Re: [squid-users] Percentege cache

From: Eliezer Croitoru <eliezer_at_ec.hadorhabaac.com>
Date: Sun, 10 Apr 2011 02:04:44 +0300

On 09/04/2011 04:22, igor rocha wrote:

> I more thank Amos, but if anyone has any other tips, information, help me
adding to Amos,
the common ISP setup is 30-50 percent but it really depends on the
knowledge and usage of the ISP cache operators.
caching using helpers instead of the ordinary setup can lead to almost
like REVERSE proxy percentage.(80-90).

the main problem is that web developers or companies are not that into
building cache aware sites.
many dynamic content based websites using dynamic headers for a static
content from unknown reasons.
i do understand some of youtube and other sites concerns about their
content but the other side is that
they do not want to even start thinking about considering helping ISP's
and companies to work it around.
(at least not on the table)

i have my own proxy that serves SOHO environment with almost specific
usage that has a hit ratio of about 80%.
(the refresh patterns are costumed and im using some helpers i have built).

Eliezer

> 2011/4/8 Amos Jeffries<squid3_at_treenet.co.nz>:
>> On 09/04/11 02:49, igor rocha wrote:
>>> ?
>>>
>>> 2011/4/8 igor rocha<igorlogos_at_gmail.com>:
>>>> Hello,
>>>> I know that does not formulate the right question, I am Brazilian and
>>>> not mastered English well, but talk about commonly used metrics to
>>>> measure the effectiveness of the cache and the amount of bandwidth
>>>> saved is hit ratio, defined as the percentage of requests that are
>>>> satisfied by the proxy as cache hits.Show me the average percentage of
>>>> index pages that are not cached,a paper.
>>>>
>>>> understand?
>> Ah. I think so.
>>
>> I'm not aware of any papers on that. It is highly variable between networks.
>> We do have two general "rule-of-thumbs";
>> * that reverse-proxy (CDN) see hit ratios usually around 80%-99% for a
>> website.
>> * that forward-proxy (ISP) see hit rations between 25% and 45%.
>>
>> with variance outside of those ranges for older Squid versions and poorly
>> written websites.
>>
>> This is general-knowledge built up from years of small talks with people
>> looking at and discussion of their cache ratios. Nothing published exactly.
>>
>> We have in recent years attempted to collects statistics on these. Which can
>> be found at http://wiki.squid-cache.org/KnowledgeBase/Benchmarks along with
>> the methodology used for collection.
>>
>> Amos
>>
>>>> 2011/4/8 Amos Jeffries wrote:
>>>>> On 08/04/11 02:48, igor rocha wrote:
>>>>>> Hello Gentlemen,
>>>>>> could anyone tell me what percentage of sites that are required for
>>>>>> the cache and actually go into the cache, can be an article that
>>>>>> talksabout it, something that helps me to have concrete statistical
>>>>>> data.
>>>>> Please explain your meaning of "required for the cache".
>>>>>
>>>>> 100% of squid cacheable sites get cached. Admin often force>100% to be
>>>>> cached.
>>>>>
>>>>> I suspect you mean something else though.
>>>>>
>>>>> Amos
>>>>> --
>>>>> Please be using
>>>>> Current Stable Squid 2.7.STABLE9 or 3.1.12
>>>>> Beta testers wanted for 3.2.0.6
>>>>>
>>
>> --
>> Please be using
>> Current Stable Squid 2.7.STABLE9 or 3.1.12
>> Beta testers wanted for 3.2.0.6
>>
Received on Sat Apr 09 2011 - 23:39:28 MDT

This archive was generated by hypermail 2.2.0 : Sun Apr 10 2011 - 12:00:02 MDT