Tom Collins wrote:
> On Sep 20, 2007, at 12:56 PM, Rick Widmer wrote:
>> If I remember right, speed was the reason for separate tables, but
>> testing showed it was not faster.  I think the single table works
>> better because all your mail users are accessing the same table, and
>> its indexes so they stay loaded all the time.  If you use separate
>> tables it is always thrashing the cache as different files need to be
>> accessed.
> 
> If we were really looking for speed, we could move to a "ng" (next
> generation) table format that was more relational.  Provide tools to
> migrate from the old to the new for those people who only access the
> data through vpopmail's APIs.  Continue to support the old method for
> people who have home-grown apps that access the data.
> 
> The domains should be in a table of their own, and the users table
> should index the domains table.  Having an index in the users table on
> an int (and the resulting size savings) would be measurable.
> 
> The biggest change would be updating the selects and inserts in the
> code.  Not a huge change -- just a join between the tables.
> 
> Something to consider, and perhaps discuss further (on this list or
> vpopmail-devel).

I understand what you're proposing, but I would suggest that it would
add complexity for little gain. Of course, that would need benchmarking
to establish which is the faster method.

If someone can give me a large dataset, I'm happy to crunch some numbers.

R.

Reply via email to