- Test - lists.nanog.org

RE: Hey, anyone who wanna help improve the net stability?
by Peter Ford 06 Sep '96

06 Sep '96

It seems the problem is number of flaps, etc. Doesn't it make sense to simply start dropping the routes that flap the most? Of course this will be biased in the /24 space but at least the rationale/solution maps identically to a solid reason. To what extent is this flapping due to basically static information being propogated? One solution would be to nail these routes ... Cheers, peter ----------

3 2

mae-west fddi saturation
by matthew＠scruz.net 06 Sep '96

06 Sep '96

I'd use the mae-west list for this, but several people have said that they're not getting messages on that list. This graph: http://www.mfst.com/MAE/west.mfs.overlay.html is showing what appears to be hard saturation. But it is at way below what I thought the saturation speed of FDDI is. Has anyone gotten a good answer as to what is broken there? -matthew kaufman matthew(a)scruz.net

1 0

Re: customers and web servers and level one naps
by Peter Lothberg 06 Sep '96

06 Sep '96

[...snip] > brought there by the customer. therefor should it really make any > difference to have the web traffic avoid the extra hops of > traversing the local loop? The local loop could be dark fiber running FDDI or OC3 packet-over-sonet, in the latter case the NAP is the bottleneck, not the local_loop. > what am I missing? Hummm.. Tempting... --Peter

1 0

Re: customers and web servers and level one naps
by Peter Lothberg 06 Sep '96

06 Sep '96

> Is PAIX a "major" exchange Gordon? What is the PAIX? (I have hosts behind a router at the exchange in my bathroom) --Peter

5 5

Re: customers and web servers and level one naps
by Peter Lothberg 06 Sep '96

06 Sep '96

> Second: allowing such a customer, or an NSP, to attach web services > directly to the FDDI ring at the NAP. This is a security problem, if there is no switch in the middle and each host are induvidually attached to the switch. Next problem is that a host needs to knew what router to send a packet to for a particular destination, so either it points default at one of the NAP routers, and packets traverse the NAP twice, or the host impleements BGP and has a full set of routes. So host at a the NAP media should be 'strongly not recomended'. An interesting scenario is, a router with two FDDI interfaces, one to the host and one to the NAP. It now comes down to if it's worth the real_estate to have the host there. --Peter

2 1

RE: Hey, anyone who wanna help improve the net stability?
by Jim Fleming 05 Sep '96

05 Sep '96

On Thursday, September 05, 1996 9:52 AM, bmanning(a)isi.edu wrote: @ > Of the class cs from the swamp how many are now being routed at the @ > defaultless core? HOW MANY ADDITIONAL CLASS Cs FROM THE SWAMP ARE THERE @ > FOR WHICH THE OWNERS COULD DEMAND ROUTING? @ > @ > In other words are these 5,000 new class c's just the beginning? Or are @ > they, hopefully the end? @ @ Lets see, what the BOE math tell us... @ @ 24,000 entries now, @ 192.0.0.0/8 is ~2million. @ @ ~2mil - 24,000 = current router death. @ @ -- @ --bill @ @ Is now when we say..."we told you so..."....??? You are seeing the great allocation policies and procedures at work... fortunately, there will be other address spaces and the IPv4 space can be wrapped up as one TWD... -- Jim Fleming UNETY Systems, Inc. Naperville, IL e-mail: JimFleming(a)unety.net

2 1

RE: Hey, anyone who wanna help improve the net stability?
by Paul Ferguson 05 Sep '96

05 Sep '96

[cc: list mercifully trimmed] Jim, Regadless of whether the address space is IPv4 or IPv6 (or IPv8 or whatever nonsense), the same problem would exist with this number of prefixes. This is *not* an issue of 'which address space is hosed', but rather, how to put into place practical allocation and aggregation strategies to keep the total number of prefixes to a minimum. It would appear that while you seem overly anxious to say 'I told you so', you've completely ignored the technical discussions on why the problem exists in the first place, and why the allocation and aggregation methodologies need vast improvement than history has shown us. This is completely orthogonal to 'other address spaces'. - paul At 05:15 PM 9/5/96 -0500, Jim Fleming wrote: > >Is now when we say..."we told you so..."....??? > >You are seeing the great allocation policies and procedures at work... > >fortunately, there will be other address spaces and the IPv4 space >can be wrapped up as one TWD... >

1 0

Re: customers and web servers and level one naps
by schnell＠gsd.sprint.com 05 Sep '96

05 Sep '96

Is the gentleman who originally posed the question talking about the NAPs awarded by NSF solicication NSF93-52 (of which there are only three), or in the generic sense (all network access points, including the PAIX). If you research the solicitation I don't believe you'll find suggestions of supporting non-NSP connections or host-based apps other than the RA. Steve At 11:07 9.5.96, bmanning(a)ISI.EDU wrote: >> >> What are the positions of Ameritech, sprint, pac bell and MFS on two >> issues? >> >> First having someone at a NAP who is there as a customer of one of the >> National service providers... and who may or may not engage in cost free >> peering with any of the other NSPs or even with other isps? >> >> Second: allowing such a customer, or an NSP, to attach web services >> directly to the FDDI ring at the NAP. >> >> PAIX is doing this. As far as I know the other major interchange provider >> are not. I am wondering why. >> > > Is PAIX a "major" exchange Gordon? > > >-- >--bill ******************************************************************************** Phone: 1.816.854.2113 Fax: 1.816.854.2201 Sprint Pager: 1.888.366.7890, PIN 398.6644 Text Page via Internet: 3986644(a)pagenet.net ********************************************************************************

1 0

wait a minute here
by Paul A Vixie 05 Sep '96

05 Sep '96

Someone just wrote me and said: > I would be happy to remove a lot of out /24's if the NIC would allocate a > larger than /32 to renumber into. The NIC does not thing this is as > important. According to what I know of InterNIC's policies, they will be glad to trade aggregated address space for unaggregated space, _in_roughly_equal_parts_, and assuming that you have efficiently used your old space. It's not that InterNIC doesn't consider this important -- rather than they are not in the routing business and the routability of addresses is not one of the criteria they can look at when making allocations. So don't put it to them in terms of routability, just SWIP your suballocat- ions and write a nice polite letter showing how some parts of the world will be better off and no part of the world will be worse off if prefixes X and Y are returned to the pool in exchange for prefix Z (of size ~X + ~Y). If you do _that_ and are told to go pound sand, I'd like to hear about it. But I know people who have successfully done what I'm suggesting so I really do not expect you to fail.

3 3

Hey, anyone who wanna help improve the net stability?
by Peter Lothberg 05 Sep '96

05 Sep '96

Hey all! We are approaching 40K routes, and this seems to make IBGP in networks with rich external connectivity pretty unhappy. Without pointing fingers, there are some 9K prefixes that could be removed by proper aggregation/reconfiguration. ICM-Stockholm-2#sh ip bgp summ BGP table version is 526856, main routing table version 526856 39321 network entries (300167/309954 paths) using 13491704 bytes of memory I suggest that everyon go look at what they can do, as we are all loosers with the currect situation. --Peter

10 12