[Catalyst] Re: Hypothetical Site and Scalability Planning

[Robert Mah]

Mea Culpa.  I was confusing memcached with mmap.  I did not realize
the original writer was referring to danga's memcached, which is
not only distributed, but distribution is its focus!

Cheers,
Rob

-----Original Message-----
From: A. Pagaltzis [mailto:pagaltzis at gmx.de] 
Sent: Sunday, October 28, 2007 10:23 AM
To: catalyst at lists.scsys.co.uk
Subject: [Catalyst] Re: Hypothetical Site and Scalability Planning

* Robert Mah <rmah at pobox.com> [2007-10-28 15:00]:
> Memcached is not distributed.

What are you talking about?

> Thus, you can't support distributed session state with it.

Well, you can't put sessions in it because it doesn't guarantee
that things you put in it will stay in it (it's a cache, not a
persistence layer), but that has nothing to do with what you're
saying.

> Having only one server is obviously bad from a scalability and
> reliability point of view. 

Not sure, again, what you are talking about. In May this year
Facebook reported that their memcached installation comprises
about 200 16GB 4-core AMD64 servers:
<http://lists.danga.com/pipermail/memcached/2007-May/004098.html>

Regards,
-- 
Aristotle Pagaltzis // <http://plasmasturm.org/>

_______________________________________________
List: Catalyst at lists.scsys.co.uk
Listinfo: http://lists.scsys.co.uk/cgi-bin/mailman/listinfo/catalyst
Searchable archive: http://www.mail-archive.com/catalyst@lists.rawmode.org/
Dev site: http://dev.catalyst.perl.org/