TELENET

Template:TOC-right Telenet was a commercial packet switching which went into service in 1975. It was the first publicly available commercial packet-switched network service in the United States; a Spanish banking network between Madrid and Barcelona preceded it by a few months. Tymnet, a competitive packet switching network, started operations slightly after Telenet.

Do not confuse TELENET, a company that ran a network, with the Internet Protocol Suite telnet protocol.

As opposed to the Internet, it used a virtual circuit rather than a datagram architecture: it set up the equivalent of telephone calls between the two endpoints, and the switching elements in the internal network remembered the status of calls. In datagram-based packet switching, routers handling packets have no memory of previous transmission; much as the post office decides how to route each envelope between airports, post offices, trucks, etc. Virtual circuit technology reflects the telephone company engineers that created it; a telephone network is defined to be intelligent, with very little intelligence in telephones. The Internet model, however, puts intelligence at the edges and makes the internal network as simple and fast as possible. Both models have advantages and disadvantages; see Multi-Protocol Label Switching for a modern approach that combines elements of the two methods.

At the time, however, the virtual circuit people insisted the ARPANET would not work, and the datagram people insisted virtual circuits could not scale.

The original founding company, Telenet Inc. was established by Larry Roberts (former head of the ARPANet), and Barry Wessler. GTE acquired Telenet in 1979. It was later acquired by Sprint and called "Sprintnet." Sprint migrated customers from Telenet to the modern-day SprintLink IP network, primarily using Cisco routers. one of many networks composing today's Internet. Telenet had its first office in downtown Washington, D.C., then moved to McLean, Virginia. It was acquired by GTE while in McLean, and then moved to offices in Reston, Virginia.

Under the various names, the company operated a public network, and also sold its packet switching equipment to other carriers and to large enterprise networks.

Coverage

Originally, the public network had switching nodes in seven US cities:

• Washington, D.C. (network operations center as well as switching)
• Boston, MA
• New York, NY
• Chicago, IL
• Dallas, TX
• San Francisco, CA
• Los Angeles, CA

The switching nodes were fed by Telenet Access Controller (TAC) terminal concentrators both colocated and remote from the switches. The switches used the X.25 protocol, which was standardized internationally by the body now known as the International Telecommunications Union (ITU). The switches intercommunicated with a modified X.25 protocol called Telenet Internetwork Protocol (TINP).

By 1980, there were over 1000 switches in the public network. At that time, the next largest network using Telenet switches was that of Southern Bell, which had approximately 250 switches. Telenet used the international standard X.75 protocol, a slight variation on X.25, to connect to other X.25 networks, much as an international telephone call may go through the equipment of different telephone companies, or an Internet packet goes through multiple Internet Service Providers on its path from source to destination.

Internal Network Technology

The initial network used statically-defined hop-by-hop routing, using Prime commercial minicomputers as switches, but then migrated to a purpose-built multiprocessing switch based on 6502 microprocessors. Among the innovations of this second-generation switch was a patented arbitrated bus interface that created a switching fabric, a shared bus in modern terms, among the microprocessors .

Most interswitch lines ran at 56 kbit/s, with a few, such as New York-Washington, at T1 (i.e., 1.544 Mbit/s). The main internal protocol was a proprietary variant on X.75; Telenet also ran standard X.75 gateways to other packet switching networks.

Originally, the switching tables could not be altered separately from the main executable code, and topology updates had to be made by deliberately crashing the switch code and forcing a reboot from the network management center. Improvements in the software allowed new tables to be loaded, but the network never used dynamic routing protocols. Multiple static routes, on a switch-by-switch basis, could be defined for fault tolerance. Network management functions continued to run on Prime minicomputers.

Its X.25 host interface was the first in the industry and Telenet helped standardize X.25 in the CCITT, which was the predecessor to the ITU.

Accessing the Network

Basic Asynchronous Access

Users could use modems on the Public Switched Telephone Network to dial TAC ports, calling either from "dumb" terminals or from computers emulating such terminals. Organizations with a large number of local terminals could install a TAC on their own site, which used a dedicated line, at up to 56 kbit/s, to connect to a switch at the nearest Telenet location. Dialup modems supported had a maximum speed of 1200 bit/s, and later 4800 bit/s.

Computer Access

Computers supporting the X.25 protocol could connect directly to switching centers. These connections ranged from 2.4 to 56 kbit/s.

Other Access Protocols

Telenet supported remote concentrators for IBM 3270 family intelligent terminals, which communicated, via X.25, to Telenet-written software that ran in IBM 370x series front-end processors.