Gauge Newsletter September 2015 | Page 14

components. In hardware, you can use a all. So I’ll give you a simple idea about tion to the attacks, obviously the major Field Programmable Gate Array (FPGA) or that. An attack like somebody is trying to focus of these research works is coming Graphic Processing Unit (GPU), and there break in. So the idea of the side channel up with how to make the system safer are different ways of achieving it. There attack is to break in and find the secret against these attacks. They are called are different projects where we are work- key that is used by a cryptosystem of counter measures. ing on different aspects of the accelera- a device. So in encryption, if you want tion. There are projects where we worked find the secret key you can do brute And Sir about your publications? on how to improve the algorithm and to force attacks and that’s the simple way You can find them on my homepage as make it faster by using GPU. We can put of doing it. But if you have a 32bit key, there is a list of about 100 of them. it parallel and make it faster. And there then you have to try for 232 combina- are places where we use FPGA to make tions, and it is not possible because it computation faster. So I am interested in takes too long time. So that’s the normal all of them and we are working on differ- way. But side channel is a different way ent projects. I have done around 25 to 30 of breaking. Instead of doing brute-force, projects on High-Performance Computing we take some measurements of a side- with students both undergraduate and channel and then we do some computa- postgraduate level. tion. So when the computation is happening, there are other things like power So sir, what did we get from those projects for our departments? Yeah, we got something like for instance I hope you know that we got a GPU Research Center from NVIDIA. It is a good addition to ESCAL. What this means is that we had to do a lot of high performance computing research and also we had to use GPU for such works. We submitted a proposal for this lab to NVIDIA. We had to tell them what we have done, and we had to present publications we made and presentations we have done on those things. We had to put them all and make a proposal. So they go through it and decide whether to give the status to our University. So to achieve it, we had dissipation happens, or time consumed. If you take an algorithm or a program, depending on the data you put, it takes different amount of time. Similarly, it will take different amount of energy, power consumption and all. So by measuring that information such as power dissipation, can I predict the key? Can I break it? That’s the idea of the side channel attack. So the interest is, not to actually break into a system, but to see whether there are problems in a system so that you can come up with a better system. Again we have done about 810 projects up to now under this topic. Right now there is a final year project, and it is called “Power Analysis Attack.” to do a reasonable amount of work. So there you measure the power con- Also, for the High Performance controller is doing the encryption. Then Computing research, we got a high performance server. It has a 32 core processor and 256GB memory. So we can use it to do parallel computing on CPU, and there were some projects we did in par- sumption of a system when the microby just analyzing the power consumption, we can figure out which key is used in the encryption. The interesting thing is that by using power measurement and the high-performance computing tech- allel computing using that server. niques, we can break a single encryption Any other recent research? 10 minutes. We also have shown that we One other research is “Side Channel Attacks”. It will take a while to explain 14 Gauge Newsletter of 128bits key AES algorithm in about can break a new cipher that is popular among embedded devices known as Speck in a small amount of time. In addi- University of Peradeniya So that’s all regarding research. And I have three more questions. Sir, there is a trend towards Software Engineering. So what do you think why students are less interested towards Embedded systems? Because of the industry trend in Sri Lanka, right? For the last 5 to 6 years, we are noticing this. When the students are in their 3rd year, they are very much interested in hardware and embedded systems. They are interested in doing embedded projects and some electronic projects, digital design projects. But when they go for training, they realize that most of them have to do what is done in the industry. It is how you perceive. However, the industry trend is changing towards embedded systems and hardware, and such companies are coming to Sri Lanka. I see it as a good trend, particularly for our graduates. Any message that you want to give to the students? Have the right attitude. You can learn anything else, everything you can learn if you have the right attitude. Believe in you, work hard and you can achieve anything you dream of. So sir, the final question. After you had finished your Ph.D., you came back here to teach. So my question is why did you choose Peradeniya again? So why lecturing in Peradeniya? That’s really an interesting question. Even I ask