[ax25-layer2] ECC vs. FCS

[Samuel A. Falvo II]

On 8/3/06, Glenn Thomas <glennt at charter.net> wrote:
> Well, the main difference is that if you send them in a second
> packet, you'll incur a total expense of two packet overheads (TXD,
> TXTAIL etc) instead of one. I agree that there's no reason you
> couldn't do it, but why would you want to if you don't have too? If

I see what you're saying, and agree in general.  However, if you have
to retransmit a frame due to ARQ, you incur these overheads anyway.
Just sending the ECC data may actually be faster because it'll be a
smaller frame.

> the issue is a too small packet size, then you have the problem of
> verifying the smaller packets, which begs the question. OTOH, the
> minimum packet size for EDAC is 26 bits. Why would you want a packet
> smaller than 26 bits long?

Obviously, for such small packets, you wouldn't -- you'd just
piggyback the ECC frame.  For larger frames that actually carry
content, however, . . .

> Sending the syndrome bits twice buys nothing and costs bandwidth. If

Who said we're sending the syndrome bits twice?

This is how it works:

1.  A sends B a data packet.  NO syndrome bits are sent.
2.  Packet is received corrupted by B -- no ack.
3.  After timing out, A send syndrome bits to B -- smaller frame, on
the assumption that errors can be corrected.
4. B received syndrome bits -- ah ha!  Queue an RR to A to ack the packet.

-- 
Samuel A. Falvo II