Re: alternate parent selection algorithms

Alex Rousskov writes:
>On Wed, 28 Oct 1998, Niilo Neuvo wrote:
>
>> As a background: I maintain a cache-server farm consisting of 4
>> machines at the moment. The overhead of sending ICP-queries back and
>> forth is stopping me from expanding the cache with more machines. We
>> deliver about 10 million queries a day. This means 7.5M*3 (22.5M) ICP
>> queries per day received by a single cache. With 8 machines and
>> double queries it would mean 17.5M*8 (140M) ICP queries, meaning about
>> 4000 ICP queries per second during peak hours. At some point this
>> system will no longer work.
>
>Use Cache Digests to reduce/eliminate ICP queries.

The other point is that I want to move the process of selecting the
cache that actually contains (or should contain) the object to the
client software.

I have a small javascript code that does this by calculating a hash
code from every url and decides what parent to use. It is vital that
the javascript uses the same algorithm as the parent cache uses.

Old customers use the parent cache and all new customers get the
javascript version of the client software. Old customers can
naturally reconfigure their browsers, but that is a very long process
and not all customers will ever get their browser configured.

-- 
     Niilo Neuvo                   >>>^<<<       niilo.neuvo@saunalahti.fi
     Technical Director              /¤\      +358 (0) 50 5611042 (mobile)
     SAUNALAHDEN SERVERI OY          | |         +358 (0) 3 31558042 (fax)
     .¸¸.·´´¯·.¸¸.·´´¯·.¸¸.·´´¯·.¸¸.oOOOo.¸¸.·´´¯·.¸¸.·´´¯·.¸¸.·´´¯·.¸¸.·´