jwb@monu6.cc.monash.edu.au (Jim Breen) (02/01/91)
A question about the REJ timer in 802.2. [Background: I have been setting the partial implementation of HDLC (ABM) as a class exercise. I have used the 802.2 Type 2 spec because it has excellent state transition tables for the protocol.] A question has arisen about the comparative settings of the acknowledgement and reject timers. As far as I can see, 802.2 is the only HDLC variant which is explicit about about having a reject timer. IS7809 (HDLC) caters for one by saying that any command which is expecting a response and does not receive it may be repeated. 1988 LAPB is similarly vague. In our simulations, it appears that the reject timer operation has the potential to collide with the operation of the acknowledment timer if they are set to the same or similar values, i.e. the RR(P=1)/RR(F=1) pair followed by a string of retransmitted I frames get nicely stuffed by the arrival of a quite unnecessary REJ. If the reject timer is set to a large enough value to avoid this, it never fires as the ack timer always leads to the retransmission. I realise the problems I am talking about are relatively rare in real life as the reject timer is only involved if the *first* I frame of retransmission is corrupted. Can anyone throw some light on this? What sorts of timer values are used in practical implementations of 802.2? -- Jim Breen AARNet:jwb@monu6.cc.monash.edu.au Department of Robotics & Digital Technology. Monash University. PO Box 197 Caulfield East VIC 3145 Australia (ph) +61 3 573 2552 (fax) +61 3 573 2745 JIS:$B%8%`!!%V%j!<%s(J