Since PSI still hosts one of the root servers (C.PSI.NET) C.ROOT-SERVERS.NET,
I hope C&W's customers understand the all ramifications of C&W's actions.
Depending on the current state of the net, it may vary from a minor reduction
in access to all possible servers to 1/13.
Unless, of course, CW or PSI has transit connectivity from someone. Can
someone on either network do a BGP lookup and see if they're seeing each
other's routes, and if so, through whom?
-C
Once upon a time, Christopher A. Woodfield <rekoil@semihuman.com> said:
Unless, of course, CW or PSI has transit connectivity from someone. Can
someone on either network do a BGP lookup and see if they're seeing each
other's routes, and if so, through whom?
We're connected to C&W, AT&T, and Sprint:
hsvcore#s ip b 38.0.0.0
BGP routing table entry for 38.0.0.0/8, version 402730
Paths: (2 available, best #2)
Advertised to non peer-group peers:
216.180.95.209 216.180.116.1 216.180.119.54
1239 174, (aggregated by 174 38.1.10.8)
216.180.119.54 from 216.180.119.54 (216.180.119.54)
Origin IGP, metric 8, localpref 100, valid, internal, atomic-aggregate
7018 174, (aggregated by 174 38.1.3.39), (received & used)
12.125.208.5 from 12.125.208.5 (12.123.21.247)
Origin IGP, localpref 100, valid, external, atomic-aggregate, best
No route to PSI through C&W.
mike@somewhere$ traceroute www.cw.net
traceroute: unknown host www.cw.net
Trying to trace out by IP address gives !H on the first router with a full
table.
Oops. Now it gets interesting...
Mike
route-views.oregon-ix.net concurs:
>
>route-views.oregon-ix.net>sh ip bgp regex ^3561_174_
>
>route-views.oregon-ix.net>
Yes, but if one or the other has backup peering, it would not look like that. It would look like _3561_.*_174_ or _174_.*_3561_ - prolly the former since AS3561 gives route-views a feed, but AS714 does not.
Looking in route-views for those two patterns, I see only a few routes under 3561.*_174_, probably leakage. There are no routes of the form _174_.*_3561_.
Since route-views does have a feed from AS3561, I would say it is official. Cable and Wireless cannot reach PSI.net.
Congratulations ladies & gentlemen. The first intentional, prolonged, significant (for some values of "significant" 
outage on the Internet. And we were all here to see it....
Wow, in the same week MAE-East dies. Sad time for the 'Net. :((
It can get much more interesting if PSI/Sprint peering was to go down, since
at least one provider (whom I won't name, but most people probably can guess
who I'm talking about) with some large datacenters (and big customers in
those datacenters) uses Sprint to reach PSI, probably ever since PSI
attempted to charge the provider with the large datacenters for peering with
them. (Another similar large datacenter provider uses Verio to reach PSI, I
believe)
I suppose now PSI gets to learn the hard way what happens when they scared
half their peers away (to be polite...), and now find that a bunch of the
other half are now turning down their PSI peering links. (BTW, has it been
established here whether PSI or CW is to blame for this?)
I really really hope that no one still resells PSI dialups, otherwise their
tech support lines could be very very busy very soon.
Oh, and FYI, a friend also in the UK using PSI reports the same thing you're
reporting... no trace of the CW network from there.
Vivien
Date: Mon, 04 Jun 2001 18:16:32 -0400
From: Patrick W. Gilmore <patrick@ianai.net>
[ snip ]
Congratulations ladies & gentlemen. The first intentional, prolonged,
significant (for some values of "significant"
Internet. And we were all here to see it....
Wow, in the same week MAE-East dies. Sad time for the 'Net. :((
Hey, Boardwatch's yearly "speed rating" is over. Another year before
anybody has to worry. *rolls eyes*
I wonder if 3561/174 are significantly perturbed on CAIDA or MIQ...
Eddy
Yes, but if one or the other has backup peering, it would not look like
that. It would look like _3561_.*_174_ or _174_.*_3561_ - prolly the
former since AS3561 gives route-views a feed, but AS714 does not.Looking in route-views for those two patterns, I see only a few routes
under 3561.*_174_, probably leakage. There are no routes of the form
_174_.*_3561_.
Ah ... me and my half-assed regular expressions. Thanks for the clean-up
work, Patrick.
The most recent showdown I can relate this to is Time Warner turning off
ABC over contract disputes. The odd thing about that was that ABC seemed
to get more of the public opinion blame than TW did ... is it possible
that PSI's distressed financial state will affect public perception of who
is at fault?
Either way, I wouldn't expect CW to turn anything back on until they get
enough customer complaints that they can't hold out any longer ... this is
certainly not the cleanest way to settle a contract dispute, and should
serve as a stark warning to anyone considering CW for transit.
-travis
this is certainly not the cleanest way to settle a contract dispute,
but it is a damn good way to invite regulation. "The stability of the
Internet is a national concern" will be the sound-bite.
> this is certainly not the cleanest way to settle a contract dispute,
but it is a damn good way to invite regulation. "The stability of the
Internet is a national concern" will be the sound-bite.
Yuck. I was interviewed by the GAO a few months back (they wanted to talk
to little players about the transit market) and was worrying that the feds
wanted to mandate interconnection policies in one form or another ... we
certainly don't want to encourage that kind of behavior. However, it
seems reasonable that if we can't regulate ourselves someone else is going
to do it for us.
-travis
Yuck. I was interviewed by the GAO a few months back (they wanted to talk
to little players about the transit market) and was worrying that the feds
wanted to mandate interconnection policies in one form or another ... we
certainly don't want to encourage that kind of behavior. However, it
seems reasonable that if we can't regulate ourselves someone else is going
to do it for us.
It seems that everyone has fogotton what the "Internet" is. The Internet is
not IP, the network protocol could change and it would still be the
Internet. The Internet is not the providers, the providers could change and
it would still be the Internet.
The Internet is a spirit and a philosophy. That spirit and philosophy is of
making a good faith effort to obtain connectivity and exchange information
with anybody else who makes a similar effort. Anyone who claims to provide
'Internet access' or 'Internet service' or to be an 'Internet' product or
service without practicing that philosophy is, in my opinion, practicing
fraud.
This applies to software, hardware, and even peering. A program is
"Internet software" if it makes a good faith effort to exchange information
with anybody else who makes a similar effort, not if it happens to work over
the machines and protocols that happen to constitute the Internet today.
DS
See PSINet's network status page (http://www.psi.com/cgi-bin/netstatus.pl5) for a possible answer.
jas
THE UNITED STATES and Offnet Connectivity
Cable & Wireless chose to terminate connectivity with
PSINet on 2June01. Over 90% of the traffic that used
to be routed through C&W is now being routed via other means
through our robust global free peering infrastructure.
The remaining 10% or so represents C&W customers that have
been deliberately cut off from PSINet by C&W. While PSINet
is ready and willing to re-establish connectivity with
C&W at any time, it is up to C&W to choose to reverse their
previous decision. In the meantime, PSINet can offer
services directly to those C&W customers that are affected.
OK, PSI seems to assert that 90% of C&W networks are still accessable from
PSI customers. NANOG research so far has determined that this is
definitely *not* the case. If anyone has evidence to support PSI's claim,
please post.
Dear PSI: this may not be directly your fault, but dammit, own up to the
scope of the issue. It's in your interest to take advantage of being the
"good guy" for once, so don't ruin it by lying about the scope of the
problem.
I don't think that this is going to be solved by C&W reversing themselves;
I think PSI is going to have to get itself some transit, and quickly.
-C
Is this the part where the people (eg: Exodus, AboveNet are the two I can
think of immediately) who were forced to get themselves some transit because
PSI wouldn't peer with them anymore go and laugh at the irony of C&W pulling
a PSI on PSI themselves?
Vivien
Uh.
Why are you yelling at PSI when you have failed to do your
own calculations? Perhaps they have taken data from
archive.route-views.org to determine what the actual loss of
connectivity was.
I don't have the time to go out and validate the PSINet claims
of how much of the net is gone for them and their (single-homed)
customers.
Perhaps someone who is more of a data processing person can go
out and provide some interesting data, such as
Number of ASNs single-homed (based on route-views data)
Top 5/10/20/25 providers based on as-path
Number of networks/ips/ASNs behind each of those top 5/10/20/25 that
would be missing connectivity.
- Jared
Is this the part where the people (eg: Exodus, AboveNet are the two I can
think of immediately) who were forced to get themselves some transit because
PSI wouldn't peer with them anymore go and laugh at the irony of C&W pulling
a PSI on PSI themselves?
I don't know about Abovenet, but when things when down between Exodus and
PSI, my impression was that Exodus just got Sprint to carry the traffic.
No new circuits, just a new path, and not a big deal because it was small
amount of traffic (rumored to be <90Mb).
Eric
I know almost nothing about bankruptcy, so I would love for someone
who does to comment. Put simply, if a company can't pay its
creditors and suppliers do they have any chance of entering into
a new purchase agreement? I know that I personally wouldn't offer
them Net-30 payment terms right now.
AboveNet did the same deal with Verio instead of Sprint... and they're still
doing it, too, just like Exodus is still using Sprint to reach PSI.
I don't think the issue here is one of circuits or anything, it's more one
of embarassment. AboveNet, PSI, and probably C&W (I'm not sure about
Exodus... ironically enough, it's the only one of these that I use) all
claim that they're "tier 1" networks. However, AboveNet has been forced to
get Verio to provide transit to PSI because of this. That, technically,
means that AboveNet is not a tier 1 by my definition (according to me, and
probably most people on this list, a tier 1 is someone who has no transit
from anyone). Now, PSI, which used to call itself "the Internet
supercarrier" IIRC (ironically, until a year or two, maybe three, ago, also
claimed their DS3 frame relay network was state of the art), may be forced
to get someone to transit the 2.5 megabits (or is my guess too high?) of
traffic to CW.
It's not likely to be a big technical deal, but the irony I find to be quite
prominent. First PSI forced others to make transit arrangements because of
their greed, and now CW is possibly making PSI do the same, for probably the
same motives.
Vivien