Report and Discussion on the / €€E ComSoc

JOSEPH D. TOUCH is with the ATOMIC-2 and Cities-OnLine projects at USClInfonnarion Sciences Institute. his is a summary of the second Gigabit Networking Workshop (GBN '95) sponsored by the IEEE Communications Society Technical Committee on Gigabit Networking (TCGN) association with INFOCOM in Ami1 1995. (Two earlier events were affiliated with ICC and LEOS Summer Topical Meetings, respectively.) We expect this to be an annual event in conjunction with INFOCOM. The workshop was handled entirely electronically: abstracts were submitted and reviewed and authors notified all by e-mail. The presentations were submitted by e-mail for inclusion in the on-line proceedings at http://info.gte.com/ieee-tcgn/conference/gbn95. In an attempt to encourage the presentation of current and ongoing work, the duration from the submission of abstracts to on-line publication before the workshop was only four weeks, with this report following three months later. The focus of the TCGN and GBN workshop is on end-to-end (layer 4) and higher issues. There are of course a number of related network layer issues of interest, particularly in considering the end-toend implications of network and internetwork architecture, but there isvirtually nocoverage ofthe physical and data link layers. Of the 27 submissions, 13 were accepted for presentation and one was used to lead a discussion session, resulting in submissions from five countries. Additionally, six were given poster space on the on-line proceedings. While a majority of the submissions had strong ATM (asynchronous transfer mode) content, quite a few had to do with TCPiIP (transmission control protocoliinternet protocol) over ATM and B-ISDN (broadband integrated services digital networks), reflecting the emerging view that the Internet and ATM communities will have to try to coexist in the future. In fact, only two of the accepted contributions were completely free of ATM content. The sessions reflects this, starting out with some pure ATM/B-ISDN topics, continuing with TCP/IP over ATM, and leading into a discussion session critiquing ATM andconsidering its role in the future Internet-based Global Information Infrastructure (GII). The afternoon was devoted to non-ATM topics, followed by a discussion session on Middlewareinwhichwe triedourbest tonotmention ATM (without always succeeding). The strong message provided by this year's contributions is that a number of researchers are concerned with the practical implications of running TCP/IP over ATM, and most of those dealing with higher layer issues use TCPIIP over ATM out of expediency.