Connectivity graph

4 views
Skip to first unread message

ritesh

unread,
Nov 9, 2008, 2:28:08 PM11/9/08
to UCR Testbed
Hi,

Great work on the testbed! I was wondering if there is any way to get
a connectivity graph of your testbed. I had a paper on wormhole
attacks in Infocom 2007, and I am trying to add some more simulations
with realistic connectivity to that paper to extend the work. Of
course, the definition of a link is important, but for my purpose any
definition, based on packet reception rate or signal strength or SNR
will do.

Thanks,

Ritesh
Stony Brook University

Konstantinos Pelechrinis

unread,
Nov 10, 2008, 1:39:35 PM11/10/08
to ritesh, UCR Testbed
Hi,

if we understand correctly you need a list of the neighbors for every node
on our testbed. A list in the form:

node ID | IDs of neighbor nodes

If this is what you need we are more than happy to help you out with this.
The way we collect this graph is by utilizing ICMP_ECHO_REQUEST and REPLY
messages. More specific, for every node we ping all the rest and we are
awaiting to see which of them reply back. By this way we can collect the
connectivity graph of our testbed.

Let us know if this is what you need.

Thanks a lot,
Kostas

Ritesh Maheshwari

unread,
Nov 10, 2008, 2:01:44 PM11/10/08
to ucr-t...@googlegroups.com
Hi Kostas,

Thanks for replying.


> if we understand correctly you need a list of the neighbors for every
> node on our testbed. A list in the form:
>
> node ID | IDs of neighbor nodes

Yes, I need exactly that.

>
> If this is what you need we are more than happy to help you out with
> this. The way we collect this graph is by utilizing ICMP_ECHO_REQUEST
> and REPLY messages. More specific, for every node we ping all the
> rest and we are awaiting to see which of them reply back. By this way
> we can collect the connectivity graph of our testbed.

This would work for my current need. So thanks a lot! :)

But if you want to be rigorous about what it means to have connectivity
graph, what it meant by a 'link' needs to be defined. This definition is
usually quite fluid and really depends on the user. So if you are doing
measurements by ping, then a summary of number of packets sent and
number of packets received on each link will be very useful.

Thanks,
Ritesh

Konstantinos Pelechrinis

unread,
Nov 11, 2008, 5:22:10 PM11/11/08
to ucr-t...@googlegroups.com, rit...@gmail.com
Hi Ritesh,

attached you can find a directory with a set of files. Every file is
named by neighbors.X. Each one of these files contains the IDs of the
neighbors of the node with ID X.

The test was basically a ping testing as mentioned at the previous e-mail.
For every node whose 1-hop neighbors we want to find, we send out 500
ICMP_ECHO_REQUEST to all of the rest nodes. Whenever we get back more
than 150 replies (i.e. 30% PDeliveryR) the link is considered feasible.

The same tests were conducted 3 times in order to verify the results.
Moreover this is the 802.11g connectivity of our testbed.

I hope these data are helpful for you.

Thanks,
Kostas


--
http://www.cs.ucr.edu/~kpele

Graph.tar.gz

Ritesh Maheshwari

unread,
Nov 12, 2008, 12:31:11 AM11/12/08
to ucr-t...@googlegroups.com
Thanks Kostas. Can I also get the (relative) location of each node in the testbed? Something in the form of:

Node-id X Y

Thanks!
-Ritesh

Konstantinos Pelechrinis

unread,
Nov 12, 2008, 2:44:10 AM11/12/08
to ucr-t...@googlegroups.com, rit...@gmail.com
You can find a map of our testbed with the relative positions of the nodes
at http://networks.cs.ucr.edu/testbed/deployment.htm . The nodes ID are
the same as the ones I send you so you can make the mapping.

Thanks,
Kostas


--
http://www.cs.ucr.edu/~kpele

Ritesh Maheshwari

unread,
Nov 12, 2008, 6:38:54 PM11/12/08
to ucr-t...@googlegroups.com
Hi,

Thanks for the pointer! I will let you know how it goes.

-Ritesh

Reply all
Reply to author
Forward
0 new messages