nameserver. it's not too late to implement it that way. no authority-only server should need a cache of any kind. the above text from marka represents a BIND implementatin detail, not a protocol requirement, evolved or not. The real fix is to get BCP 38 deployed. Reflection amplification attacks can be effective if BCP 38 measures have not been deployed. Go chase down the offending sources. BCP 38 is nearly 10 years old. my agreement with this statement is tempered by the fact that BCP38 deployment cannot be continuously assured, nor tested. therefore we will need protocols, implementations, and operational practices that take account of packet source address spoofing as an unduring property of the internet. We all should be taking this as a opportunity to find where the leaks are in the BCP 38 deployment and correct them. Mark yea, verily. and maybe track down rfc1918-sourced spew while you're at it. -- Paul Vixie
DNSSEC vs. X509 (Re: Security team successfully cracks SSL...)
Joe Abley jab...@hopcount.ca writes: On 2009-01-05, at 15:18, Jason Uhlenkott wrote: If we had DNSSEC, we could do away with SSL CAs entirely. The owner of each domain or host could publish a self-signed cert in a TXT RR, ... or even in a CERT RR, as I heard various clever people talking about in some virtual hallway the other day. http://www.isi.edu/in-notes/rfc2538.txt. i wasn't clever but i was in that hallway. it's more complicated than RFC 2538, but there does seem to be a way forward involving SSL/TLS (to get channel encryption) but where a self-signed key could be verified using a CERT RR (to get endpoint identity authentication). the attacks recently have been against MD5 (used by some X.509 CA's) and against an X.509 CA's identity verification methods (used at certificate granting time). no recent attack has shaken my confidence in SSL/TLS negotiation or encryption, but frankly i'm a little worried about nondeployability of X.509 now that i see what the CA's are doing operationally when they start to feel margin pressure and need to keep volume up + costs down. i don't have a specific proposal. (yet.) but i'm investigating, and i recommend others do likewise. -- Paul Vixie
Re: Sprint v. Cogent, some clarity facts
(note: i don't think sprint or cogent is being evil in this situation.) [EMAIL PROTECTED] writes: Has anyone heard of a backup route? With a longer path so it is never used unless there is a real emergency? Why was there no backup route available to carry the Sprint - Cogent traffic? Because there was a political failure in both Sprint and Cogent. what you're calling a political failure could be what others call a rate war. i'd imagine that cogent's cost structure lower than most networks' (since their published prices are so low and since they're not pumping new money in from the outside and since they have no non-IP business they could be using to support dumping). if cogent's rates aare also lowest then the other large networks might be losing customers toward cogent and those other large networks might feel they are hurting their own cause by peering, settlement free, with this new-economy competitor. if that's the case then the political failure you describe might be a matter of cogent saying we don't want our prices to our customers to reflect the capital inefficiencies of other networks and those other networks saying we do. note, i'm not on the inside and i'm working only from public knowledge here and so this is all supposition. but calling it political failure when these possibilities exist seems like a stretch. there'd be no other leverage whereby cogent could protect the price point its customers seem to like so well. i'm not saying a cogent customer will be glad to trade some instability to get those prices, but i am saying that if this long chain of guesses is accurate it likely also represents the ONLY way to drive efficiency in a competitive capital-intensive market. Back in 2000 it was acceptable for the big New York banks to have all their eggs in one basket in central Manhattan. In 2002, it was no longer acceptable. Do we really need a 911 magnitude of disaster on the Internet for people to wake up and smell the coffee? The Internet is no longer a kewl tool built and operated by the cognoscenti to meet their own interests. It is now part of every nation's and everbody's critical infrastructure. It needs to be engineered and operated better so that it does not end up partitioning for dumb reasons. that sounds like justification for government regulation, if true. -- Paul Vixie
Re: Sprint / Cogent dispute over?
Daniel Senie [EMAIL PROTECTED] writes: At 06:54 PM 11/2/2008, Daniel Roesen wrote: https://www.sprint.net/cogent.php ... Also in this document is a complaint that Cogent failed to disconnect. Excuse me? This was a trial PEERING agreement. That implies one or a series of point-to-point connections. That implies EITHER party can disconnect the circuits (in reality, the physical circuit doesn't even matter, just shut down the BGP session(s)). ... Not having read the contract in question, my assumption when I read Sprint's account of their depeering of Cogent was that the trial peering contract says Sprint will notify Cogent of its qualification status after 90 days; if in Sprint's estimation Cogent does not qualify, and Sprint notifies Cogent of that fact, then Cogent will either disconnect or start paying. Sprint's document's wording is careful even if their TITLE is not. If they are involved in litigation with Cogent then actual lawyers would have seen that text (if not necessarily the TITLE) before it went out. The heart of the lawsuit might be whether Cogent did or didn't implicitly agree to pay, as signalled by their lack of disconnection after their 90 day notice. None of us who aren't parties to the dispute can do other than wonder, ponder, guess. -- Paul Vixie