User:Howard C. Berkowitz

Network engineer going back before there was an Internet. Was an active participant in OSI architecture and formal testing, but have been involved in the Internet Engineering Task Force, North American Network Operators Group, and Internet Research Task Force. Published four network engineering textbooks, four IETF RFCs plus contributions to other documents such as "Future Domain Requirements" (i.e., what happens after BGP), many professional presentations.

Extensive experience in fault-tolerant networking and computing for telecommunications, healthcare and military applications. Background in biochemistry, emergency management, and continuing medical education allows me to help physicians define what they want to do with systems, and translate these to implementation requirements. Recent involvement in computing and networking for maritime navigation, safety, and fisheries management.

Long-term interest in intelligence and special operations; principal contributor to the Intelligence Cycle Management series on Wikipedia. Have done recent OSINT consulting, although for confidential clients.

Network engineering

My experience with communications standards goes back to the mid-seventies, variously with ISO/CCITT and ANSI to start, especially in network performance. I worked for GTE for a time, and had a good deal of exposure to the internals of telephone networks. As a member of the Federal Telecommunications Standards Committee (1976-1980), I got in at the beginning of what was to become OSI, and also got interested in survivable communications systems, including the (US) National Communications Systems and military networks intended to operate under the most extreme conditions. Those extremes tended to be that the network really needed to operate for 20 minutes or so, but you never knew when the 20 minutes would start, and would just have to cope with network elements randomly turning into mushroom clouds. This tied in with a lifelong interest in politicomilitary history.

The FTSC and National Communications System contributed, in the late seventies, to the ANSI Distributed Systems (DISY) architecture, which was a significant input into the OSI architecture. ISO 7498, the basic OSI Reference Model <e, was published in 1984. Even ignoring the eventual dominance of Internet protocols, there is a great deal of misunderstanding about OSI, because educators generally ignored supplemental ISO documents that clarified ISO 7498.

From 1986 to 1991, I was the first technical staff member at the Corporation for Open Systems, a not-for-profit industry research center for promoting and testing OSI and ISDN protocols. In addition to secretariat work with the various committees, I managed teams working on FTAM and X.25 test systems, and contributed to IEEE 802 test systems. One memorable experience was lecturing about X.25 testing in Japan, and had the horrible realization that my PowerPoint slides, translated into Japanese, had gotten into a different order than my English-language notes.

For around six years of my life, I explained how OSI was the answer, but eventually realized I didn't know the question.

The moving hand writes on the wall: "it's about IP, stupid"

By the early nineties, it was obvious that Internet protocols were indeed the answer, and I started to play in the Internet Engineering Task Force (IETF), North American Network Operators Group (NANOG) and the Internet Research Task Force (IRTF).

In the IETF, my main work has been in the Routing (especially BGP/IDR and OSPF) and Operations & Management Areas (especially BMWG & OPSEC), and, more as a lurker, Security and Real-time Applications & Infrastructure Area. I am an author or coauthor of RFC 1912, RFC 2071, RFC 2072, RFC 4098, and was a reviewer or contributor with many others. I've done quite a few tutorials and presentations available at www.nanog.org, and was a participant in "Team B" of the IRTF Future Domain Requirements effort, which essentially looked at the question "what comes after BGP?" Some of my most satisfying work came when I was first the product line manager for routing protocols in the carrier router group, and then in corporate research at Nortel, both working with standards and operational forums, and designing a next-generation router.

Why network architectures and standards?

The early days of computer networks were dominated by a few large companies such as IBM and DEC. In order promote interoperability and avoid a situation where a small number of vendors predominated, each with their own proprietary technology, it was necessary to introduce a set of open standards defining network protocols.

Another issue addressed by this model is maintaining the level of flexibility needed to adapt when new innovations are introduced. The earliest wide area networks (or WANs) ran over telephone lines and were used to link a small number of facilities.

Today, we rarely think about why Internet access has become so ubiquitous. Still, this is quite a technical achievement: a user may be connected to an Ethernet network, Fiber Distributed Data Interface (FDDI) or a wireless network in a coffee house. They may also use such diverse methods as DSL, cable, or dialup lines to "get online."