[omniORB] Small tweak for local object detection

bjorn rohde jensen bjensen@fastmail.fm
Thu Aug 22 14:19:01 2002

Hi guys,

  I think, it would be a good idea to add a
mechanism to specify, which ip-addresses
the local call optimisation is to be performed
on. Sort of a new transport rule, i guess. I am
especially thinking of the whole security and
interceptor thing here. The local call optimisation
does cut interceptors out of the loop, right?
  The whole idea of placing special active address
in object references sounds pretty strange to me.
IIOP, which i imagine, you are using, is build on
tcp/ip, and tcp is a 1-1 connection oriented
protocol by design and so forth. Are you doing
some sort of load balancing by means of a special
"cluster address" in a windows cluster or something?
There are several significantly better CORBA design
patterns for this sort of thing.

Yours sincerely,


> On Tuesday 13 August, Chris Newbold wrote:
>> We ran into a limitation of the algorithm used by omniORB4 in
>> giopRope.cc for determining whether or not an object reference
>> identifies a local object.
> [...]
>> The problem with the local object detection algorithm is that it
>> declares success after matching only one address in an incoming IOR.
>> Since all our IORs contain the special "active" address, omniORB always
>> thinks these objects are local.
> Unfortunately, there are some situations in which you do want to match
> on only one address. One example is where you issue a persistent
> object reference, quit, then come back with a different set of
> endpoints, but still including one you had before. In this case, you
> would still want to consider the old reference local.
> I think the proper solution to this issue is to have some sort of
> address matching mask that allows you to specify which interfaces have
> to match for a reference to be considered local. It's too late to do
> that for omniORB 4.0.0, though, unless someone submits an
> implementation in the next week or so. For now, it would be easy to
> add an option to either match one or match all.
> Anyone have any thoughts?
> Cheers,
> Duncan.
> -- -- Duncan Grisby -- -- duncan@grisby.org -- -- http://www.grisby.org -- _______________________________________________ omniORB-list mailing list omniORB-list@realvnc.com http://www.realvnc.com/mailman/listinfo/omniorb-list