Dedicated Servers + Scalable Web Architectures

[Christopher Browne]

On Mon, Jun 23, 2008 at 3:51 PM, Madison Kelly <linux-5ZoueyuiTZhBDgjK7y7TUQ at public.gmane.org> wrote:
> Lennart Sorensen wrote:
>>
>> Does anyone actually know what Web 2.0 is besides a silly marketing
>> term?
>
> Glossy icons and Iframes and javascript layering, as far as I can gather.
>
> So no, just a marketing term.

There's supposed to be something to it that's about mashing together
services that come from different places.

Wikipedia isn't necessarily "definitively correct," but it frequently
is successful at recognizing what is the commonly-perceived
understanding of things, and I think it gets the "commonly-perceived
understanding" right...
http://en.wikipedia.org/wiki/Web_2.0

Of course, that "understanding" seems to mostly consist of "silly
marketing stuff" ;-).

>> Well at least in my expeirence, MySQL doesn't scale.  Unless you almost
>> exclusively read from it it doesn't like lots of users.  Postgresql is
>> much much better that way.  I really have no idea why people always seem
>> to go for mysql first rather than looking into what is available and
>> picking the better choice.
>
> Seconding the recommendation for PostgreSQL over MySQL.
>
> MySQL is fast, that is why it is often chosen. That speed comes at some risk
> though. PostgreSQL is not perfect, but crashing it is less likely to result
> in DB data corruption. I don't mean to get this into a MySQL vs. PgSQL
> argument, others may (will) differ, just take my opinion for what it is; one
> small vote for reliability based on my anecdotal experience.

Here's a fairly fair characterization of things...
http://wiki.postgresql.org/wiki/Why_PostgreSQL_Instead_of_MySQL:_Comparing_Reliability_and_Speed_in_2007

The sorts of cases where MySQL(tm) is regarded as "lots faster" have a
lot to do with:
a) Looking at elderly versions of both MySQL(tm) and PostgreSQL.  Yes,
PostgreSQL 6.5 was dog slow, but that was 1997.

The tendancy has been that PostgreSQL has gotten both faster and more
scalable (e.g. - it plays better under light load than it used to, and
you can also get benefit from scaling up to a 16 CPU server with a
disk array); in contrast, the enhancements to MySQL(tm) haven't tended
to make it faster.  The reliable storage engines tend to leak
performance.

b) Looking at cases that do not involve scaling in any sense.  The
classic "MySQL(tm) is faster" benchmarks tend to involve running a
whole lot of queries into a single connection, which, for MyISAM, is
the only "fast" configuration.

Bump it up to 10 concurrent connections and MyISAM tends to bog down
because it uses table locking, which amounts to saying "doesn't scale
to past 1 user."

You can migrate to other storage engines, but the paucity of published
benchmarks for such cases is a (largely silent!) witness to the fact
that this pretty much leads to throwing away all of the performance
benefits you thought you were getting.  For more realistic benchmarks,
there is no indication that there is any substantive performance
difference, and as you add connections, PostgreSQL pulls consistently
ahead...
-- 
http://linuxfinances.info/info/linuxdistributions.html
"The definition of insanity is doing the same thing over and over and
expecting different results." -- assortedly attributed to Albert
Einstein, Benjamin Franklin, Rita Mae Brown, and Rudyard Kipling
--
The Toronto Linux Users Group.      Meetings: http://gtalug.org/
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists