[e2e] Regarding use of Reed-Solomon code in wireless networks

[Detlef Bosau]

Am 06.05.2015 um 18:04 schrieb Debarshi Sanyal:
>
> I understand there is a fundamental problem here: corruption or
> collision. Independent of the cause, the receiver will attempt to
> repair the packet. If it fails, the sender must transmit a fresh nonce.

Exactly.

And different from, e.g., Ethernet, you must mot switch off the
mechanism, otherwise your packet may never reach the medium - and the
receiver.

> The unfortunate downside is that if packets are lost due to collisions
> and the backoff mechanism of MAC is switched off, collisions will only
> increase.

Yes. Does WiFi have a backoff mechanism? I simply don't have it in mind.
It would be most reasonable to desynchronise competing senders.
(If you happen to know the details, please let me know. )
>
> The wormhole is assumed to be passive: it will not corrupt the packet;
> it will only tunnel it over a longer distance.

I see. And your idea is, to observe the time a packet travels and to
assess, whether this unduly long.

> This assumption effectively means only noise can corrupt the packet.

My objection is, that a normal packet, which is not redirected over a
wormhole link may see large transport times as well, when. e.g., there
is high load on a net or severe noise. So the problem is to find a
significant limit to discriminate "noise delay" or "collision delay"
from "wormhole delay".

-- 
------------------------------------------------------------------
Detlef Bosau
Galileistraße 30   
70565 Stuttgart                            Tel.:   +49 711 5208031
                                           mobile: +49 172 6819937
                                           skype:     detlef.bosau
                                           ICQ:          566129673
detlef.bosau at web.de                     http://www.detlef-bosau.de