Wireless Ad-Hoc Networks

My PhD topic had been on inventing innovative link-state routing protocols for mobile ad-hoc networks. My significant contributions had been to design loop-free, efficient routing protocols using incomplete link-state information from neighbors.This is significantly different from well-known link-state protocols like OSPF, in which complete information is used for shortest path computation. I showed that the shortest tree computation using incomplete advertised information from neighbors is an NP-hard problem and designed a protocol that forms a forwarding graph instead of tree.Significant performance improvements were achieved. Also, my node-centric hybrid routing solution is one of the early solutions to address routing in mesh networks, in which there are gateways as well as mobile nodes.This showed how proactive and reactive mechanisms can be combined to address certain real-world networks.

Publications

J.J. Garcia-Luna-Aceves and S. Roy, "On-demand Loop-free Routing using Link Vectors", IEEE Journal on Selected Areas (JSAC) in Communications, special issue on Wireless Ad Hoc Networks, March 2005

J.J. Garcia-Luna-Aceves and S. Roy, "On-demand Loop-free Routing using Link Vectors", Proc. International Conference on Network Protocols (ICNP), Berlin, Germany, Oct 5-8, 2004

S. Roy On-Demand Link-State Routing in Ad-Hoc Networks, Thesis, Jun 2003

S. Roy and J.J. Garcia-Luna-Aceves, "Node-Centric Hybrid Routing for Ad Hoc Networks", Proc. 10th IEEE/ACM International Symposium on Modeling, Analysis and Simulation of Computer and Telecommunications Systems-Workshops (MASCOTS2002 Workshop), Fort Warth, Texas, Oct 12, 2002.

S. Roy and J.J. Garcia-Luna-Aceves, "An Efficient Path Selection Algorithm for On-Demand Link-State Hop-by-Hop Routing", Proc. IEEE International Conference on Computer Communications and Networks (ICCCN), Miami, Florida, October 14-16, 2002.

S. Roy and J.J. Garcia-Luna-Aceves, "Node-Centric Hybrid Routing for Ad Hoc Wireless Extensions of The Internet", Proc. IEEE Global Telecommunications Conference (GLOBECOM), Taipei, Taiwan, November 17-21, 2002.

J.J. Garcia-Luna-Aceves and S.Roy, "Node-Centric Hybrid Routing for Ad Hoc Wireless Extensions of The Internet", Book Chapter in Mobile and Wireless Internet: Protocols, Algorithms, and Systems, Kluwer Academic Publishers, June 2003.

S. Roy and JJ. Garcia-Luna-Aceves, " Using Minimal Source Trees for On-Demand Routing in Ad Hoc Networks ", Proc. IEEE INFOCOM 2001, Anchorage, Alaska, April 22-26, 2001.

S. Roy and J.J. Garcia-Luna-Aceves, "Node-Centric Hybrid Routing for Ad Hoc Networks", Poster Presentation at MURI Workshop, UC Irvine, 2002

Shared Infrastructure For Emerging Markets

SMEs are the second highest employers of India after agriculture. SMEs flourish the best in clusters. Cluster-based approach enables the small companies to share infrastructure, cost and build a niche identity.The success stories have started from Plato, Italy and Tiruppur is our own example of a successful cluster. In recent times, govts are implementing measures to bring the SMEs together in a cluster.There is one definite advantage of being geographically close; the industry can get the facilities of shared power, road, market and other necessary parts of the infrastructure. Additionally, based on the specialized know-how of each SME, an entire eco-system can flourish in a cluster. However, all clusters are not successful and in current times, market situations are slightly different. The SMEs have better opportunities; they have access to a bigger market, but at the same time the local markets have opened and there is more competition. The orders might come in big numbers and in bigger sizes and SMEs need to deliver in time with quality products. Therefore, the current challenge for the SMEs would be to scale, maintain high quality and highly reactive.
In today's increasingly connected India with IT capital, the question arises whether IT can make a difference and help the SMEs improve productivity. There exists some IIM studies, which show how individual companies benefit by adopting the latest technology and IT. Using modernized  equipments have the direct impact on productivity. However, the picture that an SME cluster, rather than an individual company will benefit from IT solutions is still not clear. IT gives a global presence to an SME through web pages, enables remote transactions and helps formalize many stereo-typed operations. IT can make a difference, specially in the initial times to the early adopters. However, when IT becomes more pervasive, it might not be the deciding factor. In this competitive world, the local SMEs can loose out without IT, though in the long run, it would be the other factors of the ecosystem that will decide the success. The SMEs of India are extremely small compared to size of corresponding SMEs of the western world and therefore, their buying power is extremely less. In such situations, the challenge is to make IT solutions affordable for the small enterprises. One research question, which we sought to answer is whether like other infrastructure, can we develop a shared IT infrastructure. That can involve the wireless routers, the gateways, storage or software.


Publications
S. Biswas, and S. Roy, “A Shared ICT Infrastructure for Indian SME Clusters”, in IEEE/ACM International Conference on Information and Communication Technologies and Development (ICTD2007), Dec, 2007

S. Biswas, S. Roy, and S. Seshagiri, "Collaboration in Indian SME Clusters: A Case Study", Workshop on "ICT for Business Clusters in Emerging Markets", in Communities and Technologies (C&T), June, 2007

S. Roy and S. Biswas, “Collaborative ICT for Indian Business Clusters”, in Proceedings of WWW, Banff, May, 2007 (Poster)




Optical Networks

In recent years, Internet has seen phenomenal growth of 70% to 120% in traffic, mainly fueled by online file-sharing (P2P) tools (Bittorrent), online gaming, video hosting websites (YouTube), and social networking sites (MySpace, Facebook) . In future, it is expected that web traffic would further increase. According to a recent report published by Cisco, IP traffic will increase at a combined annual growth rate (CAGR) of 46 percent from 2007 to 2012, nearly doubling every two years. Main contributors to that kind of traffic would be video (IPTV, Internet TV, video sharing). This had led to a renewed spurt in building higher bandwidth network infrastructure among the telcos and hence, an increase of interest in the optical domain. There are some exciting areas to work upon currently:
1. 40G/100G MAC and PHY
2. Carrier Ethernet
3. Converged IP and optical layer for high optimization
4. Grid computing
5. Network planning for high cost optimization and optimized deployment
6. QoS to support IPTV

Publications
S. Roy, “Recent Growth in Broadband Demands and Advances in Optical Networks”, in ICCBN Bangalore, July 2008

Network Management and Planning
Routing Analytics is an important tool to envisage the actual routing paths used for data transfer. It helps the enterprises, service providers and telecom operators to monitor routes, detect faults and plan expansion. Combined with a traffic monitoring tool, Route Analytics can provide a complete view of the network traffic and the paths followed. Route analytics products passively communicate with routers in the network, and show routes computed through different routing protocols like BGP, OSPF, IS-IS or EIGRP etc. Root cause analysis of a network failure at any part of the network due to remote events can be detected extremely fast and reduces considerably the network downtime.

Network Planning is an extremely essential tool for optimized, efficient and cost-effective deployment of networks - wireless or optical. However, the challenges in each are different. As the wireless equipments are getting commoditized, and shared spectrum being used by wide number of users, the challenge in planning a wireless networks are in terms spectrum management, load balancing, coverage and interference mitigation. In contrast to that, the optical networking equipments are expensive, and hence here, the challenge is to deploy an cost-effective network that can satisfy the current bandwidth requirement as also the future projected requirements. They both pose very interesting NP-complete problems with innumerable constraints and solutions lie in mitigating them using some real-world assumptions. What-if analysis is extremely important to understand how the current network can be incrementally modified or expanded.

Misc. Topics

Trust is an essential item in real world transactions. The question arises, can an online or electronic trust be defined similarly to the real-world concept such that transactions between entities can happen uninterrupted and faster without human intervention. There are several theories on how trust can be computed, and computed trust can be used to decide future transactions and how sufficient safeguards can be incorporated in case there is a breach in trust. Trust associated with security can  be a strong tool in creating P2P systems or online procurement exchanges.

Publications
S. Biswas and S. Roy, “A Trust Based Procurement Exchange”, International Conference on Information Systems, Logistics and Supply Chain, Lyon, France, May 14-17, 2006


Professional Activities

Reviewers: Infocom 2001, 2008
                Globecom 2008
                ICCBN 2008
                WPMC 2007
                JSAC 2007
TPC: ICCBN 2008
Organizer: Co-Chair of workshop on “ICT for Business Clusters in Emerging Markets” at the conference Communities & Technologies, June, 2007 at Michigan State University
Recognition: Biographee in Marquis Whos's Who In Americal 2009              
Talks:  CITRIS-Sensorwebs (UC Berkeley) - 2003
          UCSC - CMPE150 Spring 2003
          Infocom 2001
          ICNP 2004
          C&T 2007
          ICCBN 2008

         







Hosted by www.Geocities.ws

1