Re: Best practice for BGP session/ full routes for customer
On Monday, July 14, 2014 07:32:43 PM Jeff Tantsura wrote: > Mark, > > BGP to RIB filtering (in any vendor implementation) is > targeting RR which is not in the forwarding path, so > there¹s no forwarding towards any destination filtered > out from RIB. > Using it selectively on a forwarding node is error prone > and in case of incorrect configuration would result in > blackholing. As with every feature on a router, you need to know what you're doing to make it work. Don't blame the cows if you turn on knobs you have no business using, or don't care to learn the risks of. We use this feature in our network successfully, because we know what we're doing, and care to understand the risks. If I use it in a manner other than previously directed (while I know it's a use-case, I've never heard of any vendor saying it ONLY targeted out-of-path route reflectors, but then again, I don't generally walk vendor corridors for the scoop), well, welcome to the Internet; where core routers can either be behemoths that move air the size of a football field and could be mistaken for seismic detection machines, or last generation's x86 home desktop running Quagga and grandma's health app :-). Mark. signature.asc Description: This is a digitally signed message part.
Re: Net Neutrality...
On 7/16/2014 1:57 PM, Doug Barton wrote: On 07/16/2014 08:45 AM, Eric Brunner-Williams wrote: On 7/16/14 7:50 AM, Fred Baker (fred) wrote: Relevant article by former FCC Chair http://www.washingtonpost.com/posteverything/wp/2014/07/14/this-is-why-the-government-should-never-control-the-internet/ It reads like a hit piece (by a Republican "free markets" ideologue) on a (Progressive) Democratic primary candidate for Lt. Governor of New York, not like a reasoned case by an informed policy analyst. Errr, I didn't see anything about any LTG candidates in that piece, what did I miss? I'm also curious about what it is that you think is misstated or overblown in that piece that would lead you to believe that it's a "hit piece." His use of phrases like "by a Republican "free markets" ideologue" pretty much nailed the value of his remarks for me. -- Requiescas in pace o email Two identifying characteristics of System Administrators: Ex turpi causa non oritur actio Infallibility, and the ability to learn from their mistakes. (Adapted from Stephen Pinker)
Re: Net Neutrality...
On 7/17/14, 2:15 PM, valdis.kletni...@vt.edu wrote: /me makes popcorn and waits for 4K displays to drop under US$1K and watch the network providers completely lose their shit http://www.amazon.com/Seiki-SE39UY04-39-Inch-Ultra-120Hz/dp/B00DOPGO2G $339! I use it for doing dev. It's *fabulous*. Mike
Re: Net Neutrality...
On Tue, 15 Jul 2014 13:08:58 -0600, Brett Glass said: > Estimates of the maximum bandwidths of all the human senses, combined, > range between the capacity of a T1 line (at the low end) and > about 4 Mbps (at the high end). A human being simply is not wired to > accept more input. (Yes, machines could digest more... which means that > additional bandwidth to and from the home might be useful for the purpose > of spying on us.) What does this imply about the FCC's proposal to > redefine "broadband" as a symmetrical 10 Mbps? Actually, vision is higher bandwidth than that - most VR people estimate that approaching human vision requires a gigapixel/second (at 24 bits or more per pixel) - and even that needs to play lots of eye-tracking games to concentrate the rendering on where the eye is focused. Consider how fast even high-end NVidia cards can pump out pixels and you can *still* see it's CGI. Well-shot 4K video of real objects displayed on a good monitor is *just* reaching the "it actually looks real" level - and that's a hell of a lot more than 4Mbps. And remember that bits are consumed by more than just one human per dwelling - you can have multiple people watching different things, and silicon-based consumers burning lots of bandwidth on behalf of their carbon-based masters. There's about a half-zillion ways a gaming console can burn bandwidth, for example. Heck, the Raspberry Pi under my TV can soak up more than 4Mbits/sec just doing a software update. /me makes popcorn and waits for 4K displays to drop under US$1K and watch the network providers completely lose their shit pgpL7JfQWAopv.pgp Description: PGP signature
Re: Net Neutrality...
On Jul 15, 2014, at 08:19 , Naslund, Steve wrote: > I don't believe either of those points. I will grant you that the LECs are > near monopolies in some rural areas, but these are few and far between. Yes, > a LEC may control the last mile but I can usually get circuits from a lot of > carriers. A company I work for has over 50 locations mostly in rural areas > and we do not have much problem getting Sprint and CenturyLink access > circuits to them regardless of location. In fact, we have never found a > location in the US that I can't get both of those carrier to deliver to us. > In a lot of areas there is also a cable provider available. Residential > users have somewhat more limited options but you do always have the option of > deciding where to live. Most of us in this group would consider the > broadband options available to them before they move. If you want more than 1Mbps downstream or more than 384k upstream over terrestrial facilities in most of San Jose, California (the 3rd largest city by population in the largest population state in the US and the 10th largest city in the US last I looked), then you have exactly one choice. If that's not a monopoly, I'm not sure how you define one. The situation in the vast majority of the bay area (including most of silicon valley) is the same. It's even worse in less densely populated areas in many cases, though USF has distorted that to some extent because there are rural areas where the monopoly facilities based carrier has taken subsidies to provide higher quality access than is currently available to many of us living in more urban areas. > Being a content provider has very little to do with market forces. Comcast > is, of course, a major content provider and access provider but if they limit > their customer's access to Netflix (which they have been accused of) the > customers will still react to that. The content providing access provider > has to know that no matter how good their content is, they are not the only > source and their customers will react to that. I think the service providers > are sophisticated enough to know that and they will walk the fine line of > keeping their customer happy while trying to promote their own content. It > is like saying a Ford dealer does not want to change the oil on your Chevy, > sure they would like for you to have bought from them but they will take what > they can get. How is a customer supposed to react to that? In a location where their choice is $CABLECO for 30Mbps/7Mbps vs. $TELCO for 768k/384k, how, exactly, does one react in a meaningful or useful way? Owen
Re: Verizon Public Policy on Netflix
I meant that comment as more of a snark that if someone wants to argue let's let the market take care of it then first we should reign in the govt-issued monopolies and small-N oligopolies. I just read, I could dig it up, that about 1/3 of all broadband users have one and only one provider, about 1/3 have 2, and about 1/3 have 3 or more. And a tiny sliver have zero, hence "about". There has been massive cross-subsidization from voice monopolies also. The whole thing stinks if one cherishes anything resembling a free and open market. But worse, much worse, are the vertical trusts. Comcast is the nation's major CATV provider with on demand and pay per view video. AND Comcast owns NBC Universal. This is like one company owning almost all the auto manufacturers, petroleum and gasoline companies, refineries, tire manufacturers, and the roads and road construction companies. And obtained all that by government fiat. All that's left, to beat the analogy to death, is one is more or less free to drive where they want. And now they're working on that! And it's getting worse not better (e.g., Comcast is trying to acquire #2 Time-Warner.) Shall we wait for them to merge with Verizon and then AT&T before we smell the coffee? Calling on the FCC to straighten any of this out is nonsense, they don't have the jurisdiction for starters. And, worse, the FCC's primary product is media censorship. What we need is the Dept of Justice and the Federal Trade Commission to enforce anti-trust law probably with the help of Congress (yeah good luck with that.) The FCC is what happens AFTER we admit that we WANT it all to be one big monopoly like AT&T was pre-breakup. Then of course we'd have to regulate that monopoly. That's why the FCC was created (and spectrum management.) Right now it's the worst of both worlds, they get the effective monopoly with protections and almost none of the regulation. We're in a pickle. On July 17, 2014 at 03:00 o...@delong.com (Owen DeLong) wrote: (me...) > > Let Comcast, TW, AT&T, Verizon, etc relinquish their monopoly > > protections and then perhaps we can see something resembling a free > > and open business climate evolve. Even that would deny that they > > already have become vast and powerful on these govt-mandated > > sinecures. > > The problem with this is that so long as service providers are allowed to be > facilities providers, there is an economic natural tendency to monopoly or > small-N oligopoly in all but the densest of population centers that will > result as a simple matter of external reality. It simply costs too damn much > to put facilities in for there to be large-N copies of facilities serving > the same area. > > That is one of the reasons I'm such a huge fan of home-run SWCs[1] with > large colos run by a facilities only provider, whether that FOP is a > municipality, NGO, or for profit entity (or even multiples if that were to > somehow be feasible). > > Owen > > [1] Serving "Wire" Center -- a hub where all of the fiber from a given > distribution area (of radius N where N < maximum reasonable distance served > by common transmission technologies available at the time of construction > with costs in reason for household usage. Today, I believe that's about 5km, > but it may be more). >
Re: Verizon Public Policy on Netflix
On Thu, July 10, 2014 8:01 pm, Jay Ashworth wrote: > Here's a link to a post from VZN's public policy blog, about Netflix. > (...) > > > http://publicpolicy.verizon.com/blog/entry/why-is-netflix-buffering-dispelling-the-congestion-myth And today, Level 3 responds: http://blog.level3.com/global-connectivity/verizons-accidental-mea-culpa/ BRB, I need to make some popcorn. Jima
Re: Verizon Public Policy on Netflix
On 7/13/14, 5:30 PM, valdis.kletni...@vt.edu wrote: > I've got a 50 pound bag of Purina Troll Chow to get rid of, so I'll opine > that a user on The World was more "on the internet" than your average > person stuck behind a NAT. And the most appropriate description of those > poor souls who are double or triple NATTed is "on drugs" You know I have a name for this now: (Keith) Moore's Law is that every conversation will eventually degenerate into a NAT debate. Eliot ps: with apologies to Keith.
Re: Net Neutrality...
"In truth, however, market failures like these have never happened, and nothing is broken that needs fixing." Prefixing a statement with "in truth" doesn't actually make it true, Bob. On Wed, Jul 16, 2014 at 10:50 AM, Fred Baker (fred) wrote: > Relevant article by former FCC Chair > > http://www.washingtonpost.com/posteverything/wp/2014/07/14/this-is-why-the-government-should-never-control-the-internet/ -- Ray Patrick Soucy Network Engineer University of Maine System T: 207-561-3526 F: 207-561-3531 MaineREN, Maine's Research and Education Network www.maineren.net
Re: Inevitable death, was Re: Verizon Public Policy on Netflix
On Jul 15, 2014, at 9:48 PM, George Herbert wrote: >> On Jul 15, 2014, at 5:02 PM, Brett Glass wrote: >> >> At 05:10 PM 7/15/2014, George Herbert wrote: >> >>> Layer3 runs right through Laramie. With a redundant run slightly south. >>> What conversations have you had with them?... >> >> At first, Level3 completely refused us. Then, they quoted us a rate several >> times higher than either of our existing upstreams for bandwidth. Even at >> that price, they refused to let us link to them via wireless (requiring us >> to either buy easements or buy land adjacent to their building, which sits >> on rented land). > > Local fiber provider? How does everyone else tie in to Layer3 in Laramie? > > And, find a Layer3 reseller who can handle the cost problem. There are a > bunch. I can recommend one privately if you can't find one. > > Buying retail markups from the vendor who wants to sell wholesale only does > not scale. The problem is partly a technological one. If you have a fiber span from east<-> west it doesn't make sense to OEO when you can just plop in a bidi amplifier. That OEO cost isn't "very high", but hitting every city like that becomes expensive quickly. This is why your 10G from EQUINIX-SJ to EQUNIX-ASH costs the same as the 10G loop from the DC to your local office. The cost is the OEO ends. If you're not in a fiber rich environment you are screwed. I have at&t fiber less than 1200 feet from me but they do not offer any non-dialtone services in my area. I'm all-poles to the end of the new comcast segment as well but due to a mid-part that doesn't have the density required to meet their metrics there continue to be only fixed wireless choices here. Others have suggested the UBNT gear. I'm using it myself, but I'll say.. it still leaves a lot to be desired. It's mostly meant for use in less developed countries. Their latest 5Ghz access gear often takes 6-12 months to get FCC certified to operate in the full 5ghz band. With the recent opening all the way down to 5.1 this spring with the FCC that certification process restarted. They are great for hopping short distances at high speeds in the US, but are very susceptible to interference. (The NanoBeam, now PowerBeam is a bit better). my backhaul is 3 miles and works well for my use case. Cheaper than the T1 before and higher speeds. There's a lot of people in wispa around the edges you can find doing things, and many others doing it that aren't in wispa. Most are small businesses (Some are larger) and suffer from poor business choices, but the biggest problem I see is lack of ability to get high speed access as Brett is commenting. Prices may be low at the major DCs but out in these areas expect $10/Mb or more, sometimes not including loop. - Jared
Re: Inevitable death, was Re: Verizon Public Policy on Netflix
When was the last time you did an ARIN request for resources for a large or x-large provider? I have reasonably recent (<2 years ago) experience doing requests for XX-Small, X-Small, Small, Large, and X-Large organizations, including 2 organizations that qualified for /24s (the max size for a large organization, one as an additional and one as an initial). I can tell you for certain that the scrutiny on ARIN's part has been very nearly identical in all cases. If anything, I got more scrutiny on some of the bigger requests than any of the smaller ones. Further, ARIN also has economies of scale, because you are paying for registration services, not IP addresses. No matter how much you want to keep trying to ignore that, the reality is that if anything, the bigger organizations are the ones potentially getting overcharged because the records for a /16 cost roughly the same to maintain as the records for a /48. There are some mitigating factors (numbers of SWIPs, frequency of additional requests, quality of request submissions, etc.), but exact cost accounting and billing based on it would be a bigger nightmare that might cost more to administer than is collected under the current system. While I will agree that some of the changes to the ARIN fee structure in the last round were not for the good, I really don't think your argument about price per IP has any merit whatsoever as it is completely divorced from the reality of what you are paying ARIN for. Owen On Jul 15, 2014, at 07:58 , Brett Glass wrote: > Matt: > > Here's the thing. With physical goods, there are economies of scale in > shipping and delivering them in bulk. But IP addresses are simply numbers! > Since there's already a base fee to cover the fixed costs, there's no > reason for the cost per IP to be different. And, in fact, good reason > for it not to be. Big carriers waste a lot of IPs compared to little > guys, who get disproportionate scrutiny. > > --Brett Glass > > At 12:24 AM 7/15/2014, Matt Palmer wrote: > >> While the "share of revenue" argument is bogus (as John's cup-of-coffee >> analogy made clear), you do have a point with the cost-per-IP-address >> argument: >> >> Annual Fee Max CIDR$/IP >> $500 /22 0.49 >> $1000/20 0.24 >> $2000/18 0.12 >> $4000/16 0.06 >> $8000/14 0.03 >> $16000 /12 0.02 >> $32000 > /12 Mastercard! >> >> Then again, the vast majority of businesses have discounts for volume >> purchases.
Re: Inevitable death, was Re: Verizon Public Policy on Netflix
On Jul 14, 2014, at 23:24 , Matt Palmer wrote: > On Mon, Jul 14, 2014 at 10:05:21PM -0600, Brett Glass wrote: >> At 09:40 PM 7/14/2014, John Curran wrote: >> >>> Myself, I'd call such fees to be uniform, >> >> Ah, but they are not. Smaller providers pay more per IP address than larger >> ones. And a much >> larger share of their revenues as the base fee for being "in the club" to >> start with. > > While the "share of revenue" argument is bogus (as John's cup-of-coffee > analogy made clear), you do have a point with the cost-per-IP-address > argument: > Annual Fee Max CIDR$/IP $500 /40 <0.01 $1000/36 way<0.01 $2000/32 way way<0.01 $4000/28 far <0.01 $8000/24 way far <0.01 $16000 /20 tremendoulsy <0.01 $32000 > /20 Mastercard! > Then again, the vast majority of businesses have discounts for volume > purchases. I note that even LARIAT does this. You charge $60 for > 1000Kbps, but $80 for 1500Kbps. Shouldn't that be $90 for 1500Kbps, to > ensure everyone pays the same price per Kbps? More importantly, in those cases, you are paying for units of a product. ARIN is a registry, kind of like your local DMV. Have you noticed that if you own more than one vehicle, you don't pay the same amount for those vehicles? Did you know that most DMV's have fleet registration discounts where you pay less per vehicle to register multiple vehicles? In the case of ARIN, you are not buying IP addresses from ARIN. You are paying ARIN for the service of registering the block(s) to you for uniqueness among cooperating entities. So arguing about ARIN fees in terms of cost per IP is absurd. Oh, and I've taken the liberty of correcting the prefix sizes in the above table to reflect the modern internet, rather than the antiquated prefix and pricing data presented before. Owen
Re: Inevitable death, was Re: Verizon Public Policy on Netflix
On Jul 14, 2014, at 21:21 , Brett Glass wrote: > Mike: > > An ASN is, literally, just a number. One that's used by a very awkward and > primitive routing system that requires constant babysitting and tweaking and, > after lo these many years, still doesn't deliver the security or robustness > it should. Obtaining this token number (and a bunch of IP addresses which is > no different, qualitatively, from what I already have) would be a large > expense that would not produce any additional value for my customers but > could force me to raise their fees -- something which I absolutely do not > want to do. Interesting... I, and many of my customers, have ASNs and are running BGP and haven't had to tweak or babysit it for years. It just cruises along doing the right thing. Generally, we only have to modify it when we add/move/change a peering and/or transit relationship. > Perhaps it's best to think of it this way: I'm outsourcing some backbone > routing functions to my upstreams, which (generously) aren't charging me > anything extra to do it. In my opinion, that's a good business move. That's fine, and from the rest of the world's perspective, your network is just another part of their network. You are invisible and irrelevant. > As for "peering:" the definition is pretty well established. ISPs do it; > content providers at the edge do not. I disagree. Many content and eyeball networks engage in a variety of forms of peering in various situations and for various reasons. The definition of "peering" is an exterior gateway protocol adjacency formed between two routers in different autonomous systems. (note, I use the term exterior gateway protocol in the generic sense, where BGP is the most prominent example du jour, not to specifically refer to the now antiquated EGP of days gone by). > Netflix is fighting a war of semantics and politics with ISPs. It is trying > to cling to every least penny it receives and spend none of it on the > resources it consumes or on making its delivery of content more efficient. We > have been in conversations with it in which we've asked only for it to be > equitable and pay us the same amount per customer as it pays other ISPs, such > as Comcast (since, after all, they should be just as valuable to it). It has > refused to do even that much. That's why talks have, for the moment, broken > down and we are looking at other solutions. Nope... Netflix is trying to help their customers and make it as easy as they reasonably can for the eyeball networks that serve those customers. Some less than scrupulous eyeball networks seem to be fighting a war to try and extort Netflix to subsidize their operations, and you have thus placed yourself in some interesting and dubious company by attempting to carry out a similar attempt at extortion. Perhaps you are emboldened by the success of one or more of these very large eyeball networks into thinking that this is how the world should operate. Perhaps something else drives your beliefs. Either way, I suspect that if your entire subscriber base disappeared from Netflix' customer roles, they would barely notice, if at all. OTOH, I suspect you get fairly regular complaints from your customers because you don't provide adequate bandwidth to enough of the internet to include reliable functional access to Netflix as part of your product line. Regardless of what you say in the fine print, your customers are expecting that they are buying access to the entire internet, including Netflix. They're asking for those packets from Netflix and once Netflix gets them to the front door of one or more of the ASNs advertising your customer's network numbers, Netflix has done their job. From there, your customers have paid you to take those bits and deliver them. Your failure to do so is just that... Your failure. Trying to get Netflix to help compensate you for a business model that doesn't provide sufficient revenue to correct the situation is absurd at best. Owen
Re: Best practice for BGP session/ full routes for customer
On 14/07/2014 18:32, Jeff Tantsura wrote: > BGP to RIB filtering (in any vendor implementation) is targeting RR which > is not in the forwarding path, so there¹s no forwarding towards any > destination filtered out from RIB. > Using it selectively on a forwarding node is error prone and in case of > incorrect configuration would result in blackholing. there are other drawbacks too: the difference in convergence time between < 24k prefixes and a full dfz is usually going to be large although I haven't tested this on an me3600x yet. Also these boxes only have 1G of memory might be a bit tight as the dfz increases. For sure, it's already not enough on a bunch of other vanilla ios platforms. Nick > > Cheers, > Jeff > > > > > -Original Message- > From: Mark Tinka > Organization: SEACOM > Reply-To: > Date: Tuesday, July 8, 2014 at 1:56 PM > To: "nanog@nanog.org" > Subject: Re: Best practice for BGP session/ full routes for customer > >> On Monday, July 07, 2014 08:33:12 PM Anurag Bhatia wrote: >> >>> In this scenario what is best practice for giving full >>> table to downstream? >> >> In our case, we have three types of edge routers; Juniper >> MX480 + Cisco ASR1006, and the Cisco ME3600X. >> >> For the MX480 and ASR1006 have no problems supporting a full >> table. So customers peer natively. >> >> The ME3600X is a small switch, that supports only up to >> 24,000 IPv4 and 5,000 IPv6 FIB entries. However, Cisco have >> a feature called BGP Selective Download: >> >> http://tinyurl.com/nodnmct >> >> Using BGP-SD, we can send a full BGP table from our route >> reflectors to our ME3600X switches, without worrying about >> them entering the FIB, i.e., they are held only in memory. >> The beauty - you can advertise these routes to customers >> natively, without clunky eBGP Multi-Hop sessions running >> rampant. >> >> Of course, with BGP-SD, you still need a 0/0 + ::/0 route in >> the FIB for traffic to flow from your customers upstream, >> but that is fine as it's only two entries :-). >> >> If your system supports a BGP-SD-type implementation, I'd >> recommend it, provided you have sufficient control plane >> memory. >> >> Cheers, >> >> Mark. > >
Re: Verizon Public Policy on Netflix
> Let Comcast, TW, AT&T, Verizon, etc relinquish their monopoly > protections and then perhaps we can see something resembling a free > and open business climate evolve. Even that would deny that they > already have become vast and powerful on these govt-mandated > sinecures. The problem with this is that so long as service providers are allowed to be facilities providers, there is an economic natural tendency to monopoly or small-N oligopoly in all but the densest of population centers that will result as a simple matter of external reality. It simply costs too damn much to put facilities in for there to be large-N copies of facilities serving the same area. That is one of the reasons I'm such a huge fan of home-run SWCs[1] with large colos run by a facilities only provider, whether that FOP is a municipality, NGO, or for profit entity (or even multiples if that were to somehow be feasible). Owen [1] Serving "Wire" Center -- a hub where all of the fiber from a given distribution area (of radius N where N < maximum reasonable distance served by common transmission technologies available at the time of construction with costs in reason for household usage. Today, I believe that's about 5km, but it may be more).
Re: Verizon Public Policy on Netflix
On Jul 14, 2014, at 08:17 , Dave Crocker wrote: > On 7/12/2014 3:19 PM, Barry Shein wrote: >> On July 12, 2014 at 12:08 ra...@psg.com (Randy Bush) wrote: >>> or are you equating shell access with isp? that would be novel. unix >>> shell != internet. >> >> You mean when you sat at a unix shell using a dumb terminal on a >> machine attached to the internet in, say, 1986 you didn't think you >> were "on the internet"? > > > An question with more nuance than most folk tend to realize: > > To Be "On" the Internet > > March, 1995 > http://tools.ietf.org/html/rfc1775 But the part that will really bend your mind is when you realize that there is no such thing as "THE Internet". Owen
Re: Verizon Public Policy on Netflix
On 17 July 2014 00:57, Owen DeLong wrote: > If Netflix had a closed or limited peering policy, then I'd say "shame on > Netlfix". If Netflix only peered > in an exchange point or two near corporate HQ and didn't have an extensive > nationwide network, I'd > say shame on Netflix. Reality is that Netflix is in most of the major > peering centers already and continues > to work aggressively to expand into more and more second-tier and > third-tier peering centers. I'd say > that is Netflix paying their share. Further, for providers that aren't in > peering centers Netflix is in, they > have offered a variety of alternative solutions and they pay a selection > of transit providers to move the > bits to providers they can't economically connect to directly. > Except they don't. Excuse me for talking about the world outside America. Netflix believes Denmark is an important enough market to pay for danish subtitles for their entire catalog and to have Denmark as a launch market for their service in Europe. But they can't be bothered to have a physical presence in Denmark. We have to go to a different country and a long way at that, to get to Stockholm in Sweden, where Netflix peers at the Netnod IX. Some danish ISPs do peer at Netnod, but it is only the ones that are big enough to qualify for a cache anyway. It is not economical to buy a link to Stockholm. Transit is cheaper, so that is what we are all doing. Then Netflix announces that you have to either have a cache or to peer directly with Netflix to get Super HD. This is a case of reverse net neutrality: The content provider is filtering content to ISPs that wont pay the transit bill for the content provider. "for the content provider" not "to the content provider". We pay our transit, it should not be our problem how Netflix pays theirs. Luckily this is so far only theory. We still get the Super HD. Either Netflix never implemented the policy or one of our transit providers made a deal with Netflix. I am not sure which one. But nevertheless even threatening to play reverse net neutrality games is NOT being the good guy. If transit is too expensive for Netflix, they should put in a shared cache at the danish IX (DIX) in Copenhagen. We would all be happy to peer with Netflix at that location. If Netflix chooses to host the cache at Interxion they also get access to the Netnod IX that covers Denmark and southern Sweden, a metropolitan area of more than 10 million people. Regards, Baldur