- NANOG - lists.nanog.org

Why can't all my routers do FOO ?
by George M Jones 09 Oct '03

09 Oct '03

I'm trying to finish off the operational security requirements draft (http://www.ietf.org/internet-drafts/draft-jones-opsec-01.txt) which is a collection of operational security requirements for routers and other network infrastructure. The last major change that needs to happen is splitting out the Best Current Practice (BCP) info from the other items. This is where I'd like some feedback. If you're so motivated take 5 minutes to brainstorm two lists. The first is "*security features* I use everywhere now" (logging, aaa, filters...). The second is everything else: "I can't believe no vendor does FOO...", "Vendor A does BAR, life/security would be so much better if all vendors did it" etc. I'll take your lists and try to correctly align them with the drafts. If you're more motivated, you can see exactly which features have migrated from the BCP draft (opsec-01a-toc.txt) to the "info" draft (opsec-info-00.txt) by looking at the table of contents of the work-in-progress drafts @ http://www.port111.com/opsec/ Replies can come to the list (preferred to avoid duplication, allow discussion) or to me directly. Thanks, ---George Jones

1 0

News coverage, Verisign etc.
by Howard C. Berkowitz 09 Oct '03

09 Oct '03

I have gotten a reasoned response from the technology editor of the Washington Post, and we are discussing things. While I wouldn't have done it that way, he had a rational explanation of why the story was written the way it was, and definitely indicating there will be continuing coverage of the issue. He believes there's always room for improving coverage.

13 19

NANOG 29 hotels
by Simon Lockhart 09 Oct '03

09 Oct '03

Gah, the hotels seem to have suddenly all got booked up. Can anyone recommend a hotel close to NANOG29, sub $200/night room rate, with internet access in the room, with availability 18-22 Oct? I can't find any on the list on www.nanog.org with availability (or are > $200/night). Simon -- Simon Lockhart | Tel: +44 (0)1628 407720 (x37720) | Si fractum Technology Manager | Fax: +44 (0)1628 407701 (x37701) | non sit, noli BBC Internet Operations | Email: Simon.Lockhart(a)bbc.co.uk | id reficere BBC Technology, Maiden House, Vanwall Road, Maidenhead. SL6 4UB. UK

5 5

Transit and Paid Peering Exchanges
by Daniel Golding 09 Oct '03

09 Oct '03

Doing some research.... Anyone have a list of Transit and Paid Peering exchange fabrics? I am interested in both US and EU locations, particularly in interesting sites like 111 8th Ave (NYC), Telehouse North (London) and other major telco hotel type facilities. I'll summarize for the list and repost. Thanks, -- Daniel Golding

2 4

History of unannounced changes to the DNS core
by Sean Donelan 09 Oct '03

09 Oct '03

Try to match the quotes up with the actions. Postel gave no prior notification that the test would take place, a fact that drew criticism. "It's caused a good deal of uncertainty and perceived instability in the system," said Chris Clough, a spokesman for Network Solutions, which operates under a contract with the National Science Foundation. "It's a concern about who can authorize changes over the infrastructure and traffic patterns of the Internet." - Associated Press, February 4 1998 "We're very concerned," Becky Burr, administrator with the Commerce Department, said Tuesday. "This was undertaken without consultation with the United States government." People trying to visit the popular Internic directory which checks the availability of a new Web address are unexpectedly being swept automatically instead to the home page for Network Solutions Inc., which offers to register Internet addresses with the com, net or org suffixes for $119. - Associated Press, March 24 1999 NSI, naturally, has a different point of view. "The registry information is our proprietary information." says NSI spokesman Chris Clough. "We've been providing it free to the community, but under the contract, all of the intellectual property gathered through the InterNIC process is our proprietary information." - ABC News, March 7 1999

2 1

Contacts for CHINANET-BJ?
by Tom (UnitedLayer) 09 Oct '03

09 Oct '03

Anyone got a clueful contact over there? Getting 100Mbps or so of dos from over there and I'd rather not just blackhole the /16 --- Tom Sparks

1 1

i'm missing my copy of "why a wildcard MX won't help sitefinder"
by Paul Vixie 09 Oct '03

09 Oct '03

at the meeting yesterday, verisign said they were considering the benefits of a wildcard MX RR whose target was a nonexistent name, as a way to keep smtp traffic from having to come to sitefinder for rejection. i recall a very good posting on this topic and i think it was on nanog but i can't find it now. can someone privately send it to me if you've got it? -- Paul Vixie

1 0

[nanog@Overkill.EnterZone.Net: Extensions to RFC1998 - WAS: Re: DoS Attacks]
by Haesu 09 Oct '03

09 Oct '03

Forwarding to NANOG on behalf of Mr. Fraizer. Please don't shoot the messenger for any arguable/discussions. -hc -- Haesu C. TowardEX Technologies, Inc. Consulting, colocation, web hosting, network design and implementation http://www.towardex.com | haesu(a)towardex.com Cell: (978)394-2867 | Office: (978)263-3399 Ext. 170 Fax: (978)263-0033 | POC: HAESU-ARIN ----- Forwarded message from John Fraizer <nanog(a)Overkill.EnterZone.Net> ----- X-Original-To: haesu(a)mx01.bos.ma.towardex.com Delivered-To: haesu(a)mx01.bos.ma.towardex.com Date: Wed, 8 Oct 2003 21:58:26 -0400 (EDT) From: John Fraizer <nanog(a)Overkill.EnterZone.Net> To: Haesu <haesu(a)towardex.com> Cc: nanog(a)merit.edu, a_spicer(a)bellsouth.net Subject: Extensions to RFC1998 - WAS: Re: DoS Attacks In-Reply-To: <20031009004646.GB88086(a)scylla.towardex.com> X-Spam-Status: No, hits=-2.0 required=5.0 tests=EMAIL_ATTRIBUTION,IN_REP_TO,QUOTED_EMAIL_TEXT, REPLY_WITH_QUOTES,USER_AGENT_PINE version=2.55 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.55 (1.174.2.19-2003-05-19-exp) On Wed, 8 Oct 2003, Haesu wrote: > Hmmmm? What did I miss? When did RFC1998 get updated to include Null > community? Feel free to let me know if they updated RFC on that > lately.. b/c I havent checked it in a while. > > As far as I know, my upstreams are fully RFC1998 compliant and I use them well. > > -hc > Note: please echo this to the list. I don't have post access. Ahem... Sue...Ahem... The RFC itself hasn't been updated to include a Null community but if you think about it, providing a NULL community is fully within the concept of allowing customers to influence routing policy with the use of community strings. For example: ! router bgp 65534 neighbor a.a.a.a remote-as 65530 neighbor a.a.a.a description Customer AS65530 neighbor a.a.a.a prefix-list AS-65530 in neighbor a.a.a.a route-map CUSTOMERS in ! ip prefix-list AS-65530 seq 5 permit x.x.x.x/y le 32 ! ip community-list standard POISON permit 65534:666 ! route-map CUSTOMERS permit 10 match community POISON set local-preference 500 set ip next-hop [ip address of your sink-hole] ! Of course, the rest of the route-map CUSTOMERS is going to need to do some sanity checks on the announcements you accept from the customers OTHER than blackhole requests. In our case, we pass them through a prefix-list match that includes: ip prefix-list CUSTOMERS seq 10 deny 0.0.0.0/0 ge 25 As you can see, we're doing a prefix-list check against the announcements that the customer sends us in the neighbor statement. Each customer gets their own prefix-list that covers the networks that we have LOA to accept from that customer. (Keeps boneheads from blackholing OTHER people!) The first stanza in the CUSTOMERS route-map checks for the POISON community. Any prefix that the customer sends us that includes this community will be routed to our sink-hole. The rest of the stanzas in the CUSTOMERS route-map look for other communities from the customer. One stanza looks to see if the customer is requesting us to pass their announcements of our address space on as de-aggregated announcements. If we don't see that community, they're aggregated. Other stanzas in the route-map are pretty cut and dry RFC1998. Our customers can do the following: Community Action --------------------------------------------- 13944:0 Don't announce to any peer 13944:1 Don't announce to PEERS 13944:2 Don't announce to TRANSIT 13944:3 Don't announce to CUSTOMERS 13944:20 Announce specific from EnterZone aggregate for customers who are running on our IPs. 13944:90 Set preference to 90 13944:100 Set preference to 100 13944:110 Set preference to 110 13944:120 Set preference to 120 13944:666 Poison a Route 13944:NNN0 don't announce to Peer NNN 13944:NNN1 prepend once towards Peer NNN 13944:NNN2 prepend twice towards Peer NNN 13944:NNN3 prepend thrice towards Peer NNN Any time I do any consulting on another network, I recommend that they at MINIMUM implement the Poisoned Route ability. It is not terribly difficult to do as you can see above. -- John Fraizer EnterZone, Inc (13944+$|13944+_14813+$|13944+_17266+$) PGP Key = 6C5903C4 Fingerprint = 2AA6 6614 1B5E EDD2 38AD C417 3E61 F975 6C59 03C4 ----- End forwarded message -----

1 0

More news coverage
by Howard C. Berkowitz 09 Oct '03

09 Oct '03

Associated Press at http://www.boston.com/business/technology/articles/2003/10/08/experts_descr… I am talking with the reporter, Ted Bridis. Associated Press reporters are under about as tough a deadline as TV, so it's a good fast first pass.

6 6

RE: Verisign on Process
by Howard C. Berkowitz 09 Oct '03

09 Oct '03

At 5:23 PM -0700 10/8/03, David Schwartz wrote: > > Is it possibly time to suggest that perhaps ICANN should >> call for formal separation of regiSTRAR functions from >> regiSTRY functions, and stipulate that stewards of record >> for regiSTRY functions not participate in regiSTRAR roles? > > Already done -- read the contact between ICANN and VeriSign. > >> Certainly it's been shown to be very difficult to resist >> the temptation to extend editorial control over what entries >> get placed into the DNS records as a regiSTRY if you also >> happen to be able to increase the profits from your regiSTRAR >> role. > > Sure, but you could also do it to increase the profits from >your registry >role, which seems to have been the intent of SiteFinder. > >> If the functions are stipulated to be kept separate, > > They are. > >> then we have a much better opportunity to engage a system of >> checks and balances, to self-limit potential future abuses >> like this. > > How would that stop a registry from, for example, adding a >wildcard record >that goes to pages sold to the highest bidder? Registry functions are >supposed to be wholly ministerial. VeriSign doesn't get that and ICANN is >going to have to force them to -- probably more than once. > > DS

1 0