Middlware'98 , OOPSLA'98, and SIGCOMM'98: where will the bouncing ball

Rohit Khare (rohit@bordeaux.ICS.uci.edu)
Thu, 15 Jan 1998 23:33:51 -0800


"Middleware'98 will be the premier conference on distributed systems platforms
and open distributed processing in 1998. The conference is a synthesis of the
major conferences and workshops in this area into a single international event
and representatives of ICODP (International Conference on Open Distributed
Processing), ICDP (International Conference on Distributed Platforms), DAIS
(Distributed Applications and Interoperable Systems) and SDNE (Services in
Distributed and] Networked Environments) are included on the programme

"The focus of Middleware'98 is on the design, implementation, deployment and
evaluation of distributed systems platforms and architectures for future
networked environments. Of particular interest is the application of both new
and existing architectures and platforms (such as RM-ODP, CORBA and DCOM) in
environments which may include public and private networks, overlayed wired
and wireless technologies, IPv6 and IP multicast, multimedia and real-time
information and an increasing volume of WWW and Java traffic. "

Call me a conference slut, but I don't see how to go through this grad school
thing *without* emitting a paper every six weeks or so. I'm very enamored of
my ability to think Big Thoughts... :-)

In this case, I'm trying to find a suitable host to aim a draft of the xTP
hypothesis. Generic reliable datagram services seem to have gone entirely
underground since the silent vaprous nonsuccess of VMTP ca. 1987 (Versatile
Message Transport Protocol). I still think there's a common message
microkernel below SMTP, HTTP, NNTP, FTP, &c.

But I'm only beginning down this path of inquiry. I'm not yet builidng a
system (heavens no! with this class load...) I need to do a good survey of the
set of TPs, which is a past due term paper from last month; and then begin to
synthesize that into a set of rules for how-we-evolved-so-many-TPs and into a
concrete xTP proposal, parallelling the HTTP-NG work.

March 13 is the deadline for MW'98. Program committee includes:

Peter Honeyman, CITI, University of Michigen, USA.
Derek McAuley, Microsoft Research, UK.
Robert van Renesse, Cornell University, USA.
Doug Schmidt, Washington University, USA.
Richard Soley, OMG, USA.
Doug Terry, Xerox PARC, USA.

More realistically, though, I'd like to extract out some patterns behind TP
developoment: the FloodFill and StoreAndForward distribution strategies,
RetrofittingStateWithCookies, MediaTypeMetadata, and other patterns which are
composed in the wild. A Pattern Language for Messaging Protocols. OOPSLA woudl
be due April 13th. http://www.acm.org/sigplan/oopsla/progcomm.htm

And last but not least, I still harbor dreams that Roy, Henrik, myself, or
some combination thereof will sit down to wrestle the HTTP Object Model into
an archival paper -- but SIGCOMM'98's 1/30 deadline is too close for that.

Your distributed writer,
Rohit Khare

PS. Oh yeah, C.K. Toh is presenting a talk on Wireless and Ad-Hoc networks
here at UCI this Friday 10 AM, in the EE department. Saw his talk at IETF-DC;
should be good.

PPS. Did anyone else notice the release of CSS2 last November?