Networking Concepts

Wide Area Networks
The evolution of wide area networks can be considered to have originated
in the mid- to late 1950s, commensurate with the development of the first
generation of computers. Based on the use of vacuum tube technology, the first
generation of computers were large, power-hungry devices whose placement
resulted in a focal point for data processing and the coinage of the term
data center.
Computer-Communications Evolution
Originally, access to the computational capability of first-generation computers
was through the use of punched cards. After an employee of the
organization used a keypunch to create a deck of cards, that card deck was
submitted to a window in the data center, typically labeled input/output (I/O)
control. An employee behind the window would accept the card deck and
complete a form that contained instructions for running the submitted job.
The card deck and instructions would then be sent to a person in production
control, who would schedule the job and turn it over to operations for
execution at a predefined time. Once the job was completed, the card deck
and any resulting output would be sent back to I/O control, enabling the job
originator to return to the window in the data center to retrieve his or her
card deck and the resulting output. With a little bit of luck, programmers
might see the results of their efforts on the same day that they submitted
their jobs.
Because the computer represented a considerable financial investment for
most organizations, it was understandable that these organizations would
be receptive to the possibility of extending their computers’ accessibility.
By the mid-1960s, several computer manufacturers had added remote access
capabilities to one or more of their computers.
Remote Batch Transmission
One method of providing remote access was the installation of a batch
terminal at a remote location. That terminal was connected via a telephone
company–supplied analog leased line and a pair of modems to the computer
in the corporate data center.
The first type of batch terminal developed to communicate with a data
center computer contained a card reader, a printer, a serial communications
adapter, and hard-wired logic in one common housing. The serial communications
adapter converted the parallel bits of each internal byte read from the
card reader into a serial data stream for transmission. Similarly, the adapter
performed a reverse conversion process, converting a sequence of received
serial bits into an appropriate number of parallel bits to represent a character
internally within the batch terminal. Because the batch terminal was located
remotely from the data center, it was often referred to as a remote batch
terminal, while the process of transmitting data was referred to as remote
batch transmission. In addition, the use of a remote terminal as a mechanism
for grouping card decks of individual jobs, all executed at the remote data
center, resulted in the term remote job entry terminal being used as a name
for this device.
Figure 1.1 illustrates in schematic form the relationships between a batch
terminal, transmission line, modems, and the data center computer. Because
the transmission line connects a remote batch terminal in one geographic area
to a computer located in a different geographic area, Figure 1.1 represents one
of the earliest types of wide area data communications networks.
Paralleling the introduction of remote batch terminals was the development
of a series of terminal devices, control units, and specialized communications
equipment, which resulted in the rapid expansion of interactive
computer applications. One of themost prominent collections of products was
introduced by the IBM Corporation under the trade name 3270 Information
Display System.
Remote batch transmission. The transmission of data from a
remote batch terminal represents one of the first examples of wide area data
communications networks.
IBM 3270 Information Display System
The IBM 3270 Information Display System was a term originally used to
describe a collection of products ranging from interactive terminals that
communicate with a computer, referred to as display stations, through several
types of control units and communications controllers. Later, through
the introduction of additional communications products from IBM and
numerous third-party vendors and the replacement of previously introduced
products, the IBM 3270 Information Display System became more
of a networking architecture and strategy rather than a simple collection
of products.
First introduced in 1971, the IBM 3270 Information Display System was
designed to extend the processing power of the data center computer to
remote locations. Because the data center computer typically represented the
organization’s main computer, the term mainframe was coined to refer to a
computer with a large processing capability. As the mainframe was primarily
designed for data processing, its utilization for supporting communications
degraded its performance.
Communications Controller
To offload communications functions from the mainframe, IBM and other
computer manufacturers developed hardware to sample communications
lines for incoming bits, group bits into bytes, and pass a group of bytes
to the mainframe for processing. This hardware also performed a reverse
function for data destined from the mainframe to remote devices. When
first introduced, such hardware was designed using fixed logic circuitry,
and the resulting device was referred to as a communications controller.
Later, minicomputers were developed to execute communications programs,
with the ability to change the functionality of communications support by
the modification of software—a considerable enhancement to the capabilities
of this series of products. Because both hard-wired communications
controllers and programmed minicomputers performing communications
offloaded communications processing from the mainframe, the term frontend
processor evolved to refer to this category of communications equipment.
Although most vendors refer to a minicomputer used to offload communications
processing from the mainframe as a front-end processor, IBM
has retained the term communications controller, even though their fixed
logic hardware products were replaced over 20 years ago by programmable
minicomputers.
Control Units
To reduce the number of controller ports required to support terminals, as
well as the amount of cabling between controller ports and terminals, IBM
developed poll and select software to support its 3270 Information Display
System. This software enabled the communications controller to transmit
messages from one port to one or more terminals in a predefined group
of devices. To share the communications controller port, IBM developed
a product called a control unit, which acts as an interface between the
communications controller and a group of terminals.
In general terms, the communications controller transmits a message to the
control unit. The control unit examines the terminal address and retransmits
the message to the appropriate terminal. Thus, control units are devices that
reduce the number of lines required to link display stations to mainframe computers.
Both local and remote control units are available; the key differences
between them are the method of attachment to the mainframe computer and
the use of intermediate devices between the control unit and the mainframe.
Local control units are usually attached to a channel on the mainframe,
whereas remote control units are connected to the mainframe’s front-end
processor, which is also known as a communications controller in the IBM
environment. Because a local control unit is within a limited distance of the
mainframe, no intermediate communications devices, such as modems or data
service units, are required to connect a local control unit to the mainframe.
In comparison, a remote control unit can be located in another building or
in a different city; it normally requires the utilization of intermediate communications
devices, such as a pair of modems or a pair of data service
units, for communications to occur between the control unit and the communications
controller. The relationship of local and remote control units to
display stations, mainframes, and a communications controller is illustrated
in Figure 1.2.
Network Construction
To provide batch and interactive access to the corporate mainframe from
remote locations, organizations began to build sophisticated networks. At
first, communications equipment such as modems and transmission lines was
obtainable only from AT&T and other telephone companies. Beginning in
1974 in the United States with the well-known Carterphone decision, competitive
non–telephone company sources for the supply of communications
equipment became available. The divestiture of AT&T during the 1980s and
the emergence of many local and long-distance communications carriers
paved the way for networking personnel to be able to select from among
several or even hundreds of vendors for transmission lines and communications
equipment.
As organizations began to link additional remote locations to their mainframes,
the cost of providing communications began to escalate rapidly.
This, in turn, provided the rationale for the development of a series of linesharing
products referred to as multiplexers and concentrators. Although
most organizations operated separate data and voice networks, in the mid-
1980s communications carriers began to make available for commercial use
high-capacity circuits known as T1 in North America and E1 in Europe.
Through the development of T1 and E1 multiplexers, voice, data, and video
transmission can share the use of common high-speed circuits. Because the
interconnection of corporate offices with communications equipment and
facilities normally covers a wide geographical area outside the boundary
of one metropolitan area, the resulting network is known as a wide area
network (WAN).
Figure 1.3 shows an example of a wide area network spanning the continental
United States. In this example, regional offices in San Francisco and New
York are connected with the corporate headquarters, located in Atlanta, via T1
multiplexers and T1 transmission lines operating at 1.544 Mbps. Assuming
that each T1 multiplexer is capable of supporting the direct attachment of
a private branch exchange (PBX), both voice and data are carried by the T1
circuits between the two regional offices and corporate headquarters. The
three T1 circuits can be considered the primary data highway, or backbone,
of the corporate network.
Wide area network example. A WAN uses telecommunications
lines obtained from one or more communications carriers to connect geographically
dispersed locations.
In addition to the three major corporate sites that require the ability to route
voice calls and data between locations, let us assume that the corporation
also has three smaller area offices located in Sacramento, California; Macon,
Georgia; and New Haven, Connecticut. If these locations only require data
terminals to access the corporate network for routing to the computers located
in San Francisco and New York, one possible mechanism to provide network
support is obtained through the use of tail circuits. These tail circuits could
be used to connect a statistical time division multiplexer (STDM) in each area
office, each serving a group of data terminals to the nearest T1 multiplexer,
using either analog or digital circuits. The T1 multiplexer would then be
configured to route data terminal traffic over the corporate backbone portion
of the network to its destination.
Network Characteristics
There are certain characteristics we can associate with wide area networks.
First, theWANis typically designed to connect two or more geographical areas.
This connection is accomplished by the lease of transmission facilities from
one or more communications vendors. Secondly, most WAN transmission
occurs at or under a data rate of 1.544 Mbps or 2.048 Mbps, which are the
operating rates of T1 and E1 transmission facilities.
A third characteristic of WANs concerns the regulation of the transmission
facilities used for their construction. Most, if not all, transmission facilities
marketed by communications carriers are subject to a degree of regulation at
the federal, state, and possibly local government levels. Even though we now
live in an era of deregulation, carriers must seek approval for many offerings
before making new facilities available for use. In addition, although many
of the regulatory controls governing the pricing of services were removed,
the communications market is still not a truly free market. Thus, regulatory
agencies at the federal, state, and local levels still maintain a degree of
control over both the offering and pricing of new services and the pricing of
existing services.
1.2 Local Area Networks
The origin of local area networks can be traced, in part, to IBM terminal equipment
introduced in 1974. At that time, IBM introduced a series of terminal
devices designed for use in transaction-processing applications for banking
and retailing. What was unique about those terminals was their method of connection:
a common cable that formed a loop provided a communications path
within a localized geographical area. Unfortunately, limitations in the data
transfer rate, incompatibility between individual IBM loop systems, and other
problems precluded the widespread adoption of this method of networking.
The economics of media sharing and the ability to provide common access
to a centralized resource were, however, key advantages, and they resulted
in IBM and other vendors investigating the use of different techniques to
provide a localized communications capability between different devices. In
1977, Datapoint Corporation began selling its Attached Resource Computer
Network (ARCNet), considered by most people to be the first commercial local
area networking product. Since then, hundreds of companies have developed
local area networking products, and the installed base of terminal devices
connected to such networks has increased exponentially. They now number
in the hundreds of millions.
Comparison to WANs
Local area networks can be distinguished from wide area networks by geographic
area of coverage, data transmission and error rates, ownership,
government regulation, and data routing—and, in many instances, by the
type of information transmitted over the network.