From: T Scott Saponas (ssaponas@u.washington.edu)
Date: Mon Nov 22 2004 - 07:48:20 PST
Review by T. Scott Saponas
“MACAW: A Media Access Protocol for Wireless LAN’s” examines the
performance of wireless LAN media access protocols and proposes a new
protocol. The main contribution of this paper is developing a new
wireless media access control protocol that fixes “hidden terminal” and
“exposed terminal” issues with carrier sense protocols and binary
exponential backoff’s unfair bandwidth allocation. This new protocol,
MACAW, solves the “hidden terminal” and “exposed terminal” problems in
carrier sense by using the idea of request-to-send (RTS) and
clear-to-send packets (CTS) packets from the MACA protocol. MACAW,
improves on the bandwidth allocation of MACAW by having all stations use
the backoff counters seen in packets currently in the air. This makes
it so everyone has the same backoff counter. They also fix some
contention problems by introducing a request-for-request-to-send packet
(RRTS) to allow for receiver to contend on behalf of a receiver that got
shut out in the last interval. Additionally, MACAW also adds the idea
of an ACK packet local to the wireless transmission. This is so that
packet collisions and corruptions in the wireless LAN can be detected
faster than a multiple of an RTT to the final destination.
One of the issues I had with the fair allocation mechanism was that it
is based on the receiver allocating per stream instead of per host.
They explain the need for this by describing how in most environments
there will be a base station providing access to everyone and you want
it to get much of the allocation of bandwidth since it is providing
service to so many hosts. However, I would argue you could do this
better by doing allocation per stream (like they are doing) only have
machines recognize all flows from stations besides the base station as
one flow and base station flows as multiple flows. This way one
non-base-station sender cannot hog the network by having a lot of
sending flows.
I think this paper’s contribution is quite relevant and useful. While I
wish the evaluation of their protocol involved more advanced simulation
and some results from an actual deployment, it appears promising.
Currently the Internet is edge limited. This is especially true where
the edge of the network is a wireless LAN. MACAW’s contribution is thus
quite relevant and import to today’s Internet and networks.
This archive was generated by hypermail 2.1.6 : Mon Nov 22 2004 - 07:48:27 PST