As many of you know, we at AirJaldi take pride in our tight collaboration with academia and leading research institutes, chief among which is the UC-Berkeley's TIER group. These collaborations have proved instrumental in keeping the technical edge and allowed AirJaldi to continuously improve our own and others' understanding of rural networking.

A notable  recent example is "JaldiMAC" - a study into a future wireless MAC-layer protocol coupled with a novel deployment methodology that came from AirJaldi and is being furhter developed by TIER. JaldiMAC holds the potential of dramatically improving the profitability of WISPs thereby driving broadband penetration further into sparsely populated rural areas currently not feasibly served. We look forward to testing JaldiMAC-based pilot deployments and evaluate its benefits while providing the developers with essential feedback from the ground in the near future.  TIER's work on an affordable, networked, intelligent solar charge controller is another example for such collaboration, one which is being predominantly specified and defined by AirJaldi with UC-Berkeley leading the actual R&D efforts and initial testing.

An important, if presently less known publicly, are the computer security-related studies done at AirJaldi over the past two years:
In collaboration with Prof. Vern Paxson's group from EECS, UC-Berkeley and ICSI, we have been exposing some interesting findings related to network security. 

The unique political sensitivities around the Tibetan community-in-exile in Dharamsala (where one of the networks we are researching is located), coupled with China's vivid cyber-security practices, led us to the assumption that an abundance of malware and especially targeted attacks would be found on the networks in Dharamsala. We were not wrong...
Early in 2009 we begun fishing for such attacks by monitoring the AirJaldi main Internet gateway, and indeed our work yielded some interesting catches.

In the spring of 2009, as the hunt for the Gh0stNet cyber-espionage network was underway by Cambridge University's Computer lab and others, we've also identified it on the network in Dharamsala and wrote a short draft report (never published):"The Gh0st in the Shell: Network Security in the Himalayas".

Through that initial work, we've identified two key points of synergy:
1. Unique features of the Dharamsala network makes it extremely valuable for security researchers and especially holds a high potential for identification of targeted, zero-day-attacks.
2. The collaboration with top-notch security researchers plays a major role in improving the security of the AirJaldi network, yielding peace of mind to it's subscribers.

We therefore tightened this collaboration based on a mutually beneficial model leveraging these points.
Equipment to monitor the network was sent from ICSI and they have sponsored additional gear purchased locally in India.

Nevertheless, as is often the case with computer security studies - researchers frequently struggle to balance the benefits of publicizing their work vs. the possible harm, through exposing novel findings to dark entities. As researchers, we naturally tend towards publication, yet in this particular case, we all agreed to place the security and privacy of AirJaldi's subscribers at the top most priority - at times resulting in limited or no publications.

Sadly, I cannot link to any publications at the moment - however, some are in the pipeline and have been submitted to various future conference proceedings, while others are being worked upon extensively and to be submitted shortly.

Linked here is Matthias Vallentin's draft, to which I've made some minor contributions, and which hopefully will get reshaped into publishable quality soon: "Quantifying Persistent Browser Cache Poisoning".
Also, a fascinating paper by Gregor Maier et al. studies security-related surfing habits and compare AirJaldi with an ISP in Europe.
I find this paper unique in the way the authors infer upon users behavior through network monitoring and I sense we may use similar methods for future studies of social nature regarding use of Internet and behavior of users. 

Finally, these collaborations inspired me towards a more general discussion of computing security in the developing world and how such studies could benefit from a multidisciplinary approach: accepted to NSDR2011 .

More as we publish some of the work mentioned above...

Yahel.
Berkeley, CA.