=[ x25-box.com ]=

automatic. In such cases it is difficult to say that the error rate is
unacceptable even though it is very bad. Like a crackling telephone line
it is up to the users to decide what is desirable and acceptable.

At layer three in the example, each data packet has been acknowledged
by an RR. This the normal rate of acknowledgement. Comparing the
number of Data packets with the number of Info frames, we can see that
coming from the network most of the Info frames carry real data (300
out of 360); or layer three RRs (50 out of 360). This leaves ten Info
frames carrying other types of packet.

Data travelling from the PAD has resulted in 50 packets and 300 RRs.
In this case however there are 50 additional packets carried by layer two
Info frames. What these packets might be was shown in Chapter 4;
however, they represent a significant number compared to the number of
Data packets (50 compared to 300) and it can be seen that there is cause
for worry and further detailed investigation.

5.3 More examples of statistics

Figure 5.4 shows an example of some statistics. These figures show a
lack of any information transfer but a large number of frames out of the
unit. This is due to the unit transmitting DM frames to try and establish
the link. Such transmissions are normal since it is what the protocol
requires, but it is fairly evident that the other end is not going to respond.
In this case it may be sensible to stop the operation of the port by a
command to the software, and the processor resources can then be used
more effectively.

Fig. 5.4 Example statistics

Figure 5.5 shows some higher numbers. Considering layer three first,
the number of Data packets out is much the same as the number of RR
packets in, and thus the packets are separate and individually
acknowledged. Only around half of the incoming Data packets are
acknowledged, so these are probably closer together.

The layer three totals indicate a few packets not yet accounted for.
These may be due to the following packet types: Call, Call Accept, Clear,
Clear Confirm, and Reset and Restart procedures.

The number of Info frames both incoming and outgoing is slightly
greater than the total number of packets. This is due to errors in
transmission, which is confirmed by the presence of Reject frames.

The layer two RRs indicate that many of the frames were acknowledged
by piggybacking, or else were sent sufficiently quickly to not require

НАЗАД