You are here: Computer Science > Staff > cheryl > MANET

2-26-2009

Working towards data collection on network:

Installed tcpdump v3.9.5 from http://maemo.org/downloads/product/OS2007/tcpdump/ on nokia n810

Installed wireshark 1.0.6 on PPC mac wrote how-to (the instructions provided are incomplete)

At this point we have a heterogeneous MANET (nokia N810 internet tablets, ubuntu laptop, macbook (not fully working yet. OLSR recieves but does not send HELLO and TC packets, BATMAN not yet ported), windows based eepc, plans for itouch) that can run OLSR and BATMAN routing protocols (not at the same time). An asterisk VoIP server provides SIP based VoIP over the network. Early experiments show poor performance of VoIP (unsurprising) and OLSR unable to support a SIP session over a topology change (BATMAN was able to support the SIP session over a topology change). These are very early tests and we are now working to study the network performance at various layers.

The first step is getting the network data collection set up. Wireshark is not able to capture 802.11 data on older Airport Extreme cards (Mac Book does not work, 2008 iMac does work). KisMac is able to capture 802.11 and can be saved as a Dumplog and then viewed/analyzed in wireshark.

Plans:

• MANET visualizer - layer by layer visualization of MANET performance
• BATMAN ns-2 implementation
• cooperative routing stategy (experimental compared to current MANIAC approaches) in ns-2 followed by validation on experimental live network.

6-4-2009

Bushfire sensing seed grant

Won seed funding from Universities Australia to develop linkage grant on tactical scatterable bushfire monitoring. Setting up separate site for collaboration.

Collaborative routing

Setting up iNSpect code on Mac

Possible industry partnerships with teaching

Met Derek Munneke from 2moro mobile. e-introduced to sasha and peta with Mobile Enterprise Growth Alliance (MEGA). Follow-up potential link for MCI students and MWN students interested in pursuing mobile applications.

Prof Cotz visit

Organise workshop

L&T Performance fund

Has anyone done any systematic survey of the effectiveness of feedback to students with an aim of providing feed back into teaching practice? It seems that feedback is an area that we (as a school, faculty and even a University) do not score highly in. I can clearly show how to score high or low in that area (MWN); but I'm interested in lower cost options for feedback that still maintain the *detailed personal* feedback that results in the higher level of student satisfaction with feedback. There are generic suggestions such as group feedback, which in my experience (and evidenced in SELTs) is not satisfactory to students (again, difficulty in relating to their own approach). I think the *way* in which feedback is given may not be the issue.

There seems to be a widespread acceptance that knowing *what* you did wrong helps you to do better the next time. Or that students are able to take an exemplar and work out how their own work differs from that exemplar. I actually think this is not quite correct. I think knowing *why* you did wrong helps you to improve. Approach/methodology often trip students up. I'd be interested in looking at student satisfaction with feedback, the correlation (if any) with improvement and the techniques used. This might tie in with the e-learning side too as I have a few ideas on how such feedback might be done more efficiently through myUni/Moodle.

Cisco partnership

3-3-2009

Getting OLSR running on ns-2. setdest.cc incorrectly uses god as a variable, should use God not $god_

existing patch file is out of date. corrected patch entries. communication is working but mobility is not. I expect this is configuration rather than code - confirm tomorrow and create/publish new patch

Met Linda Davis and Ying from ITR. Briefly discussed MANET monitoring tools.

5-3-2009

Fixed configuration and created clean patch for OLSR. Added to how-tos section.

2-27-2009

will need to port USB tunnel code to run BATMAN on OS X (add to projects)