Re: What is multihoming was (design of a real routing v. endpoint id seperation)
On Mon, 24 Oct 2005, Owen DeLong wrote: Yes... The network is still multihomed, but, instead of using routing to handle the source/dest addr. selection, it is managed at each end host independent of the routers. The routers function sort of like the network is single homed. It's very convoluted. That is to say the least. Offices who want to be multihomed would want to do it once for all the computers there using one device like they now can do with a router. Web farms would similarly want to do it for all the servers there again as they now do with one router or load-balancer, etc. Managing it if multihoming is entirely host-based would be hard (I note that for office multihoming you could potentially create one router that would do shim6 on its out interfaces and would do NAT between that and its inside network - but we don't want NAT for ipv6 if I understand IETF and IAB direction). So while I really do think that we need some-kind of multi6 design which works for small multi-homing networks without need for them to have to use ASN and have their routes in global BGP table (leaving all that primarily to NSPs with /32 and larger as IETF envisioned), the current shim6 design does not seem properly done to be usable for that audience and as somebody noticed yesterday it would instead be great for multi-dsl users, especially gamers and p2p. Now if we resurrected A6 with its ability to separately enter ip address with host and network parts at the dns level - then we're at least part the way done as far as multi6 multi-homing setup in dns for entire network at once. But I still don't see easy way to do it for the device management and yet another new protocol would probably be needed for automatic assignment of locators and secondary ipv6 addresses (BTW - did I hear right that there is going to be new WG related to MIP6 to work out issues of assignment and using of multiple ipv6 addresses and interfaces - IETF seems to be doing lots of things in parallel at this potential L3.5 layer that could be done lot better together as part of proper TCP/IP redesign). --- William Leibzon Elan Networks [EMAIL PROTECTED]
Re: What is multihoming was (design of a real routing v. endpoint id seperation)
On Mon, 24 Oct 2005 [EMAIL PROTECTED] wrote: A single tier-2 ISP who uses BGP multihoming with several tier 1 ISPs can provide multihoming to it's customers without BGP. For instance, if this tier-2 has two PoPs in a city and peering links exist at both PoPs and they sell a resilient access service where the customer has two links, one to each PoP, then it is possible to route around many failures. This is probably sufficient for most people and if the tier-2 provider takes this service seriously they can engineer things to make total network collapse exteremely unlikely. From RFC 3582, this is not multihoming (see the defs below). The above is referred to as multi-connecting or multi-attaching (also see RFC 4116). I agree, this is sufficient for many sites. Especially in academic world, many universities are just multi-connected, trusting the stability of their NREN's backbone and transit providers. Lots of commercial sites do it too, but some are wary due to events like L3/Cogent, L3 backbone downtime, etc. . A multihomed site is one with more than one transit provider. Site-multihoming is the practice of arranging a site to be multihomed. and: A transit provider operates a site that directly provides connectivity to the Internet to one or more external sites. The connectivity provided extends beyond the transit provider's own site. A transit provider's site is directly connected to the sites for which it provides transit. -- Pekka Savola You each name yourselves king, yet the Netcore Oykingdom bleeds. Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
Re: ICANN and Verisign settle over SiteFinder
On Tue, 25 Oct 2005, Florian Weimer wrote: http://www.businessweek.com/ap/financialnews/D8DEL2TO7.htm? campaign_id=apn_tech_downchan=tc I don't understand what VeriSign receives in return for their kowtow (under the agreement, they basically waive any right to criticize ICANN's role). They get to continue to be .COM registry forever as new agreement would extend to 2012 and then automatically extended further without formal process as it happened recently for .NET. They also are going to be able to increase registry fees for .COM by 7% per year which to put it in perspective can potentially be $2 increase 4 years from now. Two possible explanations: 2+2=5, right? :) * ICANN signalled a positive outcome of a future Sitefinder review under the new process. * ICANN promised to grant VeriSign the DNSSEC root and .ARPA maintenance without tender (the Root Server Management Transition Agreement goes into that direction; actually, the .ARPA stuff is the interesting one). * VeriSign has recognized that they couldn't win in court, and suddenly want to play nice. -- William Leibzon Elan Networks [EMAIL PROTECTED]
Re: What is multihoming was (design of a real routing v. endpoint id seperation)
From [EMAIL PROTECTED] Mon Oct 24 15:33:02 2005 Date: Mon, 24 Oct 2005 13:31:17 -0700 Subject: Re: What is multihoming was (design of a real routing v. endpoint id seperation) Stephen Sprunk wrote: [snip] Other people use this term in very different ways. To some people it means using having multiple IP addresses bound to a single network interface. To others it means multiple websites on one server. That is virtual hosting in a NANOG context. Some undereducated MCSEs might call it multihoming, but let's not endorse that here. Unfortunately, this is a common and standards blessed way to refer to any host with multiple interfaces/addresses (real or virtual). For example, from the Terminology section, 1.1.3, of RFC1122, Requirements for Internet Hosts -- Communication Layers, says, Multihomed A host is said to be multihomed if it has multiple IP addresses. For a discussion of multihoming, see Section 3.3.4 below. *sigh* Multi-homing simply means 'having external connections to more than one network' -- be it a network with multiple, disjoint, ingress/egress paths, or a host with interfaces (real or virtual) on distinct LAN subnets (even if those subnets are agregated into a single net somewhere upstream. A host with multiple adresses utilizing the _same_ netblock/netmask _should_ _not_ be called multi-homed (because there is only one path to that host), it is simply a single-homed host with multiple identities. might be called poly-ip-any or some such. grin
Re: What is multihoming was (design of a real routing v. endpoint id seperation)
On 25-Oct-2005, at 05:56, Robert Bonomi wrote: *sigh* Multi-homing simply means [...] As became clear when we wrote the draft that became RFC 3582, apparently simple terms such as transit provider and multi-homing mean surprisingly different things to different people. The important thing is not who is right, and not which definition is the best, but that everybody uses the same definitions so that they can talk to each other without running around in circles. Joe
Re: ICANN and Verisign settle over SiteFinder
* william elan net: They get to continue to be .COM registry forever as new agreement would extend to 2012 and then automatically extended further without formal process as it happened recently for .NET. They also are going to be able to increase registry fees for .COM by 7% per year which to put it in perspective can potentially be $2 increase 4 years from now. So the deal makes indeed sense from a business perspective. Thanks. Two possible explanations: 2+2=5, right? :) Oops. 8-)
Re: Scalability issues in the Internet routing system
On Mon, 24 Oct 2005, Blaine Christian wrote: As of the last time that I looked at it (admittedly quite awhile ago), something like 80% of the forwarding table had at least one hit per minute. This may well have changed given the number of traffic engineering prefixes that are circulating. Tony Yea, but that's just me pinging everything and google and yahoo fighting over who has the most complete list of x rated sites. and this probably depends greatly on the network, user-population, business involved. Is it even a metric worth tracking?
Re: Scalability issues in the Internet routing system
On Tue, 25 Oct 2005 16:28:05 -, Christopher L. Morrow said: On Mon, 24 Oct 2005, Blaine Christian wrote: Yea, but that's just me pinging everything and google and yahoo fighting over who has the most complete list of x rated sites. and this probably depends greatly on the network, user-population, business involved. Is it even a metric worth tracking? It's a fight for eyeballs, isn't it? Routing table hits caused by spidering from search engines will give a good indication of what percent of the address space the spiders are covering. Of course, you need views from a number of places, and some adjusting for the fact that the webservers are usually clumped in very small pockets of address space. On the other hand, if it can be established that 80% of the routing table is hit every N minutes, which would tend to argue against caching a very small subset, but the vast majority of the routing table hits are just spiders, that may mean that a cache miss isn't as important as we thought... Anybody got actual measured numbers on how much of the hits are just spiders and Microsoft malware scanning for vulnerable hosts? pgpuZ7kJs8OEa.pgp Description: PGP signature
IRR Coordination mailing list
This IRR Coordination mailing list was mentioned this morning during the BGP Filtering talk.We'd like to invite anyone interested in improving the trust, consistency, and coordination of IRR's to join. The archive and subscription details can be found at http://www.merit.edu/mail.archives/irrc/ -Larry Blunk Merit
Re: Scalability issues in the Internet routing system
On Tue, 25 Oct 2005 [EMAIL PROTECTED] wrote: On Tue, 25 Oct 2005 16:28:05 -, Christopher L. Morrow said: On Mon, 24 Oct 2005, Blaine Christian wrote: Yea, but that's just me pinging everything and google and yahoo fighting over who has the most complete list of x rated sites. and this probably depends greatly on the network, user-population, business involved. Is it even a metric worth tracking? It's a fight for eyeballs, isn't it? Routing table hits caused by spidering from search engines will give a good indication of what percent of the oops, I should have not replied to Blaine/Tony but directly to tony's message :( The real question was: If the percentage of hits is dependent on 'user population', 'business', 'network' is it even worth metricing for the purpose of design of the device/protocol? Unless of course you want a 'sport' and 'offroad' switch on your router :)
Re: What is multihoming was (design of a real routing v. endpoint id seperation)
Robert Bonomi wrote: From [EMAIL PROTECTED] Mon Oct 24 15:33:02 2005 Date: Mon, 24 Oct 2005 13:31:17 -0700 Subject: Re: What is multihoming was (design of a real routing v. endpoint id seperation) Stephen Sprunk wrote: [snip] Other people use this term in very different ways. To some people it means using having multiple IP addresses bound to a single network interface. To others it means multiple websites on one server. That is virtual hosting in a NANOG context. Some undereducated MCSEs might call it multihoming, but let's not endorse that here. Unfortunately, this is a common and standards blessed way to refer to any host with multiple interfaces/addresses (real or virtual). For example, from the Terminology section, 1.1.3, of RFC1122, Requirements for Internet Hosts -- Communication Layers, says, Multihomed A host is said to be multihomed if it has multiple IP addresses. For a discussion of multihoming, see Section 3.3.4 below. *sigh* Multi-homing simply means 'having external connections to more than one network' -- be it a network with multiple, disjoint, ingress/egress paths, or a host with interfaces (real or virtual) on distinct LAN subnets (even if those subnets are agregated into a single net somewhere upstream. A host with multiple adresses utilizing the _same_ netblock/netmask _should_ _not_ be called multi-homed (because there is only one path to that host), it is simply a single-homed host with multiple identities. might be called poly-ip-any or some such. grin Depends who you ask. Again, RFC1122 says (section 1.1.1), A host is generally said to be multihomed if it has more than one interface to the same or to different networks. And also section 3.3.4.1, A multihomed host has multiple IP addresses, which we may think of as logical interfaces. These logical interfaces may be associated with one or more physical interfaces, and these physical interfaces may be connected to the same or different networks. As far as a multihomed host is concerned, RFC1122 sure seems to call anything with multiple IPs multihomed. Multihomed is a trait of the host independent of any network topology around the host. But whatever. It just means people need to be clear what they are talking about when they say multihomed. As is clear from this thread, there is not clear agreement on what the precise meaning is. -- Crist J. Clark [EMAIL PROTECTED] Globalstar Communications(408) 933-4387
Re: ICANN and Verisign settle over SiteFinder
On Tue, 25 Oct 2005, Florian Weimer wrote: Two possible explanations: 2+2=5, right? :) Oops. 8-) tongue location=cheek No, you got it right. The [third] option at the end, play nice, has only a passing association to the realm of possibility. /tongue -- -- Todd Vierling [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED]
Re: ICANN and Verisign settle over SiteFinder
I don't understand what VeriSign receives in return for their kowtow (under the agreement, they basically waive any right to criticize ICANN's role). As someone else noted, a perpetual cash cow in .COM with 7%/year escalator clause. * ICANN signalled a positive outcome of a future Sitefinder review under the new process. Nope, there's this complex process with outside experts to review any new proposed sitefinder like thing. * ICANN promised to grant VeriSign the DNSSEC root and .ARPA maintenance without tender (the Root Server Management Transition Agreement goes into that direction; actually, the .ARPA stuff is the interesting one). My reading is the opposite, ICANN will create the root zone now. * VeriSign has recognized that they couldn't win in court, and suddenly want to play nice. Quite possibly and don't be silly. More concretely, they probably decided they were unlikely to win more than this agreement gives them. R's, John
fleet.navy.mil DNS / network ops contact please
Could a dns ops contact for any of these hosts please email me offlist to help troubleshoot a network reachablity / dns lookup issue from our servers. thanks -srs ccsg3.navy.mil. 30M IN NS ns2.fleet.navy.mil. ccsg3.navy.mil. 30M IN NS dnsmail.uar.navy.mil. ccsg3.navy.mil. 30M IN NS ns1.fleet.navy.mil. -- Suresh Ramasubramanian ([EMAIL PROTECTED])
Re: Scalability issues in the Internet routing system
Assume you have determined that a percentage (20%, 80%, whatever) of the routing table is really used for a fixed time period. If you design a forwarding system that can do some packets per second for those most used routes, all you need to DDoS it is a zombie network that would send packets to all other destinations... rate-limiting and dampening would probably come into place, and a new arms race would start, killing operator's abilities to fast renumber sites or entire networks and new troubleshooting issues for network operators. Isn't just simpler to forward at line-rate ? IP look ups are fast nowadays, due to algorithmic and architecture improvements... even packet classification (which is n-tuple version of the IP look up problem) is not that hard anymore. Algorithms can be updated on software-based routers, and performance gains far exceed Moore's Law and projected prefix growth rates... and routers that cannot cope with that can always be changed to handle IGP-only routes and default gateway to a router that can keep up with full routing. (actually, hardware-based routers based on limited size CAMs are more vulnerable to obsolescence by routing table growth than software ones) Let's celebrate the death of ip route-cache, not hellraise this fragility. Rubens On 10/24/05, Alexei Roudnev [EMAIL PROTECTED] wrote: One question - which percent of routing table of any particular router is REALLY used, say, during 1 week? I have a strong impression, that answer wil not be more than 20% even in biggerst backbones, and will be (more likely) below 1% in the rest of the world. Which makes a hige space for optimization. - Original Message - From: Daniel Senie [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, October 18, 2005 9:50 AM Subject: Re: Scalability issues in the Internet routing system At 11:30 AM 10/18/2005, Andre Oppermann wrote: I guess it's time to have a look at the actual scalability issues we face in the Internet routing system. Maybe the area of action becomes a bit more clear with such an assessment. In the current Internet routing system we face two distinctive scalability issues: 1. The number of prefixes*paths in the routing table and interdomain routing system (BGP) This problem scales with the number of prefixes and available paths to a particlar router/network in addition to constant churn in the reachablility state. The required capacity for a routers control plane is: capacity = prefix * path * churnfactor / second I think it is safe, even with projected AS and IP uptake, to assume Moore's law can cope with this. Moore will keep up reasonably with both the CPU needed to keep BGP perking, and with memory requirements for the RIB, as well as other non-data-path functions of routers. 2. The number of longest match prefixes in the forwarding table This problem scales with the number of prefixes and the number of packets per second the router has to process under full or expected load. The required capacity for a routers forwarding plane is: capacity = prefixes * packets / second This one is much harder to cope with as the number of prefixes and the link speeds are rising. Thus the problem is multiplicative to quadratic. Here I think Moore's law doesn't cope with the increase in projected growth in longest prefix match prefixes and link speed. Doing longest prefix matches in hardware is relatively complex. Even more so for the additional bits in IPv6. Doing perfect matches in hardware is much easier though... Several items regarding FIB lookup: 1) The design of the FIB need not be the same as the RIB. There is plenty of room for creativity in router design in this space. Specifically, the FIB could be dramatically reduced in size via aggregation. The number of egress points (real or virtual) and/or policies within a router is likely FAR smaller than the total number of routes. It's unclear if any significant effort has been put into this. 2) Nothing says the design of the FIB lookup hardware has to be longest match. Other designs are quite possible. Again, some creativity in design could go a long way. The end result must match that which would be provided by longest-match lookup, but that doesn't mean the ASIC/FPGA or general purpose CPUs on the line card actually have to implement the mechanism in that fashion. 3) Don't discount novel uses of commodity components. There are fast CPU chips available today that may be appropriate to embed on line cards with a bit of firmware, and may be a lot more cost effective and sufficiently fast compared to custom ASICs of a few years ago. The definition of what's hardware and what's software on line cards need not be entirely defined by whether the design is executed entirely by a hardware engineer or a software engineer. Finally, don't
Re: Scalability issues in the Internet routing system
Vice versa. DDOS attack will never work by this way, because this router will (de facto) prioritize long established streams vs. new and random ones, so it will not notice DDOS attack at all - just some DDOS packets will be delayed or lost. You do not need to forward 100% packets on line card rate; forwarding 95% packets on card rate and have other processing (with possible delays) thru central CPU can work good enough. It is all about tricks and optimizations - fast routing is not state of art and can be optimized by many ways. For now, it was not necessary; when it became necessary - it will be done in 1/2 year. - Original Message - From: Rubens Kuhl Jr. [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, October 25, 2005 9:21 PM Subject: Re: Scalability issues in the Internet routing system Assume you have determined that a percentage (20%, 80%, whatever) of the routing table is really used for a fixed time period. If you design a forwarding system that can do some packets per second for those most used routes, all you need to DDoS it is a zombie network that would send packets to all other destinations... rate-limiting and dampening would probably come into place, and a new arms race would start, killing operator's abilities to fast renumber sites or entire networks and new troubleshooting issues for network operators. Isn't just simpler to forward at line-rate ? IP look ups are fast nowadays, due to algorithmic and architecture improvements... even packet classification (which is n-tuple version of the IP look up problem) is not that hard anymore. Algorithms can be updated on software-based routers, and performance gains far exceed Moore's Law and projected prefix growth rates... and routers that cannot cope with that can always be changed to handle IGP-only routes and default gateway to a router that can keep up with full routing. (actually, hardware-based routers based on limited size CAMs are more vulnerable to obsolescence by routing table growth than software ones) Let's celebrate the death of ip route-cache, not hellraise this fragility. Rubens On 10/24/05, Alexei Roudnev [EMAIL PROTECTED] wrote: One question - which percent of routing table of any particular router is REALLY used, say, during 1 week? I have a strong impression, that answer wil not be more than 20% even in biggerst backbones, and will be (more likely) below 1% in the rest of the world. Which makes a hige space for optimization. - Original Message - From: Daniel Senie [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, October 18, 2005 9:50 AM Subject: Re: Scalability issues in the Internet routing system At 11:30 AM 10/18/2005, Andre Oppermann wrote: I guess it's time to have a look at the actual scalability issues we face in the Internet routing system. Maybe the area of action becomes a bit more clear with such an assessment. In the current Internet routing system we face two distinctive scalability issues: 1. The number of prefixes*paths in the routing table and interdomain routing system (BGP) This problem scales with the number of prefixes and available paths to a particlar router/network in addition to constant churn in the reachablility state. The required capacity for a routers control plane is: capacity = prefix * path * churnfactor / second I think it is safe, even with projected AS and IP uptake, to assume Moore's law can cope with this. Moore will keep up reasonably with both the CPU needed to keep BGP perking, and with memory requirements for the RIB, as well as other non-data-path functions of routers. 2. The number of longest match prefixes in the forwarding table This problem scales with the number of prefixes and the number of packets per second the router has to process under full or expected load. The required capacity for a routers forwarding plane is: capacity = prefixes * packets / second This one is much harder to cope with as the number of prefixes and the link speeds are rising. Thus the problem is multiplicative to quadratic. Here I think Moore's law doesn't cope with the increase in projected growth in longest prefix match prefixes and link speed. Doing longest prefix matches in hardware is relatively complex. Even more so for the additional bits in IPv6. Doing perfect matches in hardware is much easier though... Several items regarding FIB lookup: 1) The design of the FIB need not be the same as the RIB. There is plenty of room for creativity in router design in this space. Specifically, the FIB could be dramatically reduced in size via aggregation. The number of egress points (real or virtual) and/or policies within a router is likely FAR smaller than the total number of routes. It's unclear if any significant effort has been put into this. 2) Nothing says the design of the FIB lookup hardware has to be