Graphic-Newsletter of UH Community


Infobit Logo - Link to home page 

Link to previous articleLink to next article

A Serpent Invades Paradise
by Alan Whinery

Picture of hose.The history of "UHNet", as it is now, began in the Summer of 1987, when various faculty and staff came to the Manoa Campus on a Sunday morning with shovels, pickaxes, Ethernet supplies, and aircraft refueling hose, acquired from military surplus by Don Cole of the Chemistry Department. They came because they had been unsuccessful in convincing the UH administration that building a TCP/IP network on campus should be an institutional priority. In short, there was no budget. The digging crew consisted of a handful of scientists and technical staff, who had paid for the equipment out of their project budgets, and out of their pockets. On that Sunday, and a couple of subsequent outings, they dug trenches, buried the aircraft refueling hose, cored holes in the walls of several buildings, and used the hose as conduit to pull thick Ethernet cable from Keller Hall to Bilger, Bilger to Watanabe. The resulting backbone was called the Campus Research Network (CRN), and it joined the Computing Center, Information and Computer Sciences, Chemistry, Math, and Physics departments together by attaching to a machine in each department. BSD Unix, VAX VMS, and TOPS 20 machines served as hosts and routers, some of which had LANs behind them in their respective departments. The original link to the outside world, something of a bigger deal than it was in most places, was a 9600 baud link to NOSC Kaneohe, which was used for UUCP mail and Usenet news.

The enterprise with the aircraft refueling hose, as a grass roots exercise, should be put in perspective by mentioning that the University already had inter-building optical fiber at the time. There was already a project or two that made use of fiber to connect terminals and printers to the IBM mainframe and some of the academic Computing Center machines. This separate, parallel evolution of datacomm for academic versus administrative computing continued for about five years, until the ad hoc merger of efforts in 1993, and then the formal merger of the Computing Center and the Management Systems Office (called ITS) in 1994.

Within a year after the charter members of CRN buried the hose, Torben Nielsen, then of ICS, acquired a grant to network NASA activities in the Pacific. The effects of this project on the evolution of UHNet cannot be overstated. Dr. Nielsen made his mark as a visionary and the original UHNet engineer by bringing funding to the picture and marrying the destiny of UHNet to the interests of NASA, a marriage that leant direction and sophistication to the development of the network.

Torben’s project, PACCOM, is mentioned in some editions of John Quarterman’s book, The Matrix, which shows a snapshot of the Pacific Communications Network as a giant spider that had feet in California, Hawaii, Japan, New Zealand, Korea, and perhaps other places. Within Hawaii, PACCOM brought the vestigial Internet to Kokee Park, Kauai’s Radio observatory, the summit of Mauna Kea on the Big Island, the Galileo Spacecraft Project in Mililani, and Haleakala’s Summit on Maui. Trans-pacific links were provided by WorldComm, most were fractional T1.

After a spell of Usenet delivered by satellite, the real live Internet came to UH and to PACCOM in the form of a T1 (1.536 Mbps) link to Ames Research Center in Moffett Field California in 1991. Those were the salad days of the NASA Science Internet, SPAN, and HEPNet, which were all riders on the PACCOM link, using TCP/IP and DECNet to internetwork Hawaii scientists with their peers on the mainland and abroad. UH enjoyed the ride, and began contributing to the costs of the link, thanks to the participation of several visionaries at UHCC, including Walter Yee and Jeff Blomberg. Walter Yee is someone about whom many people will say many things, and he was the UH Computing Center personified, and in fact computing personified for two decades of UH history. His leadership was valuable in the respect that he was able to listen to people that were saying some pretty strange sounding things, and then allow them to perform miracles, when he could just as easily not have listened. As a matter of fact, he probably chose the hard way, for which the University is forever indebted to him. Walter had access to many brilliant and capable technicians, chief among them Jeff Blomberg, who assimilated mountains of new concepts and applied them to securing UH’s role in the global Internet, before anybody knew what it was.

At the same time, David Lassner was focusing on a more local application of Internet-ness. Amidst the disarray of the 1990 UH Manoa telecom project, David was working the State of Hawaii Information and Computer Services Division to shape UH’s participation in HAWAIIAN, the State’s inter and intra-island digital backbone. He worked with vendors for several months to specify the equipment that would bring Internet connectivity to almost every campus in the UH system, with T1’s provided over HAWAIIAN.

When I joined the UH Computing Center staff in December 1991, the PACCOM spider was a huge, strange beast, and UH Manoa's machine room was crowded with new network equipment boxes. There was no TCP/IP connectivity from Manoa to other campuses, although there were leased lines between the administrative VAXes, which used DECNet. I had been an Internet user for two years during my time at the Institute for Astronomy (IFA), and I had helped Pui Hin Rhoads put bags of ice on the Ethernet laser link from UH Manoa Marine Sciences Building to IFA to alleviate overheating. There was no World Wide Web. There were USENET newsgroups, e-mail, and FTP. Internet access was something one did through a Unix machine, and the idea of putting a PC on the Internet was considered cutting edge, yet strange. Dr. Nielsen showed me an "HTTP browser" program. I didn't get it.

The change in the Internet's role in global society during the 1990's occurred in tandem with the internetworking of the university's numerous facilities statewide. While HAWAIIAN took shape, UH contracted with GTE Hawaiian Telephone to bring T1 connectivity to each of the community colleges. GTE's capacity to deliver T1 circuits to each of the 10 UH system campuses was built as the circuits were ordered; the T1 to Honolulu Community college was originally delivered by microwave, because the wires were not yet in place. HAWAIIAN's T1 links gradually replaced GTE links over a period of about three years.

Dr. Ken Hensarling at HCC set up one of the first popular "Web" sites, at a time when nobody had heard of a "dot com". The "World Wide Web" was declared into existence in August of 1993. In June of 1993, Web traffic accounted for 18% of all Internet traffic to the mainland. In June of 1994, Web traffic had increased to 58% of total Internet traffic.

UH began offering unilateral "access-for-all" to all faculty, staff and students across the UH system in July 1994. Prior to this time, each individual user had to qualify for an ITS e-mail address.

That same year, implementation of administrative systems leant impetus and funding to the statewide UH network, providing high-performance routers for each of 10 campuses, as well as one for Hawaii State ICSD.

Also in 1994, the availability of frame relay, a new service from GTE, and its attractive pricing model, brought the Internet to the last three campuses in the UH system, on Lanai, Molokai, and in Kealakekua on the Big Island. Connectivity to Coconut Island and the Hawaii Institute of Marine Biology became possible when the price of a 56Kbps connection fell from over $1000 per month to $92 per month, using frame relay.

In June of 1995, the university's T1 link was becoming extremely "well used". UHNet had grown to serve over 10,000 computers, and the advent of the World Wide Web had made Internet access popular. Negotiations were begun to acquire more bandwidth.

In October of 1996, the University added a 6Mbps link to SprintLink in Stockton, California. The 6Mbps was a combination of four T1 links, which worked relatively well. In April of 1998, the 6Mbps link was becoming full, and another three T1's were added, for a total capacity of 10.75Mbps. Traffic continued to increase. In November 1998, the primary Internet link was increased to 45Mbps, now a "T3" instead of multiple T1s.

The university began participation in the UCAID Internet 2 project in April, 1999. Thanks to cooperation from the Defense Engineering and Research Network (DREN), a second T3 link to the mainland was added to connect the University of Hawaii to Abilene, one of two high performance networks which interconnect universities and research institutions together to investigate ways to build the networks of the future. Pretty much the same things that they built Internet 1 for.

Link to return to contents
Link to Infobits home page
Link to go to next article

Comments to: