Problems with US-UK Internet Connection

From
Date
Reply-To
Owing to the messages that appeared on PaleoNet in the last weeks
expressing frustration at the time it takes for US subscribers to access
the PaleoNet Pages, I though I might pass this recent message from W. Black
(copied below) along. My experience has been that the whole Internet has
become much slower and more prone to crashes over the last 12 months, and
especially since the summer. No doubt this is due to its growth. I've been
talking to Jere Lipps about getting the PaleoNet Pages mirrored at Berkeley
and I'm hoping this can come about in the near term. Until then, hang in
there. It looks like the InterNet and PaleoNet are, in some senses, victims
of their own success.

Norm MacLeod



*********************
Dear colleagues,

I am sure that many (perhaps most) of you are aware that the US Internet as
seen from JANET has become increasingly difficult to use. This note is to
assure you that we in UKERNA are aware of the problem and are considering
all possible ways to fix it.

Clearly there is insufficient capacity to meet the current user requirements
and that is being addressed - we are negotiating a further 8.5 Mbit/s of
capacity, and looking at further (diversified) connectivity. I cannot give
you a definite timescale for this, but we hope to reach agreement with our
suppliers within the next week, and should be in a position to make a
definite announcement on delivery then.

But this is not the only problem. The transatlantic fibre, PTAT, has been
broken on a number of occasions in the last 18 months. Although we get
restoration by other routes, a certain amount of disruption occurs while
this is being implemented. In addition, over the last couple of weeks, the
Sprint network in the USA, to which our transatlantic connections are made,
has also been unreliable, with a number of router failures and software
problems. At the heart of this latter problem is the difficulty routers are
experiencing in dealing with the absolute number of networks now on the
Internet. Any outage of a component on the main part of the net is leading
to serious overloading of CPUs on routers, which is turn may then fall over
and produce a flood of problems which can spread out over the entire net.
Rectification of this problem is completely outside our control, and rests
with the major networks in the USA and their router suppliers. It is
difficult to foresee how long this situation may continue.

In the meantime, we are looking at other options, but the explosive growth
in Internet usage may well be taking us towads the limits of what the
current configuration can cope with.

Regards,

Willie Black
Programme Director
UKERNA




----------------------------------------------------------------------------
Norman MacLeod
Senior Scientific Officer
N.MacLeod@nhm.ac.uk (Internet)
N.MacLeod@uk.ac.nhm (Janet)

Address: Dept. of Palaeontology, The Natural History Museum,
         Cromwell Road, London, SW7 5BD

Office Phone: 0171-938-9006
Dept. FAX:  0171-938-9277
----------------------------------------------------------------------------