We are an ISP with ~6k subscribers. I recently set up Squid on a test box
to run for a day and measure how much bandwidth it saved, and surprisingly
it cut down our internet bandwidth by about 2 megabits (out of 9mb at peak
times).
That said, we want to implement this permenantly. The response times were
fairly slow on the test box. It was a celeron 600/256mb/20gig HD, but it
was also running a few other things which is probably the cause of the
slowness. It gets about 3100 requests per minute on average.
Is the following machine too much/too little for our purposes?
P4 - 1.5ghz, 512mb RAM, 120gb IDE HD. The 120gig HD will be for cache only,
with a smaller 20g drive for the OS.
Am I correct in assuming that the larger the cache storage is, the more
bandwidth it will potentially save?
Received on Mon Dec 02 2002 - 21:37:28 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:11:49 MST