Every Manager's Guide to Information Technology:
Extract (3):
Asynchronous Transfer Mode (ATM)
ATM is a type of fast packet-switching transmission that will be the basis for most network communications by 1997. It is by
far the most important development in telecommunications since fiber optics, and, like fiber optics, it offers greatly increased speeds and lower costs. Unfortunately this term, which is widely seen as the major technical innovation in the telecommunications industry, is the same as the acronym used for an automated teller machine. Generally, the context makes it clear which term is being used, although we may see a headline within the next few years about "First National Megabank chooses ATM for ATM network."
Packet-switching is a method for sending information that divides along message into small units (packets), which it handles like a set of mailed letters, with each one containing a numbered sheet. The packets move through the network on something like a conveyor belt, mixed in with packets from other senders' messages. These packets are then routed to their destination and the full message reassembled when they all have arrived. Packet-switching allows a very large number of users to share a very high speed transmission link. The older alternative, circuit-switching, "dedicates" a link to a communications session; your phone is circuit-switched.
Packet-switched networks have historically been slow. The public data networks that use the X.25 standard for public switching allow users to operate typically at speeds of 9.6 kbps (9,600 bits per second, roughly 250 printed characters per second). The standard leased line that large companies use for their high-speed data communications operates at 56 kbps. ATM can pulse bits through the network at speeds up to 622 Mbps (millions), with much faster speeds on the horizon. Ignore the specific rates, which mean little to managers; instead, examine the scale of increase: 56 kbps is close to six times faster than9.6 kbps and 45 Mbps-a low ATM rate-is more than 800 times that.
ATM is "scaleable"; companies can use lower capacity transmission and move up to higher speeds with minimal changes.
The first demonstration of a nationwide ATM network, in November1993, linked 10 research labs. Information was transmitted at 45 million bits per second. Scientists were able to work collaboratively, sharing large volumes of data and images as if they were sitting next to each other instead of being up to 3,500 miles apart. The main conclusions from the test were that just about every existing element of network technology will need tweaking or modifying and most of them are nowhere near fast enough-as yet-to keep up with the volumes of traffic ATM can transmit.
They can also reduce their costs substantially. Initial 1994 prices for ATM equivalent to the standard private leased line were around 40 percent cheaper.
Think of ATM as a strobe light. It pulses information at fixed intervals and in fixed sizes of packet. The faster it pulses, the more packets are sent. Speed up the strobe, keeping the packet size constant, and you send more packets per second over the same transmission link.
There is a strong caveat here: ATM is the wave of the future and elements of ATM services are only sporadically being implemented now. New types of switches, fast enough to catch and transmit the strobe's pulse, are needed to handle the traffic and to connect ATM services with local and wide area networks and with another major type of fast packet-switching capability-frame relay. The telecommunications carriers (AT&T, Sprint, MCI, Bell-operated companies, and rivals) are in the early stages of rolling out ATM, which they all see as the key technology for the public data networks of the1990s and beyond. In many instances, they solve problems as they go and the articles in the trade press during 1994 that announced a new ATM product each week were full of such qualifiers as "will theoretically . . . ," "is expected to . . . ," and "initial release."
Fast packet-switching is literally the new wave of telecommunications, with ATM the key for local area networks and the public network, and frame relay a choice for large companies' backbone wide area networks (see glossary entry). The main advantage of frame relay is that while slower than ATM, it meshes easily and reliably into existing packet-switching technology and services.
The main implication of ATM for business managers is that it marks a major technical and economic shift in telecommunications that will require both complex planning-with business input and assumptions about types, volumes, and locations of information to be transmitted across the enterprise network being vital-and complex design and implementation. Stated bluntly, the network designs of 1995 are as out of date as the personal computers of 1985. Businesses will need to upgrade their networks to remain effective in an era where information now includes television, videoconferencing, document images, software programs, and any other forms of multimedia.
|