Re: cost per transaction

jonah(dot)harris(at)gmail(dot)com ("Jonah H. Harris") writes:

> On 4/20/07, Federico <rotellaro(at)gmail(dot)com> wrote:
>> I'm writing the slides for the road2pgday and I'm wondering if there's
>> some test about the cost per transaction that have PostgreSQL on OLTP.
>
> AFAIK, there aren't any audited public benchmarks for PostgreSQL. So,
> you're not going to find a TPC-C or TPC-E for PostgreSQL yet. There
> may have been one of the SPEC benchmarks that used PostgreSQL, but
> that wouldn't have been direct OLTP.
>
> Being free, PostgreSQL's cost-per-transaction would be fairly low.
> Basically, if you purchase maintenance, support, installation, or
> tuning services from a PostgreSQL support organization or consultant,
> then you're cost/transaction is really based on that fee and the
> number of your applications transactions run over the lifetime of the
> support contract.
>
> Unlike other database vendors, there's no cost/transaction for free PostgreSQL.

The last time I looked at TP$ benchmarks, they counted the cost of
both software and hardware when accounting for costs.

The fact that PostgreSQL may have no licensing fees might imply that
the *software* cost per transaction is $0. But the hardware still has
some cost, so you will have to count that...
--
(reverse (concatenate 'string "ofni.sesabatadxunil" "@" "enworbbc"))
http://www3.sympatico.ca/cbbrowne/finances.html
"Using these toolkits is like trying to make a bookshelf out of mashed
potatoes." -- Jamie Zawinski