Hi,<br><br>First of all, thanks for your explanations.<br><br><div class="gmail_quote">On Wed, Sep 8, 2010 at 3:16 AM, Juliusz Chroboczek <span dir="ltr"><<a href="mailto:Juliusz.Chroboczek@pps.jussieu.fr">Juliusz.Chroboczek@pps.jussieu.fr</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Hi Mason,<br>
<div class="im"><br>
> I would like to test babel protocol which is based on a distance vector<br>
> routing protocol.<br>
<br>
</div>Cool.<br>
<div class="im"><br>
> In case of a wired network, e.g., RIP and OSPF, link state routing protocol<br>
> is generally known to be better than the distance vector routing one due to<br>
> fast convergence and loop-free property.<br>
<br>
</div>That is common folklore, but it's not true.<br>
<br>
- link-state is not intrinsically loop-free[1], it's just OSPF and<br>
IS-IS that eliminate loops in a timely manner due to the reliable<br>
flooding algorithm they use;<br>
- while naïve distance-vector (GGP, RIP) generates loops, there exist<br>
distance-vector protocols that are mostly loop-free (EIGRP, Babel,<br>
DSDV).<br>
<br>
You may be interested in having a look at the slides of a lecture I've<br>
just given on the subject,<br>
<br>
<a href="http://www.pps.jussieu.fr/%7Ejch/enseignement/tbilisi2010.pdf" target="_blank">http://www.pps.jussieu.fr/~jch/enseignement/tbilisi2010.pdf</a><br>
<div class="im"><br>
> In a wireless network, AODV and DSR has been working better than DSDV<br>
> based on the distance vector one.<br>
<br>
</div>AODV is a distance vector protocol. DSDV's poor performance is not<br>
intrinsic to the protocol, it's due to having to wait for a periodic<br>
refresh after each starvation event (slide 46).<br></blockquote><div><br><span style="color: rgb(51, 51, 255);"><span style="color: rgb(102, 0, 0);">[msc] I've reviewed those documents and a paper. Thanks again for providing a recent paper.</span></span><br style="color: rgb(51, 51, 255);">
<span style="color: rgb(51, 51, 255);"> </span><br></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div class="im"><br>
> In addition, I also reviewed a paper which compared with various metrics<br>
> such as babel, batman, and olsr. Babel and batman outperformed olsr.<br>
<br>
</div>OLSR is a very naïve variant of link-state -- it uses unreliable<br>
flooding. I'm actually surprised it works as well as it does.<br>
<br>
> Incredible.<br>
<br>
Heh.<br>
<div class="im"><br>
> In general, for running manet's routing protocols, each wireless system has<br>
> to be set to an ad-hoc mode instead of AP or STA. In case of babel, all of<br>
> the wireless system has to be set to ad-hoc mode in a mandatory manner?<br>
<br>
> Unlike links between AP and STA, some wlan chipset working as ad-hoc mode<br>
> has not supported full throughput due to some firmware errors. In a worst<br>
> case, each wireless system cannot join the same cell infrequently. That' why<br>
> I want to avoid using ad-hoc mode due to the existence of unpredictable<br>
> characteristics. Instead of using an ad-hoc mode between wireless systems,<br>
> can I still use the babel between AP and STA (Client), on which the babel is<br>
> running separately?<br>
<br>
</div>Babel only requires the ability to transmit IPv6 multicast on a link;<br>
a Wifi link in managed mode should be fine.<br>
<br>
The issue is different: in managed mode, nodes cannot associate<br>
arbitrarily: a STA cannot associate with a STA, an AP cannot associate<br>
with an AP, and a STA can only associate with one AP at a time.<br>
</blockquote><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
But if your network topology is such that this is not a limitation, by<br>
all means, do try managed mode.<br></blockquote><div><br><span style="color: rgb(51, 51, 255);"><span style="color: rgb(102, 0, 0);">[msc] Like you commented, I've already configured a simple topology,<br><br>A-----B-----C<br>
<br>Each node has dual modes of master(AP) and sta(managed) at the same time. It is working so well. But I will have a plan to increase the size of the network and test it in details.<br><br>In addition, what about the future of Babel's recent IETF draft status? Is it submitted to the IETF WG?<br>
<br></span></span></div><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<br>
Regards,<br>
<font color="#888888"><br>
Juliusz<br>
<br>
</font></blockquote></div><br>Myongsu<br><br>