karn@THUMPER.BELLCORE.COM (Phil R. Karn) (11/29/88)
Can anyone tell me if an ARPANET PSN speaking 1822J/HDH ever generates HDLC frame aborts? (A frame abort is a series of at least seven 1-bits, without 0-bit stuffing; it causes the receiver to discard the frame without further processing). The reason I'm asking is that I have some academic curiosity about the residual bit error rate of the DDS circuit we use to talk to the ARPANET, and I wanted to rule out the PSN as a source of abort errors. The number is certainly acceptable (124 aborts over the past 7 days, out of 2.7 million frames containing 236 megabytes) but it's a lot larger than the number of CRC errors (3). One possible explanation is that the average link utilization is low, so most bit errors would occur when the link is idle. Errors that turn 0's into 1's during a flag idle sequence would show up as frame aborts. Comments? Phil