Technology Developed By Ibm Research Is Contributing To Two Major Telecommunications Projects - Providing Phone
Service To A Campus And To A Country
In Brief:
IBM's Research Division is playing a major role in two solutions that link computers and telephones. Staffers from the Haifa and Zurich research laboratories are helping Cornell University to decentralize its phone network. And a Zurich team is providing network management for o.tel.o, a new German phone company.
Yoav Medan of the Haifa Research Laboratory calls it "commputing" - the combination of computer technology and telephonic techniques that will form the basis of communications in the next century. IBM is working to ensure a leadership role in that revolution by developing two new projects, widely separated in both geography and ambition.
Each involves a search for solutions. In one, researchers at Haifa and at the Zurich Research Laboratory are helping to put together a communications network for a college campus, based on asynchronous transfer mode (ATM) technology, a field in which IBM has played a leading role. In another, the company's researchers are helping to set up a new telecommunications system for an entire country, in response to an initiative intended to open up the traditionally closed telecommunications field to competition.
The campus telephony project has the goal of reducing the reliance of Cornell University, in upstate New York, on private branch exchange (PBX) systems. While they can provide advanced voice and data communications, PBX systems have a major disadvantage. "They are completely centralized," explains Zurich's Wolfgang Kleinöder, "whereas the goal of campus telephony is to be totally decentralized." Decentralization permits students and - frequently - faculty members to move around the campus with their laptops, hoping to make contact with telephone service whatever their location. The ultimate goal is to link all computers and telephones on the campus to a single cable, rather than relying on separate lines for each.
PBXs are also expensive. A campus implementation can cost $1 million or more. So one year ago, Cornell's Information Technology Division started work on a way to reduce its reliance on PBXs by transferring much of its phone traffic to a planned campus ATM network that can both act as a campus backbone and eventually deliver data, video and voice services directly to desktops. "We're basically looking to build the ATM networks for data use as soon as possible," says Dick Cogger, assistant director for advanced technology planning at Cornell.
Transferring PBX functions to the ATM network won't be easy, though, because the campus phone system will have to work throughout the process. Hence, says Colin Harrison, a Research staff member at the Thomas J. Watson Research Center who worked on the project at Zurich, "the changeover to ATM has to be incremental. The phones in the new system must work with those in the old." Indeed, critical telephone services will remain on the existing voice switches until overseers of the project regard the new infrastructure as sufficiently reliable.
Cells In Frames
To meet that requirement, Cornell first set up a project called Cells in Frames. Its aim: to devise a means by which the frames in which data is sent through the Ethernet technology used in its current local area network (LAN) can incorporate the packets of data - or cells - transmitted via ATM. With that fundamental work now completed, says Zurich scientist Lucas Heusler, "Cornell is keen about getting it into its net. It is asking IBM to put in an ATM backbone and enhance the Ethernet solution to implement the Cells in Frames idea."
To develop the solution, IBM is adapting technology from the Haifa and Zurich labs. A key contribution from Haifa is a gateway, running on an RS/6000 SP® or PowerPC® workstation. The gateway connects to both a PBX, via ISDN lines, and to the ATM
network. "The gateway will make the ATM network appear to be an ISDN network to the PBX," explains Pnina Vortman of the Haifa lab.
The two labs have also contributed to the network by making it compatible with ITU H.323, an emerging industry standard that applies to both telephony and video transmission, and that has recently been upgraded to support ATM. IBM will provide the middleware for a workstation-based H.323 terminal and an H.323-based gateway. It contains a protocol that makes the audio information stream smoothly through the computer and over the network, an area where the Haifa lab is drawing on its long experience with real-time software.
The Zurich lab, meanwhile, is developing signaling procedures for the campus network. "We're working to make everything you get on a PBX available in a decentralized way over a packet-based ATM network," explains Heusler. In particular, says Harrison, the Zurich team has implemented the signaling part of H.323. "We changed to the H.323 signal standard for Internet telephony," adds Vortman. "This gives the possibility of interoperability with other standard phones."
Open To Competition
While Cornell's telephony plan is effectively a pilot project for new technology, Kleinöder is working on a full-scale system with a rapidly approaching due date. On January 2, 1998, telephone services in the 15 member countries of the European Union (EU) will become open to competition. Several new competitors hope to gain significant market shares from telecommunications monopolies in individual EU countries, by a combination of technology and market strategy.
One participant in the race is o.tel.o, a German company that specializes in data services, mobile communications, cable TV and corporate networks. Now, o.tel.o is preparing to take on Deutsche Telekom, the company that has long had a monopoly on telephone service in Germany. As a small part of a project run by IBM's global telecommunications industry solutions unit, Kleinöder and a Zurich team are working on telecommunications network management for o.tel.o's system. IBM is due to deliver the first version of its technology in early October, a scant three months before the start of serious competition.
"We will use several SP machines to manage the system," says Kleinöder. "This is the largest system ever based on IBM telecommunications management technology." The technology enables operators to manage devices for voice, X25, frame relay and ATM networks from different vendors. Zurich has focused on toolkits to develop "agents" that interface with various types of telecommunications equipment.
However, Zurich lacked experience with customers' real-world problems. "What we had was the technology," explains Kleinöder. "We hadn't developed the solutions." Now, the collaboration with o.tel.o gives IBM and the Zurich lab the opportunity to convert its basic research into a real solution - a solution
focused on one of the most competitive markets in the "commputing" arena.
