- From: <hallam@zorch.w3.org>
- Date: Thu, 11 Apr 96 14:40:54 -0400
- To: "James E. Calloway" <jcallowa@nando.net>, www-logging@w3.org
- Cc: hallam@zorch.w3.org
>> 2. If not all servers are running exactly the same time it can be >> misleading to log seconds if we start to combine logs from different servers. >> Also, what time is logged? The beginning or the termination of the >> request? It might be wise to have both times. >That shouldn't be a problem in this day and age when even desktop PCs can >synch up to SNTP servers. Sys Admins should be synching their servers as >a matter of course. Global synchronization is IMHO much less important at the second by second level. In any case we run synchronisation within 2 seconds across the Atlantic so its really not that big a deal. Analysis of results on a single server at the second by second level would be very interesting and worthwhile IMHO. Knowing that one got 20 requests a second for 15 seconds is quite interesting. Knowing that one got 300 in a minute less so. When we tunned the CERN server we looked at the number of connections per second that it peaked at then looked at the bottlenecks. Phill.
Received on Thursday, 11 April 1996 14:40:56 UTC