OLSR HNA support is officially up on ns-3-dev. If you don't wish to pull the latest ns-3-dev snapshot, feel free to apply the following two patches to your source tree: 1) HNA support, mixed interfaces support and helper improvement for OLSR. (Patch) 2) MID fix for mixed interface integration. (Patch) Once you've applied the patches, you may check out examples/routing/olsr-hna.cc which illustrates how to use the HNA API.

The last few weeks have been full of surprises! My research team got a poster accepted at "Design Automation and Test in Europe (DATE) 2010", which is quite a prestigious international conference. The poster covers our recent work in speeding up Logic Simulation of digital circuits using General Purpose Graphics Processing Units (GPGPUs). Although the work is still in progress, it presents some of our (promising) initial results as we explored novel approaches to simulating circuits on a GPGPU. This of course, came as a surprise to all of us because we weren't sure if our work was good enough for a conference like DATE! But oh well. :) Meanwhile, something that really hit me hard was when another one of my papers didn't get accepted into an IEEE sponsored conference. In the paper, I'd proposed a simple, straightforward and highly resource efficient countermeasure to the Collusion Attack against OLSR based Mobile Ad Hoc Networks. The simulation results (ns-3 FTW) were promising as well. What made me feel worse was that none of the reviewers rejected the paper! The paper got ratings of 5/5, 4/5 and 3/5 from each reviewer respectively, but citing some weird reasons, the paper was ultimately not accepted. And sadly, none of the reviewers gave any useful comments as well, two of them _completely_ missing the entire point behind the paper! Maybe that's how research works. Not everyone gets what you're doing, and not everyone's equally convinced about the same piece of work. Something you strongly claim might be turned down by others from the community. Similarly, something that you do may impress others more than yourself. Fun isn't it? :) There are a few more surprises but that's as much as I'd like to talk about here. :) Meanwhile, it's still unclear what I'll be doing for the next one year. Until last week, I was worrying myself to the last nerve, since I'm barely two months away from graduation. Strangely enough, I'm convinced that something good will come up for me soon. Maybe next week, maybe next month, maybe even a day after I graduate. And no, I don't remember any accident that could have rewired my brain cells to make me this optimistic. It just happened I guess! With many lessons learnt, and many habits un-learnt, seems like I've changed quite a bit over the past one month. Oh well, I'd rather be a flowing river than a puddle of stagnant water. After all, it's the latter that attracts the mosquitoes. Right? :)

Another one of my patches was committed into ns-3-dev today (Bug 788). This was just a small bug which I'd stumbled upon while working on fixing Bug 407, which is to add HNA support to ns-3's OLSR implementation. This is slightly more tricky than it seems, purely attributed to the fact that RFC 3626 does not talk about storing Netmasks in the routing table. Hence, it's upto the implementation to work around this flaw. Anyhow, I should be completing this sooner or later. In other news, I finally laid my hands on the devil because of whom I had to cancel my GRE and TOEFL, along with my plans to apply for Fall '10 MS courses at foreign universities, thus leaving me lost for a whole year! Yes, I'm talking about my damn passport which had to be re-issued because of some mistakes I never knew about. Oh well, I guess every dark cloud should have a silver lining. Right?

I always feared this day would come, and it finally has. So with a heavy heart, I bid farewell to Sun Microsystems. Thanks to all those of you who made my one and a half years at Sun a wonderful experience. To wrap it up, all I've got to say is, "until we meet again!"

My first post of the year! So here I am, whiling away the last day of my winter vacation at home. December '09 has been rather fruitful, I learnt a lot about NS-3 and even became a contributor to the project. I also managed to make a good deal of progress with my work on Security Issues in Mobile Ad hoc Networks. :)

Anyhow, the only problem with working from (my) home is the dial-up connection I'm stuck with. My laptop doesn't even have a dial-up modem and hence, I'm forced to use my trusty (maybe not) 8 year old Windows box, which has housed and nurtured generations of malware. There's this weird one that forces a click on a particular section of the screen, another one that blocks keyboard input for a while (and surely records/sends it somewhere) before allowing me to continue, and a million others. This system won't even boot from a USB which ruined my efforts to install Linux on it. And yes, I'm too lazy to go out of home and buy a blank CD/DVD. Anyhow, the dial-up is so slow, I always have a Solitaire window open to keep my fingers busy while my browser struggles to load pages over the connection that gives me blazing speeds of around 1KB/sec. This obviously implies that I've played Solitaire a LOT over the month, and this allowed me to make a few observations about it in the progress. Here goes:

1) No matter how many times you hit F2 (New Deal) under a second, the deal won't change and you'll end up with the same decks of cards. This kind of implies that their new deal function (which would be using a randomise function) is using the current system time for a seed value.

2) This observation was made possible thanks to all the malware who've worked persistently to slow down my system. Upon opening a new card from the deck on the top left (after there's another card already open on the same deck), and then hitting undo, you can see (only on my machine of course, because mine's the only slow system on the planet) the entire set of cards upto the previous one flashing by, one after the other. So my guess is, all decks are being maintained as linked lists with a pointer to the card on top and the one just before it. When you opt to undo your action, the entire linked list is traversed (and shown on screen) upto the previous one with the help of the 'previous-card' pointer. Guess they took advantage of processor speeds to hide it from the user under normal circumstances. :P

I might be wrong (my guesses usually suck), so feel free to correct me. And do let me know if you've made any observations as well. :)

Happy New Year and uhm... Happy Solitairing. :)