- Test - lists.nanog.org

maximum active vlans in a crisco 6509
by Bob Biver 21 Jun '00

21 Jun '00

the docs say max 250, is this informational or a limit of spanning tree? also, is anyone running more that 250 ? thx bob ________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

7 7

DACS
by Bob Biver 21 Jun '00

21 Jun '00

Digital Access and Cross Connect System. A TSI switch that switches DS-0 channels among DS-1 data streams. Computer controlled and designed for keyboard entry as opposed to automatic switching. There are many uses, including semi-automatic test boards, digital circuit concentration, trunk switching, and alternate routing. The term "DACS" was extended for DS-0/ DS-1 switching but has been applied to DS-1/ DS-3 switching. Many large "DACS" units include combined switching systems that allow DS-0 switching among DS-3s, always with demultiplexing of the DS-3 data streams. IXC and LEC carriers offer partitioned DACS service under customer control, which allows the customer to control a Virtual Private Network (VPN) within the carrier's network. ________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com

1 0

RIPE NCC CIDR blocks - smallest allocations
by Joao Luis Silva Damas 21 Jun '00

21 Jun '00

During the RIPE 36 meeting in Budapest there was a request to publish the list of CIDR blocks allocated by the RIPE NCC and the smallest allocation issued from each of them. This is it: 62/8: /19 193/8: /29 # Contain small PI assignments 194/8: /29 # Contain small PI assignments 195/8: /29 # Contain small PI assignments 212/8: /19 213/8: /19 217/8: /20 We shall put this information permanently on our website. regards, Joao Damas Head of External Services RIPE NCC

1 0

Re: DACS failure in Tulsa, OK
by Mufti Ahmed 21 Jun '00

21 Jun '00

Hi Sean what is a DACS ( Data Access Control System)? I'm asking because in Reuters we use that as a permissioning system for Market Data clients. Mufti Nayeem Ahmed Network Systems Engineer Market Data Networks Reuters America Inc. (212)-603-3595 1-800-2REUTER 1-800-PAGE-MCI pin# 1569803 To: nanog(a)merit.edu cc: (bcc: Mufti Ahmed/NYC/US/Reuters) Subject: DACS failure in Tulsa, OK Southwestern Bell lost a DACS in its Tulsa-Elgin central office yesterday. When it lost utility power, approximately 627,000 customers lost service throughout northern Oklahoma. This affected Internet, cellular, 911, and other telephone services in the area. ----------------------------------------------------------------- Visit our Internet site at http://www.reuters.com Any views expressed in this message are those of the individual sender, except where the sender specifically states them to be the views of Reuters Ltd.

2 1

DACS failure in Tulsa, OK
by Sean Donelan 21 Jun '00

21 Jun '00

Southwestern Bell lost a DACS in its Tulsa-Elgin central office yesterday. When it lost utility power, approximately 627,000 customers lost service throughout northern Oklahoma. This affected Internet, cellular, 911, and other telephone services in the area.

1 0

RIPE NCC ADDRESS POLICY CHANGES
by Joao Luis Silva Damas 21 Jun '00

21 Jun '00

*** CHANGES IN RIPE NCC ADDRESS POLICY *** During the RIPE 36 meeting in Budapest (16-19 May 2000) the LIR working group agreed that the RIPE NCC will: - lower the minimum allocation from a /19 to a /20. - this change will take effect starting August 1 Allocations of this size will initially be issued from CIDR block 217/8. Minimum allocations do not constrain requests for address space that can justify bigger blocks of addresses. ******************************************************************************** ** If you use prefix based filtering in your routers, please note this change ** ******************************************************************************** regards, Joao Damas Head of External Services RIPE NCC

1 0

Global Crossing
by Shawn McMahon 21 Jun '00

21 Jun '00

Anybody seeing any problems with Global Crossing (gblx.net) today?

1 0

using IRR tools for BGP route filtering
by Peter Francis 20 Jun '00

20 Jun '00

How would one extract the following information from the IRR: the list of AS's which are non-transit customers of a given backbone provider Peter Francis Sr. Network Engineer SoftAware Inc.

4 3

FW: Private and Public Peering
by Lane Patterson 20 Jun '00

20 Jun '00

Hansen, to directly answer your original question: 1. if ISP A peers with ISP B through a switch run by an exchange point operator, this is generally considered to be public peering. The same switch may be aggregating many other ISP's giving O(N^^2) peering relationships. 2. if ISP A peers with ISP B by connecting to ISP B's switch (which then may aggregate into ISP B's router), this is considered to be private peering, and scales to O(N) peering relationships for ISP B. This seems to be a more common way for ISP's to privately aggregate smaller (10 mbps, 100 mbps) peers to a switch, which can then use a GbE trunk to a router. However, I believe the question limits the answer and that we should think about operational peering interconnection as more than public vs. private: 1. Public Peering in Public Places: Peering via FDDI, ATM, Ethernet 3rd party exchange switches. Great for cost-effectively aggregating larger numbers of smaller peers (and scaling multicast trees). 2. Private Peering in Public Places: cross connects between colocated peers, ala Equinix, PAIX, ... bilateral control over fewer large flows, generally for Tier 1<->1 and Tier 1<->2. 3. Private Peering between Private Places: circuit from ISP POP A to ISP POP B. These aren't exact. There are a number of operational or technical gray areas: - POP-friendly buildings like Westin Bldg in Seattle can act as a "partly public space" via fiber pulls and cabling "meet-me" rooms - DPT/SRP peering ring is a multiaccess "public peering" fabric but generally seen as a way to interconnect a small number of private POPs on a ring (although extending a ring to a "public place" could be a variation) - 3rd party DACS (for SONET) and eventually OXC (for lambda trades) can scale/channelize large numbers of private cross connects. Is this a form of private peering, or public? This may depend on operational, business, and technical definitions, but it's probably a mix of 1. and 2. Cheers, -Lane > From: Scott Marcus [mailto:smarcus@genuity.com] > Sent: Monday, June 19, 2000 5:46 AM > > > At 21:42 06/18/2000 -0700, Steve Feldman wrote: > > > >The closest to a real distinction that I've been able to come > >up with is whether or not a third party is involved > >*and* whether packets (or cells) are switched. > > > >For example, peering through a FDDI, Ethernet, or ATM switch > >is always called "public" (unless perhaps the switch is > >owned by one of the peers). And peering through a wire or > >SONET/SDH circuit seems to always be called private, > >even though the data might pass through SONET/SDH > >multiplexers, cross-connects, and switches operated by > >a third party... > > > Steve, this happens to be true in most cases, but I would > view it as being > sort of coincidental. > > What most people term "public peering" is effected at a public traffic > interchange point, where many providers appear, or can choose > to appear. > Note that the decision whether to interconnect is still, in almost all > cases, a bilateral business decision between each pair of > providers. The > word "public" is thus something of a misnomer. > > It's usually implemented over FDDI, ATM, whatever, because the traffic > interchange point needs to implement any-to-any connectivity. > That's a > matter of engineering practicality, but the definition should > not rest on it. > > What most people mean by "private peering" is a direct interconnection > between two providers. That's most often implemented over a circuit > between the two, without either deploying equipment to the other's > premises; again, however, that's simply a matter of engineering > convenience. These connections are conceptually point-to-point. > > Hope this was helpful, > - Scott >

2 1

Re: uunet fiber cut (pgh, pa)
by Sean Donelan 20 Jun '00

20 Jun '00

On Mon, 19 June 2000, deeann mikula wrote: > does anyone have the scoop on the "major fiber cut" in the UUnet > backbone affecting areas around pgh, pa? Worldcom has a fiber cut affecting at least 11 OC48s between Pittsburgh and Washington DC. Folks have reported that UUNET's Pittsburg hub was isolated from the rest of the network. The best estimated time to repair is sometime overnight. Also Sprint reloaded most of their routers recently, and there have been sporadic reports of problems in different Sprint cities yesterday. I don't have good information about the reasons.

2 1