|PPNCG Information Pages : STOP PRESS! and Contacts : London area problem in summer 2000 - details|
|Chronic problem(s) in the London area during working daytimes...
This has been ongoing for quite a number of weeks now. A report was sent to JANET operations on 24 July, at which time we were told that network staff were aware of a couple of points of congestion in the London area and that upgrades were on order. However, the symptoms continued, and in the last three weeks or so were getting worse.
Based on observations on the PPNCG's network plots, and a number of complaints from users, and ad hoc investigations by the PPNCG and by staff at RAL, we could see:
 there was serious packet loss on one of the routes between JANET and
the London MAN network. This affected a subset of routes to/from
London sites from/to non-London sites.
On connection paths that traverse both hazards (e.g RAL to IC Physics), the effect was particularly horrendous. By comparison, the plots from DL to IC, although unpleasant, were noticeably less bad, as they seemingly missed the first hazard.
RAL's networking contact reported the problem on behalf of their London-based users, and it was decided that we would hold-off from placing a further report on behalf of the PPNCG, pending some response to RAL's report.
There has now been a quick and positive response, to the effect that a fibre problem has been fixed and also a routing problem corrected. Throughout the working day of Friday 1st Sept the previously-affected link has now recorded clean results, so this looks like an efficacious solution to the problem.
Looking back over the plots that cover the period of this problem, I have to express astonishment that users were willing to accept this very unsatisfactory level of network service without making more vociferous complaints. This does seem like the right moment to make the point that the PPNCG works best when it responds to user needs. There is no point in reporting a poor quality network link merely because it is there, if none of our users are relying on it. So I'd like to make the point again that users who are experiencing significant problems with their research as a result of network problems are invited to raise the matter via the hepgripe address, as this helps the committee to make representations that are in accordance with user requirements.
As for the second point, apparently the external connection which serves IC Physics is due to be upgraded shortly; no immediate relief is expected with regard to that specific problem, we need to await the upgrade.
|[Append to This Answer]|
|Previous:||Bulk transfer rates from ESnet (article of June 14th)|
|Next:||RAL-CERN Line closed down|