030.50.2 Gigabit Networks by Craig Partridge There are now strong signs that research and development in gigabit networking has begun to make that critical move from paper studies to real systems of fiber, silicon and lasers. First, we're beginning to see a flood of papers from researchers who have built at least some hardware. Two different papers on recently constructed ATM (Asychronous Transfer Mode) interfaces were published at SIGCOMM '91. The folks at Cambridge University have been reporting interesting work with their 600Mbit/second prototype of the Cambridge Backbone Ring. And IBM-Zurich demonstrated its gigabit LAN at a telephony conference this fall. (Indeed, it is somewhat sobering truth that we may be innundated with gigabit literature in 1992.) A second, and possibly more important, development is that the computer communications community has begun to get seriously involved in the process of making standards for future gigabit networks. In particular, the computer communications community began to get involved in Broadband ISDN (B-ISDN). The standards process for Broadband ISDN (which is built on ATM) has been underway in CCITT for a few years now. B-ISDN is the technology that the telephony community plans to deploy in the coming decades to serve a wide range of voice and data communications requirements. Before late 1990, almost no one in the computer communications community (people worried about making computers communicate) had looked very closely at the B-ISDN standards. By late 1991, the computer communications folk had taken a close look at B-ISDN and concluded that it didn't adequately support the transmission of datagram traffic (a big concern for protocols like IP!). So in late 1991 a consortium of computer companies (Apple, BBN, DEC, IBM Motorola/Codex, Sun Microsystems and Xerox) brought their concerns before the US ANSI T1S1.5 standards committee in the form of a proposal called SEAL to augment the B-ISDN standards. ANSI and CCITT are trying to wrap up the initial B-ISDN specifications so this late appearance by the computer companies could have caused problems. But instead, in a pleasant case of the standards process working as it should, the consortium's contributions were welcomed and the committee agreed to ask for a change to the ATM header to support the SEAL proposal. This change will make it possible to experiment with implementing datagram protocols over B-ISDN. (Some of this implementation work will be done in conjunction with the IETF. Send mail to atm-request@bbn.com to join the IETF discussions). The plan is to take the results of this experimentation and incorporate it into the next standards release.