[brite-users] Re: Huge node sets and Dijkstra's SP algorithm

Vladimir Blagojevic vladimir@cs.yorku.ca
Sun Mar 2 20:50:01 2003


Anukool,

> Sorry for the late reply.  My responses below:

NP.

> I'm not sure I understand what  you mean by "simple path between each
> two ASes using only inter-AS edges".

It means that you can reach any AS u from any other AS v by traversing
only inter AS edges.

I don't want to traverse intra AS edges while computing path from AS u to AS
v. I am currently using workaround by having router nodes in each AS
generate edges to all other router nodes in that AS using inter AS edge
type. I know it is not that clean solution but it enables me to find path
from AS u to AS v by using only inter AS edges.

Hope it makes more sense now :)


> The bottom-up approach should label routers which connect to other routers
> in a different AS as RT_BORDER.  In anycase, one can easily determine this
> by looking at the parent AS for each router.  Suppose we are given a
> router-router edge (u,v) and suppose u's parent AS is A and v's parent AS
> is B.  Then u and v are RT_BORDERS.     I will point you to the BRITE
> manual for more information on this.   You can also take a peek at the
> Java source code, as a lot of these things are commented carefully.

I'll look into the source. As things stand right now with the publicly
 available version of BRITE all nodes listed under nodes section of
generated topology file are RT_NODE type. You can verify this yourself.

I agree that the soluion is trivial - I am just stating how current
version of BRITE is behaving.

> Rocketfuel specifically maps ISP topologies by using a lot of clever
> heuristics to guide the probing.  As a result, ISP maps obtained by
> rocketfuel are more complete when compared to their skitter
> counterparts according to the rocketfuel sigcomm'02 paper.

But there is not drop in module in BRITE to use rocketfuel traces? I
suppose you are currently working on this?

Thanks,
Vladimir