[Cosmo-dev] Re: [service-dev] Deployment of Cosmo persistence load distribution

[Lisa Dusseault]

On Jul 18, 2006, at 8:08 AM, Brian Moseley wrote:

> i think i might have been a little pessimistic. the issue is one of
> concurrent writes to a resource or collection, and how we manage that
> situation. the more i think about it, the less i see need for state in
> cosmo to deal with it. we probably just need to implement locking and
> encourage our clients to use it.


It's probably hopeless to count on clients to implement locking.   
It's an optional part of CalDAV and WebDAV and many clients just  
don't or won't bother.  So how do you manage concurrent writes when  
the client doesn't implement an optional feature?  One way is for the  
server to implement a write request as an implicit LOCK -> [write] ->  
UNLOCK, waiting if necessary for the LOCK to succeed before  
proceeding with the other requests and waiting for the whole thing to  
work before sending the successful reply.

Lisa
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osafoundation.org/pipermail/cosmo-dev/attachments/20060720/c0f730ae/attachment.htm