CSCITA 2017

2nd International Conference On
Communication Systems, Computing & IT Applications

Venue: Hotel Kohinoor Continental, Mumbai | 7 - 8 April, 2017

St. Francis Institute of Technology
(Approved by AICTE, Affiliated to University of Mumbai, ISO 9001:2008 Certified, NBA Accredited)

Keynote Addresses & Invited Talks:

Keynote Addresses
1: Mr. Rajkumar Kettimuthu
Research Opportunities in the Changing Landscape of Computer, Network and Data
Abstract: Growing volume and velocity of data generated by scientific experiments is creating unprecedented demand for computer and network capability and capacity. To meet these demands, government agencies are making investments to push the next frontier of high-performance computing—namely, exascale, which is expected to be attained early in the next decade. As many science experiments use geographically distributed computing resources and move large volumes of data among these resources, significant investments are being made to build high-speed wide-area networks. There is a growing need for on-demand computing. Commercial clouds leverage the law of large numbers and some degree of overprovisioning to enable immediate scheduling of at least modest-sized tasks. Supercomputing facilities, which still follow traditional model of batch scheduling, are looking at transformations to support on-demand jobs. Recent developments in Software Defined Networking, Network Function Virtualization, and related technologies hold an enormous promise in programming the network on-demand to meet the requirements of scientific experiments. In this talk, the challenges that the changing landscape of data, compute, and network present and the research opportunities they provide will be discussed.
Speaker: Rajkumar Kettimuthu
Rajkumar Kettimuthu received the B.E. degree from Anna University, Chennai, India, and an M.S. and Ph.D. from the Ohio State University, Columbus, OH, USA, all in Computer Science and Engineering. Since 2003 he has been working at Argonne National Laboratory, where he is currently a Computer Scientist in the Mathematics and Computer Science Division. He also holds joint appointments at The University of Chicago as a Senior Fellow in the Computation Institute and at Northern Illinois University as a Graduate Faculty Scholar. His research is focused on software-defined networking, intersection of HPC and cloud, data-intensive computing, and high-speed transfer of large-scale data. He has co-authored more than 90 articles in the above-mentioned areas. He is a recipient of R&D 100 award. He is a senior member of both IEEE and ACM.
2: Dr. Sunil Kumar Kopparapu
Recognizing Emotions in Spontaneous Speech
Abstract: Emotions when explicitly demonstrated by an actor are easy for a machine to recognize by analyzing their speech. However in case of day to day, naturally spoken spontaneous speech it is not easy for machines to identify the expressed emotion even though emotion of the speaker are embedded in their speech. While there are several reasons for this, one of the prominent reasons for this is that people, especially non-actors, do not explicitly demonstrate their emotion when they speak, thus making it difficult for a machine to recognize the true emotion implicitly expressed in their spoken speech. In this talk, based on some of our previous published work we identify the challenges in recognizing emotions in spontaneous speech and discuss the different approaches that we have explored in the lab which has helped in reliably determining the emotions expressed in spontaneous speech. The talk is aimed to motivate young researchers to get an insight into the challenges and hopefully getting them to work on this problem.
Speaker: Dr. Sunil Kumar Kopparapu
Dr. Sunil Kumar Kopparapu (Senior Member, IEEE; ACM Senior Member India) obtained his doctoral degree in Electrical Engineering from the Indian Institute of Technology, Bombay, India in 1997.

Between 1997 - 2000 he was with the Automation Group, Commonwealth Scientific and Industrial Research Organization (CSIRO), Brisbane, Australia working on practical image processing and 3D vision problems, mainly for the benefit of the Australian mining industry.

Prior to joining the Cognitive Systems Research Laboratory (CSRL), Tata Infotech Limited, as a Senior Research Member, in 2001, he was associated with the R&D Group at Aquila Technologies Private Limited, India, as an expert for developing virtual self line of e-commerce products.

In his current role as a Principal Scientist with the TCS Innovations Labs - Mumbai, he is actively working in the areas of speech, script, image and natural language processing with a focus on building usable systems for mass use in Indian conditions.

He has coauthored a book titled Bayesian Approach to Image Interpretation and more recently a Springer Brief on Non-linguistic Analysis of Call Center Conversation apart from several patents, journal and conference publications.


Invited Talks
1: Dr. Bernard Menezes
Recent Advances in Cache-based Side Channel Attacks on Cryptographic Algorithms
Abstract: Through much of the history of cryptography, attacks on cryptographic algorithms have focused on cracking hard mathematical problems such as the factorization of very large integers (which are the product of two very large primes) and the discrete logarithm problem. More recently, however, side channel attacks have gained prominence. These attacks leak sensitive information through physical channels such as power, timing, etc. and, typically, are specific to the actual implementation of the algorithm. In addition to the software libraries used, the processor and memory architecture (different Intel x86 and ARM releases, for example) and operating system versions also dictate the success of the side channel attack.

An important class of timing attacks are those based on obtaining measurements from cache memory systems. They exploit the fact that access times to different levels of the memory hierarchy vary by 1 – 2 orders of magnitude. Cache-based side channel attacks have targeted widely used cryptographic algorithms such as RSA, DSA (and its elliptic curve version) and AES (the Advanced Encryption Standard). The latter is a relatively new algorithm for secret key cryptography, now ubiquitously supported on servers, browsers, etc. Almost all software implementations of AES including the cryptographic library, OpenSSL, make extensive use of table lookups in lieu of time-consuming mathematical field operations. This talk focuses on a cache-based side channel attack which aims to retrieve the encryption key of a victim performing AES.

We consider two possible scenarios. The first involves a victim process running on behalf of a data storage service provider who securely stores documents from multiple clients and furnishes them on request after due authentication. The same key or set of keys is used to encrypt documents from different clients prior to storage. In the second scenario, two entities, A and B, exchange encrypted messages. The victim is on B’s machine and decrypts blocks of ciphertext received from A. In both scenarios, we assume that the attacker or spy is hosted on the same processor core as the victim and that their executions are interleaved. Moreover, both attacker and victim use the OpenSSL library. So only a single copy of OpenSSL is resident in main memory and is mapped to the virtual spaces of both, attacker and victim.

We design and implement a multi-threaded spy process - the spy threads execute in round robin fashion. During each run of a spy thread, the cache lines containing the AES tables are flushed out of cache. When CPU control returns to the victim, it brings in some of the evicted line(s). When control returns back to the spy, it determines which of the evicted lines were fetched by the victim by measuring the time to access them. To complete the cycle, it then flushes out the AES tables from cache before relinquishing control of the CPU. Our algorithms to deduce the AES key use two crucial inputs. The first of these is either the plaintext or the ciphertext. In Scenario I, the attacker could pose as a customer to the data storage service provider and request that his documents (plaintext) be securely stored. The attack in Scenario II makes the reasonable assumption that the ciphertext to B can be eavesdropped upon. The second crucial piece of information is the line (or block) numbers of AES table entries accessed by the victim. Several entries in the AES table are placed on a single cache line and the spy provides a set (not list) of lines accessed. The absence of temporal information makes it especially challenging to deduce the key especially for sets with larger cardinalities.

Given the above two inputs, our main contribution is the design and implementation of a suite of algorithms to deduce the AES key. Our algorithms are simple, are elegantly expressed using relational algebraic operations and run in a few minutes. Above all, our attack is highly efficient – we demonstrate recovery of the full AES key given only about 6 – 7 blocks of plaintext or ciphertext (theoretically even a single block would suffice). This is a substantial improvement over previous cache-based side channel attacks that require between 100 and a million encryptions.
2: Dr. Krishna Mohan Buddhiraju
Recent Advances in Satellite Imaging
Abstract: Image processing requires detailed knowledge of the process of the image acquisition, and properties of the sensors involved in generating the image so that proper pre-processing, and analysis algorithms can be developed. It is equally valid for images acquired from spaceborne platforms on Earth orbiting satellites. Some of the defining properties of images acquired based on the principle of remote sensing from space of the Earth include their spatial resolution, spectral resolution, radiometric resolution and temporal resolution. Another important aspect of space based imaging is passive – natural solar radiation reflected by objects on Earth surface is sensed by the imaging system versus active – objects on the Earth surface illuminated by energy sources onboard the satellite, particularly in the microwave region of the electromagnetic spectrum. The received signal in the optical/infrared regions of the EM spectrum is handled differently compared to the signal in the longer wavelengths of microwaves where both amplitude and phase of the received signal are recorded and used. The advances in imaging due to improvements in sensor technology make it possible to undertake exciting new analysis paradigms and applications. In the optical/infrared portions of the EM spectrum, image acquisition has developed to the extent that the sensor system can focus on as small areas as 1 ft x 1 ft on the ground, enabling extraction of information at a very fine detail. On the other hand, spectroscopy which used to be the sole purview of chemistry laboratories, is made possible in imaging from space with hyperspectral sensing technology that allows studies on vegetation health, water quality, monitoring harmful algal blooms in oceans, search for traces of precious minerals in the top soil for mineral prospecting, etc. The latest development in imaging, though not from spaceborne platforms, is due to Unmanned Aerial Vehicles (UAV) wherein very low altitude imaging at extremely high spatial resolution helps developers of townships, airports and airfields make the calculations with high precision. An exciting aspect of UAV based imaging is that both hyperspectral sensing with high spatial resolution can be achieved that enables study of features based on their spectral response as well as physical geometry, topological relations between neighboring objects, and so on. Some of the interesting problems that are being pursued with remote sensing based imaging include content based image retrieval, watermarking, classification using deep learning methods, three-dimensional modelling of an area using multi-view imaging, radar interferometry, and change detection.
3: Dr. Satyanarayana Bheesette
Leveraging advances in ICT for experimental science research
Abstract: Advances in information, computer and communication technologies - together, revolutionised the design, deployment, functional efficiency, cost and throughput of rest of the technologies. Just on similar lines to the way they catalysed and influenced a phenomenal growth of diverse areas such as industry, education, agriculture, business, infrastructure etc., ICT started playing decisive roles in the conduct and quality of modern experimental science research. And in fact, there is more than this unidirectional cause and effect relation in case of science and ICT. Basic science gives birth to technologies and engineering. The technologies we are used to and getting benefited from as well as the engineering which helped us build useful gadgets and products were all result of the seeds of basic science which discovered laws of nature and explained the basic phenomena of the universe in which we live. But interestingly the technology and engineering also helped science with many major modern discoveries. Besides, need to build highly complicated modern science experiments also motivated development of state-of- the-art technologies. World Wide Web, GRID computing, fibre optics, super conductivity, silicon and VLSI technologies, novel computer architectures etc. are only a few of the examples in this respect. This talk will introduce the interdependence of science, technology and engineering. It will dwell - using some case studies, on the crucial roles of information computer and computer technologies for cutting-edge basic science research.
4: Dr. Supratim Biswas
Automatic Parallelization of Programs for Contemporary Architectures
Abstract: The computational capabilities of the multi-core and GPGPU architectures of today are enormous. Applications from diverse domains such as computational fluid dynamics, graphics and image processing, genetics and cell biology, exist that admit large scale parallelism. Harnessing the computing power of these immensely powerful contemporary architectures has remained as difficult as ever. Various lines of research are under investigation in this area to address the gap, a) using advancements in design of parallel algorithms, b) invention of parallel programming languages, c) design of system software tools for efficient execution of the parallel programs on a given architecture, and d) automatic generation of efficient code for contemporary architectures for existing sequential programs.

This talk focusses on the last approach. The two key research objectives are, (a) automatic detection and extraction of parallelism from sequential programs, and (b) exploitation of the extracted parallelism on the multi-core architecture. The basic issues are illustrated with simple examples. The state of art of research in both theory and practice is explained and the challenges of the immediate present and the near future are stated.
5: Mr. Nimish Somaiyya
TBA
Abstract: TBA