Review: MACAW. Di Niu

Review: MACAW

Reviewer: Di Niu

This paper proposes MACAW: a media access protocol for wireless
LAN's by making a good amount of modification to MACA, a media access
protocol for PARC's radio network. Some basic and important issues
regarding the design of wireless MAC protocols are discussed as MACAW
is presented in detail.

In the design of MACAW, a multiple access rather than token-based
access control is first justified. The background of its predecessor
MACA is then provided. In designing of MACAW, some important issues
have been discussed in the paper, including the RTS-CTS-Data-Ack
scheme, the DS packet, and backoff algorithms.

To solve the hidden terminal and exposed terminal problems, MACA
adopts an RTS/CTS protocol which compensates the insufficiency of
CSMA in wireless media access control. However, in MACA when data
packets suffer from losses, the error has to be recovered by the
transport layer, which incurs a significant wait. MACA solves this by
introducing an ACK packet. MACA also poses a problem of unfairness
among users. Loser of an RTS contention may continue to lose it in
the following contentions. MACAW solves this by using a back-off
counter called DS packet that is embedded in packets, so that all the
users have the same back-off counter. This scheme (the DS packet)
aided with the ACK packet also help to solve the problem of exposed
terminals which was not solved by MACA. Furthermore, a terminal may
not hear the DS so that it may not get a fair access to the media.
This is solved by further introducing an RRTS packet in MACAW.
Multicast issues have also been considered, though without
significant details.

The paper discussed in detail the design of a wireless MAC. However,
its problem is that it's too complicated. For example, what if we
lose RRTS? In fact, the problem with any complicated solutions is
that the complication itself could further introduce error.
Received on Mon Sep 25 2006 - 23:35:01 EDT