Re: iSCSI: Markers

[Jonathan Stone <jonathan@dsg.stanford.edu>]

....This may be a good point to observe that the original
implementation of COBS was for a wireless radio link which struggled
to deliver 19.2kbps of goodput to any individual end-station.

I believe the first two (or three, depending on how one counts the
appletalk implementation) of COBS all found the change in buffer sizes
diffiult to integrate into other operations, and they all implemented
COBS encoding/decoding as a copy between encoded and decoded buffers,
not as an in-place operation.

Even if a COWS data-touching loop is folded into another data-reading
loop, COWS encoding/decoding still dirties all the cache lines being
encoded or decoded, and will therefore has the same overhead as an
additional copy.

Performing COBS encoding/decoding in an outbouard accelerator also
makes it impossible for those of us who don't trust our I/O busses to
perform a `trust, but verify' integrity check: e.g., recomputing the
TCP/IP checksum over the data stream as the HBA-cum-NIC deposits it
in host memory.

Indeed, the original COBS implementation(s) provided only framing;
COBS framing was wrapped outside an end-to-end integrity check.

I beleive Stuart is still on vacation, but I can ask him to comment on
these issues when he gets back, if that would be worthwhile.