RE: [manet] Re: Comments Solicitation on a new draft for MANETs

Inline [RRR].

From: manet-bounces at ietf.org [mailto:manet-bounces at ietf.org] On Behalf Of Nour-eddine.KETTAF at uha.fr
Sent: Monday, June 26, 2006 6:33 AM
To: Alicia Triviño
Cc: manet at ietf.org
Subject: [manet] Re: Comments Solicitation on a new draft for MANETs

Hi Alicia,

Thank you for your interest :)

>Dear Noureddine:

>I have read your draft and, as you suggested, I enclose my comments.

>In my opinion, the main disadvantage of your routing protocol is related to
>the measurement of the QoS parameters. I mean, how can you know the channel
>bandwidth in wireless networks? As far as I know, this parameter varies
>with the movement of nodes. If the reservation of the resources is
>performed, how can you assure it is maintained? Additionally, the
>interferences of nodes should be taken into account.

The proposed protocol deals mainly with the IP layer, however, the problem of bandwidth estimation is related to the MAC layer which is not actually the scope of MANET WG. In our draft we assumed that the estimation is based on an efficient and reliable mechanism to avoid the problems related to hidden and exposed nodes. On the other hand, we provided a solution for this problem (available bandwidth estimation, mobility, interferences, etc.) I will communicate it to you asap.

[RRR] Alicia is right that the throughput capacity (bandwidth) of the MANET also depends on the speeds of the mobile nodes (published simulation results have confirmed this fact). Now, if the physical/MAC layer bandwidth changes due to the speeds of the mobile nodes, the IP layer bandwidth will also change. So, the bandwidth estimation between the mobile nodes will be a function of speeds of the mobile nodes, multi-path interferences, and many other factors. The speed of the mobile nodes is another serious contributor in effecting the throughput capacity in MANET. In technical term, people also provide requirements in MANET like mobile through capacities in different speeds of the mobile nodes.

[RRR] So, we expect that the problem that you have solved for the bandwidth estimation in the IP layer will include the effects of speeds of the mobile nodes, multi-path interferences, and other factors. Is this the correct assumption.

[RRR] Is it possible to send your paper to this list just to know how you addressed the problem of estimation of the bandwidth in MANET?