Re: Starting to see the light and IOR through webservers

[Michael Meeks <michael ximian com>]

Hi Peter,

On Tue, 2003-04-29 at 18:19, Peter Van Osta wrote:
> Thanks again to you all, form what I read about transferring images I
> will use "cross-over" client/servers, where the server can send iamge
> data to the "client" by a small server running inside the client that
> accepts the images "pushed" from the server during a process.

	That sounds ideal :-) I'm glad it's coming together; don't stop asking
questions though :-)

> An idea came up while thinking about the availability of the IOR in a
> network. When running servers and clients on different machines in a
> network, the servers could send their IOR string to a webserver (e.g.
> Apache), running a small database (e.g. MySQL) where the IOR strings
> could be stored. A client could ask the webserver for an IOR and
> connect.

	Right; there are several suggested solutions for this; the problem
being that you will always need a 'first IOR'; and that has to be
bootstrapped somehow outside of the CORBA framework - just a problem
that we can't solve generally.

> I intend to run a pool of automated microscopy severs, each with a
> machine attached to it (microscope reader, robot arm, ...). Beside the
> machines other servers will be available to do large volume analysis.
> Providing a central webserver where the IORs are registered would
> facilitate the managment of this ORBit server pool. Is this feasible or
> is there a better alternative for this ?

	Well; I'd tend to go for having a single IOR (the server) to register
with; and if the model is more complex - it's prolly worth using the
Naming service (although it's a total pain to use ;-). We provide a
sketchy impl. of the naming service inside ORBit2; you still have to get
an IOR to the naming service though - so perhaps it doesn't win you
much.

> By the way our robot arm is being controlled through an ORBit link to
> the main software and works fine (not all worries and trouble coming
> from here ;-) ). We even automated a Phytron stepper motor by using
> ORbit.

	Nice :-) I'm sure James will be pleased to know his work is useful in
that context.

	Regards,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot