[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

[Freedombox-discuss] Libre Planet Followup

James, Charles, and Paul,

Here are some quick and hopefully helpful thoughts and comments on your
discussion from yesterday and today.




On Sun, 20 Mar 2011 at 09:44:56 -0400, James Vasile <james at hackervisions.org> wrote:
> Meshing is hard.  Nobody I met knows anybody who is nailing mesh
> networks.  I'm going to get all the mesh heads together soon for a
> real conversation to see if we can work towards a recommendation on
> the most promising avenue.

In order to really move forward on this front, you're going to need at least
preliminary answers to the following questions:

   1. what network scenarios are most important to you?

       (this includes sub-questions like:

         * what physical and logical media do you expect to use to build your

         * what sorts of network engineering and operations support do you
           expect to be able to rely on?

         * what are the capacity, latency, jitter, and uni/broad/multi-cast
           requirements of the protocols that you want to run on top of the

   2. what user experience do you want to offer when the mesh is misbehaving?

   3. what sorts of adversarial behavior do you expect to need to tolerate?

On Mon, 21 Mar 2011 14:51:59 -0500, Charles N Wyble <charles at knownelement.com> wrote:
> Um.... *waves*.  I guess I need to get out more. I've built a few mesh
> networks over the past year. It's not that hard (it used to be quite
> difficult, but the underlying bits have really matured).  Us mesh heads hang
> out at villagetelco.org and a few other places (olsr.org, batman.org)  :)  We
> have an annual gathering already, http://battlemesh.org/

I don't see much in the way of comprehensive performance data on display at
these sites. However, at olsr.org, I found the following graphs of node counts:


To my surprise, the timeseries for "number of nodes as a function of date"
depicted in both of these charts have some very high-frequency components.

Do you know why this might be?

(Also, are there any available graphs for other observables like total
capacity, per-node capacity, utilization, latency data, signal quality, or
route churn, either for these olsrd deployments or for the other mesh protocols
and implementations that you linked to above?)

On Mon, 21 Mar 2011 at 16:08:26 -0400, James Vasile <james at hackervisions.org> wrote:
>Can you tell me the largest mesh actually created with these mature
>bits?  I'm told there are scaling problems because the routing
>difficulty grows faster than the nodes (e.g. it's not O(n) but more like

As I wrote to Luca last month, this paper:

   "Capacity of Ad Hoc Wireless Networks", Li et al. 2001,

is a nice introduction to the subject.

(Also, if you want some second opinions, these friendly folks [1] seem like they
might be good people to talk to...)

[1]: http://datatracker.ietf.org/wg/manet/charter/

On Tue, 22 Mar 2011 06:56:39 +1030, Paul Gardner-Stephen <paul at servalproject.org> wrote:
> I am aware of ~1000 nodes, which is actually more than you need if you
> combine mesh as the "last mile" with what infrastructure happens to be
> available to gate traffic onto an internet (not necessarily the
> internet) to get it from place to place on the mesh.
> The mesh itself can be continuous and much larger, provided that each
> node only cares about a limited number of the most local nodes.
> This limits the routing complexity to be more or less constant for the
> local mesh, and O(n) at worst for the wormholes between regions on the
> mesh.

The largest mesh of which I am aware is the mesh of autonomous systems (ASes)
that forms the Internet itself. Here's a nice picture from 2009:


Unfortunately, though, it's a point-to-point mesh and it *still* requires quite
a bit more capital, care, feeding, and upkeep than the sorts of mesh that I
think you're aiming for...

On Sun, 20 Mar 2011 at 09:44:56 -0400, James Vasile <james at hackervisions.org> wrote:
> Michael Stone pointed me to <a
> href="http://www.cs.umbc.edu/~finin/home/heilmeyerCatechism.html";>Heilmeyer's
> Catechism</a>.  Those are some good questions.

There seem to be a couple of versions of the questions floating around;
however, for what it's worth, I happen to prefer the wording used in the
Wikipedia article on Heilmeier [1], which I have excerpted below:

   Heilmeier's Catechism:

   * What are you trying to do? Articulate your objectives using absolutely no

   * How is it done today, and what are the limits of current practice?

   * What's new in your approach and why do you think it will be successful?

   * Who cares?

   * If you're successful, what difference will it make?

   * What are the risks and the payoffs?

   * How much will it cost?

   * How long will it take?

   * What are the midterm and final "exams" to check for success?

(I particularly appreciate that this version places on avoiding jargon.)

[1]: http://en.wikipedia.org/wiki/George_H._Heilmeier#Heilmeier.27s_Catechism

On Sun, 20 Mar 2011 at 09:44:56 -0400, James Vasile <james at hackervisions.org> wrote:
> Scores of people expressed interest in further volunteering.  I hope
> to see them in IRC and the email discussion soon.  Rob Savoy, in
> particular, is a fascinating individual who could teach us all a thing
> or two about development.

Minor correction: Rob's last name is spelled "Savoye".

Reply to: