Re: Efficient way of global concurrency control/serializability in federated databases??

I can't tell if this discussion is still going on but, anyway, here it goes "the" replacement (it's actually a generalization): Paxos Commit http://research.microsoft.com/users/lamport/pubs/pubs.html#paxos-commit

On Oct 11, 11:24 am, "Jan Hidders" <hidd..._at_gmail.com> wrote:
> Marshall wrote:
> > On Oct 9, 1:40 pm, "Jan Hidders" <hidd..._at_gmail.com> wrote:
> > > anonym wrote:
> > > > Hi,
>
> > > > I am preparing an analysis report and need some help.
> > > > Currently, what is the most efficient / most used way of ensuring
> > > > global concurrency control / serializability in federated
> > > > databases/multidatabases ??http://en.wikipedia.org/wiki/Two-phase-commit_protocol
>
> > Word on the street is that the world is giving up on two phase commit.Well, I would say that is a bit of an exaggeration, in the same sense
> that "XML databases are going to replace relational databases" is a bit
> of an exaggeration. :-)
>
> > Although I'm not sure I agree, I can see some of the motivations.
> > The thing that bothers me, though, is that any potential replacements
> > are necessarily application-specific, and can't be handled in a
> > fully automatic way.It's not a replacement. We are facing new types of problems that
> require new types of solutions under new sets of assumptions such as
> that for certain operations there are other operations that "undo"
> them.
>
> > Any good papers to read about alternatives to 2PC?Hmm, sorry, not of the top of my head.
>
> -- Jan Hidders
Received on Sat Oct 14 2006 - 14:43:58 CEST