Mesh Network FAQ

  This page is monitored by the OLPC team.

These are questions about the Mesh Network used by the XO laptop. A related page is Mesh Network Details.

How many servers per school?

How will the school servers in one school be connected, and how many users each are they expected to support?

The ratio of students to school servers is currently planned to be no more than 100:1. The interconnection between servers will be best possible: Cat3/5 or powerline is strongly recommended, with a fallback to wireless if necessary.

Each server will support at least two WiFi access points (the Marvell modules), with up to five or six access points possible. The optimum deployment scenario probably provides two or three meshes per school (on channels 11, 1, and 6). Each server provides access points on two or more meshes (i.e. there are multiple servers/access points on each mesh.)

How is the mesh channel for a laptop to join chosen?

You have to spend some time on every channel and estimate how heavy the traffic is from the RREQ and RREP packets.

This is tricky, as these packets are not sent to the host. One way to determine this indirectly is to examine the forwarding table: when RREQs are received, reverse routes are created. Alternatively, one could monitor the forwarding statistics of the mesh interface (ethtool -S msh0). These are all "passive" detection methods: will not work if all the hosts in the mesh are silent.

Yet another alternative would be to run daemons attached to the mesh interface, such as http://www.cozybit.com/projects/lsmesh

This is now answered by the Mesh Network Details

How can we determine whether a channel has other active mesh users?

There are no beacons currently (although they are in the implementation plan) so you really have to listen for mesh traffic.

How many radios can one mesh channel support?

How will a laptop decide to join another mesh channel if the current one gets too busy?

Is there an answer ?

Will physically adjacent laptops be on the same mesh ?

Is it possible that two children sitting next to one another are on different channels and therefore cannot "see" each other on the net?

One of the roles of the School server will be to bridge between mesh clouds running on different channels. How do you decide what mesh (channel) to join was a previous question.

What about WiFi encryption ?

Does the mesh part of the firmware use the same encryption settings as the b interface? Do we care about making a 'private mesh' with WPA-PSK or WEP or something like that?

Yes, the mesh uses the same encryption.

Down the road, we care, and will probably use WPA-PSK.--Michalis

I will argue that link layer encryption is the wrong place to protect secrets. If an application handles private or sensitive data, it should apply encryption at that time (e.g. HTTPS). My concern is the management overhead of the authentication server for WPA-PSK. The ability of devices other than XO laptops to join the school network will be supported. --Wad 00:21, 22 February 2007 (EST)

How are school servers (network gateways) discovered?

Will servers send out some sort of announcements to allow the laptops to find them automatically, or must we cache a DNS name or IP address for the server?

They will act as gateways and respond to RREQs for a reserved anycast address.--Michalis

This is two different questions. At the layer 2 networking level, the laptop is looking for a default gateway. This will be supplied by a mesh network configuration server running on all school servers, which responds to a request for a particular anycast address with its IP address. The laptop will then use DHCP to request an IP address, netmask, domain name servers, and default domain.

At the services level, each laptop will be associated with a particular school server that contains a student's journal and backups, through a mechanism which is still being defined.--Wad

CUWin

Why not go for an open-source mesh network?

Have you considered following the CUWin Mesh protocol instead of 802.11s? 802.11s is mostly a very complex compatibility standard for mesh which is more suited towards commercial metro coverage of a city. As a wireless engineer who has been part of the IEEE Standards development process, I can tell you that 802.11s compatible equipment is probably not the right approach.

CUWin on the other hand, is a decentralized, community-owned mesh network that fosters democratic cultures and local content. The connections are not made to funnel up to an ISP, but instead across town to other CUWin nodes. Each node is encouraged to host its own mail sever, host it's own local content, etc. The equipment is relatively very inexpensive (see below)

Kgrr 11:43, 22 March 2007 (EDT)

Meraki

Meraki Minis and Outdoor are inexpensive have been built to run CUWin - $50-$100. It would make sense to be compatible with this hardware since it's inexpensive and open source. This is an inexpensive way to build a network across a town. This would very much needed in order to connect a communities of laptops together.Kgrr 11:43, 22 March 2007 (EDT)