- Test - lists.nanog.org

RE: Faster 'Net growth rate raises fears about routers
by Richard Noonan 04 Apr '01

04 Apr '01

My experience with these issues (as a customer): I have a corp lan and space in 6 colo's around the US. Connectivity between corp and the colo's is mission critical. Initially this was addressed by two T1's each from different providers and we advertised a /24 for the corp. This was hugely benficial b/c just prior to one of the providers imploding problems arose with the circuit that they could not manage to resolve with the telco (probably b/c they weren't paying their bills). Data could still get down the line at a poor rate and I was able to maintain it as a backup to the good circuit until another was provisioned. I was able to do this effectively and quickly b/c I had a measure of control over my routing. I intended to get the 2nd circuit from another provider, but my existing (good) provider asked for the opportunity to convince me otherwise. They happen to be a smallish, local ISP and I happen to deal directly with one of their principals who is an excellent network engineer. He came in and made one of the most detailed and lucid proposals I have ever seen from a vendor. The bottom line was, I have one fiber access into the building. Nothing to be done from me to the CO. But, he was able to get me a 2nd circuit through Worldcom (first being Verizon) take it to another POP which is in-turn connected to a different set of carriers than the other. "Oh, and by the way, this 2nd circuit with us will be dirt cheap." Sold, with caution. I still maintain a BGP session to him and advertise the /24. Why? If this turns out not to have been such a great idea, I can cancel that 2nd circuit, go back to plan A, with minimal re-engineering. If things continue to run smoothly for another 6 to 12 months, then maybe I'll feel confident enough to turn over all routing control to him. My points: We have colo'd as much as possible, but in the end we still have a critical need to reach those facilities. Maitaining an EBGP session and advertising a non-aggregated block gives me control and flexibility. However, the right provider can possibly provide another solution. We should see more providers marketing this and maybe I wouldn't be so cautious. But, a big part of why I was even willing to experiment with this was what I felt to be the quality of engineering on the ISP's side and total understanding of the situation I was in (in addition to having direct access to their knowledgable, senior staff at any hour). This is, in my experience, simply not the norm and an ISP customer is forced to take the matter into their own hands. Hope I haven't been too redundant, but half my time is spent as exactly the type of customer you're all discussing. -Rich -----Original Message----- From: Travis Pugh [mailto:tpugh@shore.net] Subject: Re: Faster 'Net growth rate raises fears about routers On Tue, 3 Apr 2001 hardie(a)equinix.com wrote: > Why are customers spending money on this? My belief is that they > want more say in their own fate. That may express itself as > a desire for redundancy in the case of catastrophic business failures, > better ability to express their own routing policies, or a simple > worry that they won't get the best price if they have only one supplier. > At the core of this, though, is a desire for more control over something > that they see as increasingly important to their own fate. > <text removed> I'm at a multi-POP network in Boston. We've had great luck selling customers a Verizon circuit into one of our POPs and a Worldcom circuit into a different one. It costs more, but they don't have nearly the exposure of a single circuit customer. However, if you're not set up to do this, the appropriate level of paranoia calls for circuits to two different providers. Maybe if SPs really addressed availability requirements of their customers, it wouldn't be such an issue. -travis

1 0

Re: Verio's NOC is closed?
by Sean Donelan 04 Apr '01

04 Apr '01

On Tue, 03 April 2001, doug(a)safeport.com wrote: > > Anyone know what is going on here? I'll let Verio explain, if they want to. Sometimes external factors require a company to relocate the folks who answer the phones.

1 0

Re: Faster 'Net growth rate raises fears about routers
by Greg Maxwell 04 Apr '01

04 Apr '01

On Tue, 3 Apr 2001, RJ Atkinson wrote: > At 08:37 03/04/01, Greg Maxwell wrote: > >Replace the internet with a highly aggregated IPv6 network > >which uses transport level multihoming and you gain a factor > >of 1000 improvement at core routers (and 100,000x further > >from the core where you no longer need to be default-free) > >and still have the oppturnity for a further 5x by going > >to a state-of-the-art CPU (providing that your cpu speed > >reasoning is valid). > > Precisely which "highly aggregated IPv6 network > which uses transport level multihoming" is one talking about ? > What's the RFC on this ? It's possible to 'solve' these problems in the future: Forbid IP level multihoming for IPv6 which crosses aggregation boundaries. I.e. absolutly no multihoming that inflates more then your providers routing tabling, connect to whoever you want, but no AS should emit a route for any other AS without aggregating it into their own space without a special agreement of limited scope (i.e. not globally!) > AFAIK, IPv6 multihoming is identical to IPv4 multihoming, > with all the same adverse implications on the default-free routing > table -- hence the creation of an IETF MULTI6 WG to try > to change this. If I've missed some recent advance in the > IETF specifications, please share the details (preferably citing > RFC and page number :-) with the rest of us. IPv6 multihoming *is* the same. What I argue is that: Routers are the wrong place to do multihoming for anything but Tier-1 connectivity. Multihoming belongs in the end node. SCTP (2960) is a transport level protcol which offers what amounts to a superset of TCP. One of it's features is multihoming (2960; section 6.4). With such a protcol it is possible to accomplish all of the realibility benifits of IP multihoming while achieving much greater scalability, flexibility, and performance. We need to stop looking at IP addresses as host-identifyers (thats what DNS is for) and look at them as path-identifyers. I'm not sure of an RFC detailing specifics of an Internet architecture based on these concepts, thats one of the reasons I asked here. I'll troll around more and see if I can find anyone working on such a document and if no one is, I'll create one.

8 13

RE: Faster 'Net growth rate raises fears about routers
by Richard Welty 04 Apr '01

04 Apr '01

Travis Pugh <tpugh(a)shore.net> wrote: > Verizon even has a diversity > option on their tariff, although there's some doubt as to > whether or not > they bother to make sure it is really diverse. i have had some discussions about getting such diversity with Verizon folks in upstate NY, and it was clear that their engineers understood the issues, giving us all the right answers about how they were going to do it. there still would have been some risks involved; the diverse circuits would have crossed at one specific location in the RPI tech park (they pointed this out explicitly), and the connection to the IXCs on the other end of the local Sonet ring would have been in a single CO, and thus still a single point of failure, but this was really hard to avoid. they were quite honest with us, actually, which was a complete suprise given that it was the phone company (we had previously talked to other Verizon reps who were total idiots, so if someone disagreed with me on this assessment, i would expect that they were dealing with one of the idiots and never got past that to the folks who knew what they were doing). since we didn't go forward with the project in question, i have no idea how it would have worked out in practice. richard

1 0

Correction - gTLD Informational Message]
by Brad Verd 04 Apr '01

04 Apr '01

Please note that the correct date is Wednesday evening the 4th of April.

1 0

RE: Faster 'Net growth rate raises fears about routers
by Roeland Meyer 04 Apr '01

04 Apr '01

> From: Travis Pugh [mailto:tpugh@shore.net] > Sent: Tuesday, April 03, 2001 10:34 AM > I'm at a multi-POP network in Boston. We've had great luck selling > customers a Verizon circuit into one of our POPs and a > Worldcom circuit > into a different one. It costs more, but they don't have nearly the > exposure of a single circuit customer. However, if you're > not set up to > do this, the appropriate level of paranoia calls for circuits to two > different providers. Maybe if SPs really addressed availability > requirements of their customers, it wouldn't be such an issue. The problem with this, if done, is that we back right into the other problem of prefix filtering. If the customer has a /19 or /20, there is generally no problem. But, if it is the usual case (/24) then only one of the upstreams can aggragate the routes up. What is the other ISP to do? How would this be made to work? BTW, this is exactly the reason we weren't fully multi-homed yet. Yes, greg described a way where both interfaces (end point) were NAT'd. However, I have a concern with brittleness and tinker-factor there.

3 2

Need Help
by doug＠safeport.com 04 Apr '01

04 Apr '01

Your network center is really closed?? I do not have my customer number with me. You can lookup by company name: Safeport Network Services. My first question is when will you be back in the office? _____ Douglas Denault doug(a)safeport.com Voice: 301-469-8766 Fax: 301-469-0601

2 1

Provider change (was RE: Faster 'Net growth rate raises fears about routers)
by Sean Donelan 04 Apr '01

04 Apr '01

I'm sure the telesales people are making those phone calls as this thread goes on. The FDIC has this down to a science with banks. The bank closes on Friday evening, the FDIC swoops in, and by Monday all the accounts are moved to a new bank. FEDWIRE reroutes all the traffic to the new service provider, and the checks "in-float" don't even bounce. Besides the potential route table explosion as the network is deagregated, another problem is how do you verify who has what authority to change which circuit id, ip address, etc. Does everyone just start grabbing address blocks, do providers include the CIDR block as a line item on the invoice. On Tue, 03 April 2001, Randy Bush wrote: > what if an isp with a lot of dedicated line customers in old A space takes > the dive and a few thousand of their customers come to isps each asking to > have 38.42.666.0/24 or whatever routed? > > mirjam says > o route them now, forcing instant renumbering would add insult to injury > o give them the normal 90 days to renumber > > and she's never led me wrong so far.

1 0

RE: Faster 'Net growth rate raises fears about routers
by Roeland Meyer 03 Apr '01

03 Apr '01

Sorry, I thought we were discussing all multi-homing. Your example doesn't help the business whose ISP suffers a business failure (such as DSLnetworks, Flashpoint, et al), only the case where the access provider fails SLA. To put it bluntly, a single circuit, to a single business, while annoying, doesn't cause wide-spread outages when it fails. It is thus, the lesser-order case. The case I thought was under discussion is when an ISP dumps something on the order of 10^3 or more customers when they fail. I understand that NorthPoint abandoned ~100,000 customers when they sold their backbone to AT&T and AT&T didn't pick up the subscribers. I will wager that many of them were /24s. DSLnetworks had over 700 Covad customers, FlashPoint was larger. For various definitions of "wide-spread", this is a much larger issue than a broken copper-pair. I suspect that it also has a much higher likelyhood of occurance. Especially, in the current business shakeout. Guess what ... it won't stop. This sort of problem will be with us forever. We should find a solution ... someday ... ya think? Business failures are on one side of the problem and CIDR aggregation is on the other. > -----Original Message----- > From: RJ Atkinson [mailto:rja@inet.org] > Sent: Tuesday, April 03, 2001 11:26 AM > To: Roeland Meyer > Cc: nanog(a)merit.edu > Subject: RE: Faster 'Net growth rate raises fears about routers > > > At 13:55 03/04/01, Roeland Meyer wrote: > > >The problem with this, if done, is that we back right into > the other problem > >of prefix filtering. > > No. These are separate tail circuits to separate POPs > of the same ISP. So that one ISP only needs to advertise its > fully aggregated prefix. So the problem you postulate does > not arise in this particular situation. > > > What is the other ISP to do? > > You didn't read closely enough. 2 tail circuits, > 2 POPs, but only 1 ISP was the scenario outlined. It works > quite well, provided one picks the ISP thoughtfully. > > Ran > >

4 3

RE: Faster 'Net growth rate raises fears about routers
by Roeland Meyer 03 Apr '01

03 Apr '01

> From: Bill Woodcock [mailto:woody@zocalo.net] > Sent: Monday, April 02, 2001 2:51 PM > > On Mon, 2 Apr 2001, Roeland Meyer wrote: > > From what I just went through (DSLnetworks v Covad), > this is straight > > bulldust. Had I been properly multi-homed, MHSC would > NOT have been down for > > 5 weeks. > > Patrik said ISPs, not IAPs. So, please explain to me how not being multi-homed is anything other than a bad-thing and high-risk? No, I am not including colo, because it is assumed that you know what their arrangements are before you "buy". Reputable colos are multi-homed, in spades.

9 9