Flow and Error Control

Flow and Error Control

Flow control and error control at the X.25 packet layer are virtually identical in

format and procedure to flow control used for HDLC, as described in Lesson 6.

A sliding-window protocol is used. Each data packet includes a send sequence

number, P(S), and a receive sequence number, P(R). As a default, 3-bit sequence

numbers are used. Optionally, a DTE may request, via the user-facility mechanism,

the use of extended 7-bit sequence numbers. As Figure 9.19 indicates, for 3-bit

sequence numbers, the third and fourth bits of all data and control packets are 01;

for 7-bit sequence numbers, the bits are 10.

P(S) is assigned by the DTE on outgoing packets on a virtual circuit basis; that

is, the P(S) of each new outgoing data packet on a virtual circuit is one more than

that of the preceding packet, modulo 8 or modulo 128. P(R) contains the number of

the next packet expected from the other side of a virtual circuit; this provides for

piggybacked acknowledgment. If one side has no data to send, it may acknowledge

incoming packets with the Receive-Ready (RR) and Receive-not-Ready (RNR)

control packets, with the same meaning as for HDLC. The default window size is 2,

but it may be set as high as 7 for 3-bit sequence numbers and as high as 127 for

7-bit sequence numbers.

Acknowledgment (in the form of the P(R) field in the data, RR, or RNR

packet), and hence flow control, may have either local or end-to-end significance,

based on the setting of the D bit. When D = 0, (the usual case), acknowledgment is

exercised between the DTE and the network. This communication is used by the

local DCE and/or the network to acknowledge receipt of packets and to control the

flow from the DTE into the network. When D = 1, acknowledgments come from

the remote DTE.

The basic form of error control is go-back-N ARQ. Negative acknowledgment

is in the form of a Reject (REJ) control packet. If a node receives a negative

acknowledgment, it will retransmit the specified packet and all subsequent packets.

Packet Sequences

X.25 provides the capability, called a complete packet sequence, to identify a contiguous

sequence of data packets. This feature has several uses. One important use

is by internetworking protocols (described in Part IV) to allow longer blocks of data

to be sent across a network with a smaller packet-size restriction without losing the

integrity of the block.

To specify this mechanism, X.25 defines two types of packets: A packets and

B packets. An A packet is one in which the M bit is set to 1, the D bit is set to 0, and

the packet is full (equal to the maximum allowable packet length). A B packet is any

packet that is not an A packet. A complete packet sequence consists of zero or more

A packets followed by a B packet. The network may combine this sequence to

make a larger packet. The network may also segment a B packet into smaller packets

to produce a complete packet sequence.

The way in which the B packet is handled depends on the setting of the M and

D bits. If D = 1, an end-to-end acknowledgment is sent by the receiving DTE to the

sending DTE. This is, in effect, an acknowledgment of the entire complete packet

sequence. If M = 1, there are additional complete packet sequences to follow. This

enables the formation of subsequences as part of a larger sequence, so that end-toend

acknowledgment can occur before the end of the larger sequence.

Figure 9.21 shows examples of these concepts. It is the responsibility of the