home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.isis
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!ames!agate!spool.mu.edu!caen!batcomputer!cornell!ken
- From: ken@cs.cornell.edu (Ken Birman)
- Subject: Re: Groups in a mixed media setting
- Message-ID: <1993Jan26.214328.14986@cs.cornell.edu>
- Organization: Cornell Univ. CS Dept, Ithaca NY 14853
- References: <1993Jan26.165131.28041@comp.lancs.ac.uk>
- Date: Tue, 26 Jan 1993 21:43:28 GMT
- Lines: 64
-
- In article <1993Jan26.165131.28041@comp.lancs.ac.uk> nigel@comp.lancs.ac.uk (Nigel Davies) writes:
- >...
- >1) I seem to remember that there was some talk of collaboration between APM
- > (the developers of ANSAware) and Cornell a year or so ago. Did this ever
- > come to anything ? (The most recent version of ANSAware still does not
- > provide support for groups so I am guessing not ...)
-
- When ISIS and APM became commercial, they sort of went different directions.
- We are now working on extending ISIS into a Corba-complient environment,
- which we call RDOM, but have not been working directly with ANSA/APM, although
- the ASNA ideas certainly have had a huge impact on our thinking.
-
- >2) It is some time since I have seen any ISIS documentation. Does ISIS support
- > hierarchies of groups, and if so what sort of event ordering guarantees
- > does it provide ?
-
- Yes; I'll ask Robert or Brad to answer on this because they did the work.
-
- >3) One of the reasons for using the group abstraction for describing multicast
- > /multidrop continuous media configurations was to provide the programmer
- > with similar tools for carrying out both invocation and continuous media
- > transmission. However, judging by the recent postings there may not be a
- > requirement for multidrop invocation groups.
-
- We are more and more interested in this. Right now, you do this (you can do
- this) by binding each of a set of groups to different transport protocols.
- For example, the ipmcast code I just added to Isis is a "transport" for
- multicast messages. One could bind some groups to video transport, some to
- audio, some to packets. Isis then would synchronize between the groups,
- so that you see what you would expect in terms of ordering properties.
- In Horus we are carrying this even further -- but starting with ISIS V3.0.8
- where the isis_transport stuff is really solid and used, you could get the
- same effect.
-
- The other side of the coin, though, is that this raises a huge number of
- hard technical questions and much research is needed. The whole issue of
- mixing real-time and non-real-time software (as one would be doing here)
- is in need of much study and thought. So, we may have more of a framework
- for solving such problems then a solution, at first.
-
- One hope is that we may be able to draw on technology from INESC, which
- has more background in such mixed mode systems, and we are trying to
- standardize interfaces now so that INESC software and ISIS software can
- inter-operate. INESC has done a lot in these sorts of multimedia and
- real-time communication problems, and we would like a way to benefit without
- redoing the same work. The hope is that in our new system, Horus, we
- will be able to do so.
- >
- >4) We are now considering an implementation of the platform using mobile
- > radio as the communications medium (supporting a _very_ limited range of
- > media types). Does anyone have any experiences in implementing group
- > protocols over a radio link ? More generally, does anyone know of any
- > distributed systems work aimed at such an environment ?
-
- Xerox Parc is doing work on this. The company that sells Isis plans to
- explore this area too, and also ATM communications. I am especially
- interested in hand-held devices and how they might be connected to "news".
- But, we don't have anything concrete here yet. In fact, I think the
- devices out there right now are still just a little too weak to use for
- something really fancy. But, they are close...
- --
- Kenneth P. Birman E-mail: ken@cs.cornell.edu
- 4105 Upson Hall, Dept. of Computer Science TEL: 607 255-9199 (office)
- Cornell University Ithaca, NY 14853 (USA) FAX: 607 255-4428
-