August 2001 - Test - lists.nanog.org

RE: multi-homing fixes
by Roeland Meyer 28 Aug '01

28 Aug '01

|> From: John Fraizer [mailto:nanog@Overkill.EnterZone.Net] |> Sent: Tuesday, August 28, 2001 11:56 AM |> |> > Multihomed customers generate LESS technical support rather than |> > more? Maybe we should urge more people to multihome... |> |> "Multihomed customers generate LESS technical support rather than |> more?" Perhaps in an outage instance they do. When they |> have a routing |> issue, it can become a finger-pointing match between their |> upstreams and |> thus, it generates 100% more technical support for the upstream who |> doesn't have the problem. only until the mal-functioning provider is replaced.

1 0

Re: DNS Parser
by Paul Vixie 28 Aug '01

28 Aug '01

> Does anyone know of a simple script that will go through a list of BIND > primary files, or more and output a list of output like this: > > secondlevel.domain --> address > thirdlevel.secondlevel.domain --> address > > and so on for several thousand domain files. I agree with the comment that you ought to use AXFR rather than parsing zone files, since the zone files can have all kinds of irregularities that make it a complex job to parse them, whereas an AXFR client will have predictable output. Much as it pains me to recommend the use of "nslookup" by anybody for anything, it has a reasonable way of doing this: nslookup <<EOF | awk yada yada server $MASTER ls -t A $ZONE EOF If you prefer "dig" then you'll shortly be cursing me for its non-optional shorthand pretty-printing, but once in 1995 when I got done cursing myself for this misfeature, I wrote $BIND8/contrib/misc/normalize_zone.pl, as in dig @$SERVER $ZONE axfr | bind8/contrib/misc/normalize_zone.pl | grep ' IN A ' Hope this helps.

2 1

RE: multi-homing fixes
by Roeland Meyer 28 Aug '01

28 Aug '01

1 0

RE: multi-homing fixes
by Roeland Meyer 28 Aug '01

28 Aug '01

|> From: Leo Bicknell [mailto:bicknell@ufp.org] |> Sent: Tuesday, August 28, 2001 8:12 AM |> |> On Tue, Aug 28, 2001 at 08:00:40AM -0700, Sean M. Doran wrote: |> > I urge the rest of you to invoice him for your |> 50-cents-times-n, today. |> |> In the spirt of the market as of late.... |> |> Roland, you don't want Sean to route your prefixes, you want me to |> route them. As such I'm going to give you a deal, just 25 cents |> per prefix! |> |> (I'm sure no one else will want to price it lower. :-) Considering our current upstream arrangements, please respond to RFP posted at http://www.rotflmao.bid. At the moment, we are practicing safe exchange using NAT, with multiple providers, but we are not getting full satisfaction from the relationships. Not all of our services are being utilized to the full enjoyment of our customers. Some hardcore services, like SSL, are not even possible with these relationships. We are looking for two, potentially long-term, relationships that will culminate in mutual satisfaction, where we will be safely protected from any mis-judgements of either of the parties involved. We are looking for a bundled price that includes all connectivity, interchangably with either of the parties, or both. We are looking for two lowest bidders. We don't tolerate incest so those IP-addrs have to be as widely separated as possible, for at least two generations. Two ISPs with the same upstream wont cut it. Even if those upstreams share a common upstream, it is grounds for disqualification. We are looking for good lineage, there is just too much risk of inheriting downstream business problems and we have to protect the well-being of our current customers and future customers that we may both generate. Please direct bids to mailto:harry@palmpilot.rotflmao.bid.

1 0

Re: multi-homing fixed
by smd＠clock.org 28 Aug '01

28 Aug '01

| I've asked this of a number of people now, but how many providers | have multiple POP's in a city that are _completely redundant_? | That is, they can operate _fully_ with one POP out of service? The suitcase nuclear bomb that takes out my facility in one location is very likely to take out my other facility in the same metro area. And probably all the customers who would remotely care. | Even if they have two pops, many of those cities won't have redundant | long haul capacity. The bigger problem is actually that many of those cities will have zero customers who are willing to pay extra to be connected to a router in POP-A and a router in POP-B, and I know of no way to make such a connection cost the same or less than a single connection to only one of those routers. So, while it's _possible_ to make the Big Red Switches' states very nearly totally invisible to any customer, it's not often done. But, hey, deep-pockets people have found out that several suppliers are eminently capable of engineering (nearly) all the resilience and redundundancy that the deep-pockets can manage to pay for... Sean.

6 6

Re: multi-homing fixed
by Rachel Warren 28 Aug '01

28 Aug '01

> > A bomb threat is definately something to worry about, though. > > Once upon a time, while talking with a customer about a particular LA > datacenter's fault tolerant features (you know, generators, UPS, HVAC, > diverse fiber entrances, etc...), they very calmly asked me what > procedures we had in place in the event that a major earthquake struck > the southland and dropped LA into the ocean or otherwise put the city > underwater. I told them very bluntly that if an earthquake of that > magnitude were to hit LA, I would *not* be overly concerned with their > servers/circuit/etc., assuming I was still breathing. They were quite > taken aback. ;) I guess I should have been more clear. I'm not talking about nuclear bombs. I was talking about regular explosives (for lack of better wording) - pipe bombs and the latter. The type of quake you are talking about is up in the nuclear-bomb stratosphere, and that's not something that would keep me up at night worrying. Plus, I live on the east coast. We don't have to worry about these things called earthquakes for the most part. ;-) Rachel -- Women and elephants never forget. - Dorothy Parker

3 2

RE: multi-homing fixes
by smd＠clock.org 28 Aug '01

28 Aug '01

"David Schwartz" <davids(a)webmaster.com> writes: | I'm not sure I believe that this tragedy of the commons exists where people | route on allocation boundaries. If I make Sprint carry an extra route just | for my little network, that helps all Sprint customers reach my little | network. Cool, so you can get them to contribute a fraction of the 50 cents you owe me? Since I live downstream from Sprint, I tell you what: I'll give you the first flap for free. And that's a generous offer, given that your little network represents approximately zero percent of my traffic, and I doubt I'll be complaining to my other personalities that I can't get at your fascinating content without subsidizing your network's globally-visible dynamicism. Sean. (who does, incidentally, subscribe to some pay-for-play web sites, for example, which helps them pay for a part of THEIR dynamicism. i must remember to ask for a 50 cent * n discount at renewal time...) P.S.: the point here, since it's easy to miss, is that a clearing house function is useful for containing complexity of negotiation, and is currently being done by the RIRs on behalf of their owners business opportunity: a tool which reliably lets one decide if prefix x is worth carrying in all its flapping glory, or whether it should be bitbucketed until a bill is paid

2 1

wow, this is off-topic but very interesting (nanog bounce messages)
by smd＠clock.org 28 Aug '01

28 Aug '01

Wow, it seems that I'm a male chauvinist pig. Sean. | From: System Attendant <MAIL-SA(a)syndesis.com> | To: "'smd(a)clock.org'" <smd(a)clock.org> | Subject: ScanMail Message: To Sender, sensitive content found and action t | aken. | MIME-Version: 1.0 | | Trend SMEX Content Filter has detected sensitive content. | | Place = davids(a)webmaster.com; patrick(a)cybernothing.org; randy(a)psg.com; | rmeyer(a)mhsc.com; nanog(a)merit.edu; | Sender = smd(a)clock.org | Subject = Roeland Meyer owes me money! (was re: multi-homing fixes) | Delivery Time = August 28, 2001 (Tuesday) 10:16:36 | Policy = Sexual Discrimination | Action on this mail = Quarantine message | | Warning message from administrator: | Sender, Content filter has detected a sensitive e-mail. |

2 1

RE: multi-homing fixes
by smd＠clock.org 28 Aug '01

28 Aug '01

AVG writes: | Therefore, the cost of extra prefix is definitely not $0.50; it is _much_ | higher. Quite likely, considered globally. But personally, I'll be happy to get Roeland Meyer's prompt payment of the 50 cents (per flap, per router, per prefix) he owes me, as a start in discovering a market price for my routing slots. I urge the rest of you to invoice him for your 50-cents-times-n, today. Sean.

2 1

RE: multi-homing fixes
by smd＠clock.org 28 Aug '01

28 Aug '01

| ARIN could decide to issue /32's to dialup customers so they could | change providers without renumbering and it would do no good since nobody | would carry those routes. This is an excellent suggestion, and I have made it to RIPE people from time to time when this issue comes up. Some slow-start fine-tuning means that technonerds with DSL, cable, dialup and so on and several PeeCees might grow into a /28 someday! (Or is that a /48?) | [It would] harm the net as a whole if RIRs adopted a | microallocation policy that ... resulted in ... allocating | non-routable IP space. Not really. It would force some issues that people would rather not deal with. Most notably, it would force the issue of how providers exchange reachability information, which has languished while the same providers have focused on how providers exchange traffic attracted by that reachability info. Right now, the RIRs seem to understand how the conversations along the lines of "if you send me <prefix x> i will route it to you as long as you route <prefix y> to me" enough that they don't actually have to happen bilaterally. Maybe that could change, given that at least one of the RIRs is developing some opinions on how those conversations SHOULD work rather than how they WOULD work. Any catharsis of spurious morality couched in "for the good of the Internet" (or the small guy or the small ISP or the big banks, or whoever your favourite cause is) rhetoric is always welcome, by me. | Don't make any sense. That was my point in replying to you. There is no | right to a route in my router. If you want a route in my router, you better | find out what routes I'm willing to carry and under what terms. Is it still, loosely, "did you get this prefix/mask allocated by an RIR"? Or are things breaking down already? Sean.

1 0