Re: Reliability of RPC services

[Bas Wijnen]

On Sun, Apr 23, 2006 at 11:23:59AM +0200, Tom Bachmann wrote:
> Bas Wijnen wrote:
> > On L4.X2, we planned to have a task server doing several things. 
> > [...]
> > I think a similar system could be used to solve the problem in this case.
> 
> no no no. This leads directly to session based protocols we definitely
> want to avoid.

It all depends on how you look at it.  If you get the death-notification link
before every RPC and unlink it after it, then it is simply part of the RPC and
doesn't imply sessions.  If you really do have a session, then an obvious
optimization is to not unlink and relink it between RPCs.

Both this approach and the send-once capability can be used to implement both
session-based and sessionless protocols.  They can be compared for expected
performance in both cases, and then we can decide which is better to use.

(Of course there may be more options than just these two, but I'm ignoring
that here.)

Thanks,
Bas

-- 
I encourage people to send encrypted e-mail (see http://www.gnupg.org).
If you have problems reading my e-mail, use a better reader.
Please send the central message of e-mails as plain text
   in the message body, not as HTML and definitely not as MS Word.
Please do not use the MS Word format for attachments either.
For more information, see http://129.125.47.90/e-mail.html

signature.asc
Description: Digital signature