<br><font size=2 face="sans-serif">Gabor,</font>
<br>
<br><font size=2 face="sans-serif">Our main design constraint when we were
planning the mesh implementation was power consumption.</font>
<br><font size=2 face="sans-serif">For a laptop mesh to be dense enough,
it is imperative that the user doesn't worry about battery consumption
when the laptop does nothing else but forward traffic for other users.</font>
<br><font size=2 face="sans-serif">Existing laptops require something around
10W to forward packets wirelessly. Our laptop can do it with ~3W
using the GX2 and a conventional laptop WiFi card, however given that we
only have a ~25Wh battery, that was still to high for our goal.</font>
<br>
<br><font size=2 face="sans-serif">Thus, we had to utilize an embedded
WiFi module with its own CPU and memory and Marvell's 88W8388 was pretty
much the only choice. To give you an idea of the power savings, we can
forward frames with less than 400mW allowing for our NiMH battery to keep
a laptop running as a wireless router for 2 days.</font>
<br><font size=2 face="sans-serif">802.11s helps keep functional separation
between the GX2 and the ARM core on the Marvell chip since it operates
on layer 2. So frames are handled by the 8388 and packets by the
GX2 (we are not running the TCP/IP stack on the 8388). This has created
some confusion because 802.11s does ad-hoc routing using MAC (layer-2)
address, however that routing is transparent to layer-3.</font>
<br>
<br><font size=2 face="sans-serif">Marvell already had a full network stack
running on the 8388 so they only had to add the 802.11s layer to it (they
are paying the people who do that development) and that's what they are
going to release.</font>
<br><font size=2 face="sans-serif">with the intention for that code
to become a module for d80211.</font>
<br>
<br>
<br><font size=2 face="sans-serif">There is a bigger question as to whether
it is practical to expect complete open source firmware on the 8388 and
my take on that is that it is, however it will take time that extends well
beyond OLPC's deadlines. </font>
<br>
<br><font size=2 face="sans-serif">Best,</font>
<br>
<br><font size=2 face="sans-serif">M.</font>
<br>
<br>
<br>
<br>
<br>
<br>
<table width=100%>
<tr valign=top>
<td width=40%><font size=1 face="sans-serif"><b>"Gabor Dolla"
<agdolla@gmail.com></b> </font>
<p><font size=1 face="sans-serif">12/31/2006 11:25 AM</font>
<td width=59%>
<table width=100%>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">To</font></div>
<td><font size=1 face="sans-serif">"Michail Bletsas" <mbletsas@laptop.org></font>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">cc</font></div>
<td><font size=1 face="sans-serif">devel@laptop.org, gnu@toad.com, jg@laptop.org,
"Rob Savoye" <rob@welcomehome.org></font>
<tr valign=top>
<td>
<div align=right><font size=1 face="sans-serif">Subject</font></div>
<td><font size=1 face="sans-serif">Re: OLPC Laptop - an open source substandard
?</font></table>
<br>
<table>
<tr valign=top>
<td>
<td></table>
<br></table>
<br>
<br>
<br><font size=3>Hi<br>
<br>
</font>
<br><font size=2 face="sans-serif">end of February. In parallel we are
working with Marvell to release the 802.11s source code under GPL.</font>
<br><font size=3><br>
Who is going to ravel out the 802.11s code from the tangled source ?<br>
I'd help if needed..<br>
<br>
What I can not see is how the Linux running on the main CPU and the Marvell
chip with its own OS and ip stack work together ? <br>
Usually the main OS handles the routing, ARP, etc...<br>
If the Marvell chip does the routing what the Linux does ? <br>
<br>
Happy New Year to you, too<br>
<br>
Gabor</font>
<br>
<br>