Trial Balloon: A new DataStore Architecture?

Rob Browning
31 Oct 2000 10:54:49 -0600

Derek Atkins <warlord@MIT.EDU> writes:

> I personally dislike CORBA.  My reasoning is two-fold:

I don't actually know enough about CORBA to dislike it :> but I've
messed with RPC systems enough to sympathize strongly with your

I mostly just brought it up as another one of those things we should
at least evaluate.

> Another potential problem is the security of CORBA.. Namely, there
> is none. :) I would personally insist on data encryption and strong
> (kerberos-level or greater) user authentication.

Don't worry, I'm sure there will be plenty of people, me at the top of
the list, jumping up and down and pulling the fire alarms if anyone
tries to get something lame in on this front.

> Frankly, I don't think that designing a real protocol would be
> difficult.  I also don't think it needs to happen right away.  I
> think we can architect and implement the data model using local
> storage before designing the network protocol.  At least a set of
> access requirements should come first.

Yes, and I'd imagine how hard the design of the protocol would be
dependsin in great measure on how much you want out of it (failure
semantics, etc.).

> > (Fingers tired... stopping typing now :>)
> I'd like to continue this discussion...  When your fingers relax a bit
> :)

OK.  I need to stop with the mail for a while and get back to some
actual coding, so I can finish up some things and move on to other
things like this :>

But I think we should keep talking about this, so badger me in a bit
if I trail off.  I also think, as I've mentioned, that it should be
considered in parallel with deciding what we should do about the
top-level "dataset" in the engine, and how that relates to the
existing Session object.  Some of my recent things make minor
improvements, but we need much more than that.

If you get a chance, you might want to look at Session.h and

Rob Browning <> PGP=E80E0D04F521A094 532B97F5D64E3930