Blogger : Benjamin Ms blog
All posts : All posts by Benjamin Ms blog
Category : WSCF/WCF
Blogged date : 2005 Jul 31
Kirk Allen Evans has an
interesting post about proposing a distributed architecture for a client where
he finds, after doing performance testing, that the speed difference between different
transports reduces as the amount of work done within each call increases. Based primarily
on this finding he's recommending ASMX web services as a better strategy
than using COM+/Enterprise Services for his client.
I remember coming across this arguement at TechEd 2004 in Amsterdam where Christian
Weyer spoke about some Thinktecture research
which showed that the difference in speed between Web Services and other distributed
technologies (remoting, enterprise services/COM+) was significant if the work done
in the distributed call was minimal, but once a reasonable amount of work was being
done the performance difference became less and less important. Worrying about
the speed of the transport without thinking about the amount of work done in each
call is a little bit like worrying how fast a car can rev if you have it in neutral.
Of course, this is all keeping with the long-known
guidance from the Indigo team, which is to start with ASMX and only go to other
options if you need specific features that they provide. Richard
Turner's recent white paper provides more details.