Reiner Ludwig wrote:
> Berthold,
>
> without having read your paper it seems that there is still some
> controversy when it comes to WLAN links:
I think there is no controversity, because both referenced papers have
investigated different WLAN technologies having different MAC and
link-layer protocols.
We have used the proprietary WaveLan and Arlan. Both technologies use
CSMA/CA. WaveLan has no error control mechanism, whereas Arlan uses a
built-in send-and-wait protocol.
Authors of the referenced papers only used WaveLAN. In addition Eckardt
and Steenkiste implemented a send-and-wait protocol on top of the MAC,
but switched off WaveLan's CSMA/CA backoff algorithm to emulate the PCF
of IEEE 802.11. In fact, they used a polling mechanism, which requests
each mobile station to send data. This reduces the packet delay and its
variation, which may lead to a different TCP performance. In our
measurements we took Arlan unmodified (send-and-wait as link-layer
protocol and complete CSMA/CA for access). This is closer to the IEEE
802.11 DCF.
Berthold
>
>
> - Eckhardt D. A., Steenkiste P., Improving Wireless LAN Performance
> via Adaptive Local Error Control, In Proceedings of IEEE ICNP 98.
>
> also look at TCP over a reliable WLAN link and did not find any
> spurious timeouts. Only when they drop 100 consecutive packets out of
> every 500 (!) do they capture spurious timeouts.
>
> - Balakrishnan H., Padmanabhan V., Seshan S., Katz R. H., A Comparison
> of Mechanisms for Improving TCP Performance over Wireless Links, In
> Proceedings of ACM SIGCOMM 96.
>
> also look at WLAN and also don't find any spurious timeouts.
>
> Are you maybe looking at a different WLAN implementation?
>
> ///Reiner
>
>
> At 05:18 AM 2/3/99 , Berthold Rathke wrote:
> >To contribute to the discussion of TCP performance over reliable link
>
> >protocols
> >in a wireless LAN environment we will present our measurements.
> >
> >The results show, that TCP performs spurious retransmissions
> degrading TCP
> >throughput.
> >More informations can be found at
> >http://www-tkn.ee.tu-berlin.de/bibl/ours/TKN-01BR.ps.gz
> >
> >Berthold
> >
> >Technical University Berlin
> >Telecommunication Networks Group
> >Sekr. FT5-2
> >Einsteinufer 25
> >10587 Berlin
> >Germany
> >
This archive was generated by hypermail 2b29 : Mon Jan 28 2002 - 09:12:20 EST