home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Magnum One
/
Magnum One (Mid-American Digital) (Disc Manufacturing).iso
/
d7
/
comproto.arc
/
PAPER-01.TXT
< prev
next >
Wrap
Text File
|
1991-04-28
|
18KB
|
352 lines
COMMUNICATIONS PROTOCOLS FOR PERSONAL COMPUTERS
Hayes Microcomputer Products, Inc.
705 Westech Drive
Norcross, Georgia 30092
INTRODUCTION
Since the beginning of human communications, man has defined
"protocols," or sets of rules, to provide for the orderly interchange
of information. These have included agreement on sounds, languages,
alphabets, vocabularies, grammar, syntax, punctuation, and
formatting. The precision required in electronic communications
needs similar, very carefully defined and precisely implemented
protocols.
Such protocols begin with agreement on the binary
representation of data. Standards such as the American Standard Code
for Information Interchange (ASCII) and the Extended Binary-Coded
Decimal Interchange Code (EBCDIC), define specific bit patterns that
are interpreted as particular characters. These are the "alphabets"
of computers, and provide the basis for the orderly use and exchange
of information.
Encoded characters can be transferred between computers using
two methods: parallel or serial. Parallel communications requires a
separate wire for each bit. The ability to send all of the bits of a
character simultaneously makes parallel the preferred method of
transfer when high speed is required. When such speed is not
required, or when it is not feasible to have eight individual
circuits between the devices, serial communications is used. In
serial communications, each character is sent as a series of bits one
at a time on a single wire between the systems and the systems
assemble or disassemble characters to create this stream of bits.
ASYNCHRONOUS AND SYNCHRONOUS COMMUNICATIONS
There are two basic methods of communicating characters
serially between computers: asynchronous and synchronous.
Historically, they have been implemented separately. Today, we bring
their application fields together in the personal computer
communications environment. In asynchronous communications, each
character is transmitted as an individual entity, separated from
preceding and following characters by an indeterminate period of
time. Special bits are sent preceding and following the actual data
bit of each character to synchronize the receiver with the
transmitter. These additional bits result in at least a 20 percent
overhead of non-information bits.
In synchronous communications, data are grouped and sent as a
continuous series of characters called "blocks" or "frames." Rather
than synchronize the receiver at the beginning of each character, a
special bit pattern is sent at the beginning of each frame. The
following data can be separated into characters because of the fixed
timing relationship between them. Using this method, overhead becomes
a smaller factor as block sizes increase, resulting in more efficient
use of the communications channel.
Historically, the asynchronous method has been used for
interactive terminal communications in which throughput is limited by
the capabilities of the terminal hardware and the operator's typing
speed. The synchronous method has been predominant in
computer-to-computer communications where high volume of data is the
rule and high speeds are required. As corporations rely more on
personal computers, and the need for sharing data between them and
corporate mainframes increases, the trend will be toward higher speed
for microcomputer communications, and the greatest efficiency
provided by synchronous communications. This separation of
asynchronous and synchronous communications into exclusive
environments has changed and, today, they have into one, more
comprehensive communications environment.
Computers are able to communicate a bit of information by
asserting a voltage level on a wire connected to the intended
recipient. The most basic "physical" communications interface
standards, such as RS-232, specify which voltage levels correspond to
which bit values, what type of connectors are to be used, and the
ancillary control and timing signals and their meanings. A direct
connection of this sort is well suited to short distance between
nearby systems.
MODEMS AND DATA COMMUNICATIONS
When connecting to widely separated systems, direct wired
connections are not feasible. The cost of the physical circuit is
prohibitive, and the incapability of sharing the circuit makes it
impossible to spread this cost over multiple users. most common
solution for long-distance data communication is to connect through
the public telephone network. Virtually any site can be reached, and
the cost is low compared to a direct wired connection because the
facilities are shared by many users.
The public switch telephone system represents a significant
capital investment and connects to almost every part of the world
with a steadily improving level of quality. The telephone system is
designed to carry sounds (voices) rather than voltage levels, and it
is therefore not possible to communicate through that system as one
would over a wire. Instead, it is necessary to convert the digital
voltage levels into sounds (modulation), and to convert them back to
digital signals at the other end (demodulation). The device that
accomplishes this task is known as a "modem" which is a contraction
of modulator-demodulator.
Early modems required total manual control - the human
operator had to physically dial the telephone, and switch to "data
mode" or place the telephone handset into a device that transferred
the sounds to and from the telephone. As modems came into wider use,
and use by non-technical personnel demanded simplication of the
communications process, Hayes Microcomputer Products, Inc. introduced
its innovative modem series. A Hayes modem can be directly connected
to the telephone system, eliminating the error-prone acoustic
connection. It also permits the terminal operator or computer to
control modem functions, such as dialing a number or switching from
command to data mode from the keyboard or with software. The Hayes
Standard "AT" Command Set has since become the standard method of
controlling modem functions from software. The "AT" refers to the
letters that precede each command string that allow the modem to
detect the baud rate and parity the computer is using. The Hayes
Standard "AT" Command Set has been adopted by many modem
manufacturers, thus claims such as "Hayes compatible" are frequently
seen.
The Hayes Standard "AT" Command Set specifies a method of
controlling modem functions from the attached equipment, but does not
define the means of signalling between modems. Protocols such as the
Bell 212 standard established by AT&T for use on its domestic
circuits, and the V.22 standard established by the CCITT for
international circuits ensure that the operating capabilities of the
telephone network are not exceeded, and provide a standard modulation
method so that the receiving modem can decode the transmitted
signals. These standards, among others, are implemented by Hayes so
that Hayes modems can be used to communicate with those produced by
other manufacturers that follow the same standards.
While standards such as Bell 212 and CCITT V.22 provide for
signalling in such a way that the receiver can usually decode the
sounds accurately, it is possible for the sounds to be corrupted by
interference (noise, background hiss, or pops and clicks) so that the
received data does not match the transmitted data. Noise is a
characteristic of the public switched telephone network and system
implementers have to be aware of the existence of transmission
errors.
In interactive applications, where the typist sees an echo of
the results of the communications, these problems can be corrected
immediately. Even so, the frequent introduction of errors can be
very frustrating and confusing. In file transfer applications, the
user does not see the data directly, and it is not feasible to
examine it for errors. Errors might go uncorrected, resulting in the
data being useless as received.
Additional protocols are used to ensure correctness of data
delivered by the physical devices. The next section describes the
protocols that Hayes has selected for this higher-level
error-control, and our justification for this decision.
ERROR-CORRECTING PROTOCOLS
Various protocols have been defined by computer users,
manufacturers, and international standardization bodies. These
include protocols for transferring files between microcomputers (such
as Hayes Verification and XMODEM), for block-mode communications
between terminals and mainframes, between modems, between mainframes,
to public data networks, and between public data networks.
Hayes selected the protocols it feels are best suited to
simplify migration to the communications environment that will exist
through the eighties and beyond. The following points were critical
to the decision:
- For speed, efficiency, and integration with mainframe
communications networks, synchronous communications must be
supported.
- An easy migration path for current users of asynchronous
equipment to the synchronous environment must be provided.
- Bidirectional, error-free communications must be provided
with minimal overhead or loss of performance.
- Interactive, file transfer, and other applications must be
supported. - Considering the worldwide nature of communications, and
the desire to minimize barriers to free flow of information, the
protocol must be recognized by international standards organizations
and accepted in all countries.
- Information regarding the protocol must be publicly
available so that implementation will be compatible and supportable,
further minimizing barriers to data interchange.
- The protocol must be compatible with the planned Integrated
Services Digital Networks (ISDN) of the future, again providing an
easy migration path for users.
THE X.25 PROTOCOL
The International Telegraph and Telephone Consultative
Committee (CCITT) in 1976 established Recommendation X.25 as a
standard for communications between terminals and public data
networks. X.25 uses bit-oriented synchronous communications to
provide an error-free, bidirectional communications service to
communicating applications. The CCITT enhanced X.25 in 1980 and 1984
to provide increased flexibility and to address technology such as
satellite communications, proving the intention of maintaining X.25
as a "living" specification. ISDN protocols compare closely with
X.25, making adaptation of X.25 products to ISDN a simple task. X.25
is public information, and, being an international standard, all
implementations can therefore be compatible.
For these reasons, Hayes has selected X.25 as its standard
for error-correcting protocols in future hardware and software
products. Companion standards, such as X.32 (defining the method by
which a dialed connection can be made with a network) will also be
adhered to. The protocols will be implemented both in modems (so
that existing applications can take advantage of an error-free
connection) and in software (so that new applications can take
advantage of advanced X.25 features such as multiple simultaneous
connections). Other advanced products are planned that will
implement X.25 and deliver its advantages to the communications user.
Hayes will continue to support the Hayes Verification
Protocol, XMODEM, and other accepted file-transfer protocols for the
microcomputing community. We will support and contribute to the
development of international standards in the area of file-transfer
protocols (such as the proposed FTAM standard) and intend to
implement protocols in our products.
LAYERS OF CCITT PROTOCOLS
X.25 is a "multi-layer" standard, meaning that it consists of
separate distinct parts each designed to perform a specific function.
The "packet layer" provides multiple circuits (the capability for one
terminal to communicate with several hosts simultaneously) or for
several terminals to share a single physical circuit to a network and
then on to the same or different hosts. Information at the packet
layer is transferred in "packets," which contain the user data plus
routing information necessary to get the data to its proper
destination.
The packet layer uses the services of the next lower layer,
the "link layer." The link layer establishes a single communication
path between adjacent nodes. The basic unit of information transfer
between link layer entities is the "frame." A portion of an "I"
(Information) frame contains the "packet" (if the packet layer is
being used). The link layer protocol used in X.25 is known as
"LAPB," which stands for "Link Access Procedure-Balanced." It is
"balanced" because the stations on each end of the link have equal
rights to send an appropriate frame type at any time, unlike a
"master-slave" relationship in which the slave can only transmit when
"polled" by the master.
In situations where only one connection exists between two
stations, and there is no need to route information to other stations
beyond those two, it is possible (and desirable) to use only the link
layer of X.25 (LAPB) rather than the entire packet layer, in order to
reduce protocol overhead. The new Hayes V╨series system modems,
which provide a single-channel error-control link, use only LAPB
without the overhead of the X.25 packet layer. Hayes future V-series
system products will implement the packet layer as well and provide
multiple connection capability to the user.
The upcoming Integrated Services Digital Networks (ISDN) use
a protocol very similar to LAPB for communication on the "D" channel.
This protocol, known as "LAPD," provides multiple virtual circuit
capability at the link layer, in addition to supporting the X.25
packet layer. Hayes has implemented LAPD in our prototype ISDN PC
card, and it will become increasingly important to other ISDN-related
products in the future. LAPD has also been discussed in the CCITT as
a possible standard for modem error-correction, and Hayes would
support this standard, if adopted, by designing modems which would
support both the new CCITT standard and the present V-series system
LAPB protocol.
ASYNCHRONOUS FRAMING TECHNIQUE
Although the overwhelming trend is toward synchronous
transmission, asynchronous hardware and software currently dominate
the personal computer marketplace. X.25 as it is currently defined
requires a synchronous data transmission link, and thus cannot be
used with most of the existing equipment. To address this need,
Hayes has proposed (in the CCITT and the International Standards
Organization) Asynchronous Framing Technique (AFT) which can be used
with any bit-synchronous protocol such as LAPB, LAPD, or IBM's SDLC.
AFT replaces the lowest-level framing and error detection portions of
these protocols, adapting them for operation over asynchronous
transmission links while retaining all of the formats and
error-correction and frame sequencing procedures. Higher layer
protocols, such as the X.25 packet layer, can be used without
modification on an AFT-adapted link layer.
The advantages to this approach, as opposed to adoption of an
entirely different non-X.25-compatible protocol, are many. AFT
provides an easy migration path to synchronous X.25 and can be used
as a stopgap measure until synchronous becomes more widespread. It
requires minimal modification to existing implementations, making it
easy for a vendor who already provides X.25 to add support for
asynchronous transmission. AFT takes advantage of the wide knowledge
and experience with now present in the worldwide data communications
industry. Existing certification procedures and equipment for X.25
are easily adapted to AFT. Familiarity with the protocol will
shorten the development cycle for adoption and implementation of
X.25/AFT.
Hayes will support X.25/AFT, in addition to X.25, in those
future products in which transition from asynchronous to synchronous
communications is important, including both software and advanced
hardware products. The goal is to move to the speed and efficiency
of synchronous communications as quickly as possible, and X.25/AFT
facilitates this by providing a compatible interim asynchronous
standard.
THE HAYES COMMITMENT
As the leader in communication hardware and software products
for microcomputer systems, Hayes takes seriously its responsibility
to analyze present and future requirements, and to provide products
for our customers to accomplish their communications and data
management tasks. As we grow into a world of high-speed data
communications between systems produced by various manufacturers on
many continents, the adoption of and adherence to international
standards becomes increasingly important. Our commitment to quality,
service, and value will allow our customers to continue to benefit
from Hayes products in the future.