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@xxxxxxxxxxxxx>:
On 09/04/11 02:49, igor rocha wrote:
?
2011/4/8 igor rocha<igorlogos@xxxxxxxxx>:
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
