TCP-group 1991
[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
MAC sublayer protocol for TCP/IP
- To: tcp-group@ucsd.edu
- Subject: MAC sublayer protocol for TCP/IP
- From: johnson@nisc.jvnc.net (Steven L. Johnson)
- Date: Fri, 11 Jan 91 16:32:02 EST
With the growing possibility for a 56K TCP/IP backbone (and/or network)
in the area, I've been giving some preliminary thought to improving the
throughput of medium and high speed RF packet links, especially under
heavy load.
I'm assuming that the configuration would be a string or mesh of hubs
with omnidirectional antennae talking to leaf nodes with gain antennae
(for the most part). It is my understanding that under these conditions
collision detection is impossible and carrier sense is of little value
since the only other station that a leaf node can reliably talk to is
a hub. I imagine that for effecient communication that the hub should
arbitrate access for its channel (i.e. freq within a limited geographic
area).
Do any of the current RF networking protocols or programs provide this
type of Medium Access Control (MAC)? If so, where could I find descriptions
of the protocols?
If not, is there interest in developing a new MAC sublayer protocol that
would improve the reliability and effeciency of TCP/IP? I envision a
"token star" protocol, similar to IEEE 802.4 "token bus" but MUCH simpler
because the majority of the protocol runs in a centralized hub and the
concern with realtime response is unnecessary. I anticipate congestion
on certain moderate speed links to be high, and a token based arbitration
scheme would keep the network from collapsing under heavy load.
-Steve