Re: Informix vs. Sybase vs. Oracle vs. (gasp) MS SQL Server

David Williams wrote:
>
> In article <yuten43d90u.fsf_at_mew.corp.sgi.com>, Pablo Sanchez
> <pablo_at_sgi.com> writes
> >>>>>> "Greg" == Greg <greg_at_cyberramp.net> writes:
> >Greg>
> >Greg>
> >Greg> Same reason practically noone writes assembler anymore. Works great if
> >Greg> you futz with it enough. No one can afford to futz with it enough
> >Greg> though. It's just not economic to do so. Unfortunately for Sybase,
> >Greg> they're learning this the hard way.
> >Greg>
> >
> >... let's make it easy... why is Sybase's TPC-C's higher
> >than Informix? They're both OLTP? As I said, row-level vs
> >page-level is hype.
>
> page level locks give HIGHER PERFORMANCE yes BItUT the reduce
> concurrency and hence LOWER RELIABILITY.

Be careful with your wording...locking granularity has nothing at all to do with reliability (indeed, table locking vs. column locking has the same reliability when properly implemented). It is all simply a matter of concurrency and efficiency.

> It's a tradeoff and
> most of the time people go for reliability over a little more
> performance. 90% of performance tuning is
>
> a) reducing disk I/O by only selecting/updating the minimum
> amount of data.
> b) database schema / index usage.

Of course, it could be argued that b) is a subset of a). But I digress :)

-scott

-- 
Scott C. Gray                 gray_at_voicenet.com     "my keybard is
brken"
Sybase Professional Services  scott.gray_at_sybase.com
   http://www.voicenet.com/~gray/sqsh.html