Re: 回复:Re: what means Cache-off?

From: Joe Cooper <>
Date: Wed, 10 Apr 2002 02:50:35 -0500

It's in the FAQ under the section labelled "12.34 What does
--enable-heap-replacement do?" (the FAQ is well worth reading...covering
almost anything you could want to know about Squid): wrote:
> Thanks, Joe pal!
> Where can I find paper from the Hewlett Packard as you mentioned?
> Best regards,
> George, Ma
> ----- 原文 -----
> From: Joe Cooper
> To:
> Cc:
> Subject: Re: what means Cache-off?
> Sent: Wed Apr 10 14:08:36 CST 2002
>>I have no idea. Alex mentioned in his reply on this topic a couple of
>>options. And it wouldn't be difficult to script something in perl that
>>will read an access log and fetch every object fetched during the day.
>>Probably 5-10 minutes if you just want to test hit ratio (and don't care
>>about latency or throughput issues).
>>That said, Polygraph can provide a quite realistic workload without
>>relying on the quirks of the real internet--polymix-4 has no flaws as
>>far as I can see for realistically reproducing the workload a production
>>cache would see. Have you read the paper from the Hewlett Packard folks
>>who first implemented the heap-based policies? They've already done a
>>pretty solid bunch of research into the impact of the various policies.
>> How the policies perform is mostly a known quantity.
>> wrote:
>>>Thanks, Joe pal!
>>>Your reply has clarified my doubts. I still have a question. If I want to use
>>> historical log-files to test the performance of my cache, which do you think
>>> is the best tool for I to use? My aim to test the LRU, GDSF and LFUDA algorithms
>>> ( The hit rate of them.)
>>>Best regards,
>>>George, Ma
>>Joe Cooper <>
>>Web Caching Appliances and Support

Joe Cooper <>
Web Caching Appliances and Support
Received on Wed Apr 10 2002 - 01:53:17 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:15:06 MST