Tuesday, April 17, 2007

A MAC parameter optimization scheme for IEEE 802.11e-based multimedia home networks

Minyoung Sung; Namhyun Yun

They calculated some delay(rigid) vs throughput(elastic) curves by means of simulation. Then they propose that the AP, depending on the accepted rigid flows and the required delay, decides the most appropriate parameter set.

It did not give me a very good impresion. It reduces the number of possible flows/bitrates. Gives examples that are surprising (required delay 1ms); It obviates the jitter... however it is very similar to the thing we want to do...

A CAPWAP Architecture for Dynamic Configuration of QoS MAC Parameters

A. Banchs∗ , G. Iannello† , P. Serrano∗ and L. Vollero

Our focus has been until now on a single-cell scenario. However, as explained in this article, the aparition of the CAPWAP standard will allow the centralized management of a number of STAs. The QoS management could then also be centralized, providing a better scenario for handling handovers. In the last part of the article, the authors propose a number of rules to determine EDCA parameters. However, it is not clear to me how the first and the second parts are linked.

Friday, April 13, 2007

Experimental Evaluation of TCP Performance and Fairness in an 802.11e Test-bed

Anthony C.H. Ng, David Malone, Douglas J. Leith

The authors build a testbed and compare the results with the ones obtained from analytical models. Additionally they present the uplink/downlink impairment performance (without any citation) and implement and validate a solution presented in other paper (this time there is a citation).

Tuesday, April 10, 2007

Understanding TCP fairness over Wireless LAN

Saar Pilosof, Ramachandran Ramjee, Danny Raz, Yuval Shavitt and Prasun Sinha

Yet another performance imparement due to the .11/TCP combination. In this case, the culprit is the AP buffer limited size. Discarding a downstream data packet has a greater impact than discarding a (down-going) ACK of an upstream flow.

The anomaly is observed, analyzed and simulated. A solution is also suggested, involving the manipulation of the TCP packets.

The 3G IMS

I got a few ideas from that book. Concerning QoS, I discovered that there are headers that explicitely detail the required bandwidth for the requested flows. I understood when the resource reservation is performed and how a S-CSCF can apply a policy to an INVITE and reject such invite according to the requested codecs.

Concerning the location/position information, I discovered that IMS messages convey the cell information. However, these headers are removed as soon as the messages are forwarded to foreign networks.