facility components involved in the data transfer. Flow control must operate at several
peer layers of protocols, as will be discussed later.
Error control. Error control is a technique that permits recovery of lost or errored
packets (frames, blocks). There are four possible functions involved in error control:
1. Numbering of packets (frames, blocks) (e.g., missing packet).
2. Incomplete octets (a bit sequence does not carry the proper number of bits, in this
case 8 bits).
3. Error detection. Usually includes error correction (see Section 10.5.3).
4. Acknowledgment of one, several, or a predetermined string of packets (blocks,
frames). Acknowledgment may be carried out by returning to the source (or node)
the send sequence number as a receive sequence number.
Open Systems Interconnection (OSI)
Rationale and Overview of OSI
. Data communication systems can be
very diverse and complex. These systems involve elaborate software that must run on
equipment having ever-increasing processing requirements. Under these conditions, it is
desirable to ensure maximum independence among the various software and hardware
elements of a system for two reasons:
To facilitate intercommunication among disparate elements.
To eliminate the "ripple effect" when there is a modification to one software element
that may affect all elements.
The ISO set about to make this data intercommunication problem more manageable. It
developed its famous OSI reference model (Ref. 20). Instead of trying to solve the global
dilemma, it decomposed the problem into more manageable parts. This provided standard-
setting agencies with an architecture that defines communication tasks. The OSI model
provides the basis for connecting open systems for distributed applications processing.
The term open denotes the ability of any two systems conforming to the reference model
and associated standards to interconnect. OSI thus provides a common groundwork for
the development of families of standards permitting data assets to communicate.
ISO broke data communications down into seven areas or layers, arranged vertically
starting at the bottom with layer 1, the input/output ports of a data device. The OSI
reference model is shown in Figure 10.20. It takes at least two to communicate. Thus we
consider the model in twos, one entity to the left in the figure and one peer entity to the
right. ISO and the ITU-T organization use the term peers. Peers are corresponding entities
on either side of Figure 10.20. A peer on one side of the system (system A) communicates
with its peer on the other side (system B) by means of a common protocol. For example,
the transport layer of system A communicates with its peer transport layer at system B. It
is important to note that there is no direct communication between peer layers except at
the physical layer (layer 1). That is, above the physical layer, each protocol entity sends
data down to the next lower layer, and so on to the physical layer, then across and up to
its peer on the other side. Even the physical layer may not be directly connected to its
peer on the other side of the "connection" such as in packet communications. This we
call connectionless service when no physical connection is set up.
However, peer layers
must share a common protocol in order to communicate.
Connectionless service is a type of delivery service that treats each packet, datagram, or frame as a separate
entity containing the source and destination address. An analogy in everyday life is the postal service. We put
a letter in the mail and we have no idea how it is routed to its destination. The address on the letter serves to
route the letter.