dce/rpc "client" api

Luke Kenneth Casson Leighton lkcl at samba.org
Wed Aug 23 06:47:16 GMT 2000

On Tue, 22 Aug 2000, Gerald Carter wrote:

> Luke Kenneth Casson Leighton wrote:
> > 
> > by the same token, using one smbd process for *all* smbd 
> > code is equally not scalable.
> Nah...Big yeah...a lot to understand....yeah...but 
> not scalable...I don't think so.
> Jeremy's comments were about load and performance.  

i do not think so.  i recall the set of assumptions that i managed to find
out he had about the daemon architecture, and he assumed that it was
single-process, multiple sockets.  it is not: like smbd, it is
select/fork, multiple-process, single-socket-per-client-process.

> You're now trying to equate KLOC with number of clients.  Those are
> different things.


what is KLOC?

> ok...I remember this thread and I remember the fact that
> dce/rpc sessions are generally short lived

for this particular reason, dce/rpc is generally inefficient, however that
is actually the fault of the developers writing the client-app and
sometimes the fault of the client dce/rpc transport implementation.

or, more specifically, if the app-developers are not aware of the network
traffic their functions generate, their app may be either slow or
massively inefficient.

More information about the samba-technical mailing list