A key challenge in building next-generation data networks is to satisfy the increasing demand for data traffic and extended coverage for people and places without a corresponding rise in energy and carbon emissions. There will be an explosion in wireless connectivity due to Internet-of-Things (IoT) devices, and these devices consume more energy for communication than for local processing.
One approach for prolonging the life of these devices is to increase the amount of computation at each node, such as by adding signal processing at the nodes, in order to reduce the number of bits that need to be communicated. Similarly, in distributed computing frameworks such as MapReduce, one can increase the computation load of the Map functions in order to reduce the communication overhead.
Multihop wireless transmissions allow for nodes to transmit at low power levels and reuse the same frequency in different physical locations. In situations where battery life is important and communication to distant nodes should be avoided, distributed solutions are feasible. Distributed algorithms perform suboptimally when compared to a global centralized algorithm but can be implemented efficiently.
The Distributed Computing and Networking R&D Group at INSOFE, led by Prof. Ashwin Ganesan, conducts research in all areas of networking, including interconnection networks, sensor networks, wireless networks, mobile computing, and IoT. The group also investigates research questions in data science. The group’s research focus is on the theoretical foundations of distributed systems and networks, and in particular on the communication-computation tradeoff mentioned above.
Consider the following fundamental problem in networking: what are fundamental limits to the performance that is achievable with some given amount of resources? More specifically, if each node in the network has information about only its local neighborhood, then what are the limits to performance? Given that each link in a wireless network has a certain minimum-bandwidth quality-of-service requirement, is it possible to determine in a distributed fashion whether the given set of flow rates is feasible?
In recent research by the group, new distributed algorithms have been proposed for these problems and the tradeoff between the level of decentralization and the performance of these distributed algorithms has been quantified. Despite the large volume of prior work on this topic, this work has succeeded in cutting new ground – in analyzing models that have not been considered in the previous literature and in obtaining interesting and nontrivial performance bounds. These research results were published in the journal IEEE/ACM Transactions on Networking in February 2020 and will be presented at the International Conference on Distributed Computing and Networking in January 2021.
If you would like to collaborate with us or join our doctoral program, feel free to contact us on LinkedIn.
- Conduct research in all areas of networking, including wireless networks, sensor networks, mobile computing, interconnection networks, and IoT
- Investigate the design and performance of distributed algorithms and resource allocation and scheduling algorithms
- Conduct research in data science, with a focus on graph algorithms in data mining and networked structures in economic and social networks.
- A. Ganesan, “Performance guarantees of distributed algorithms for QoS in wireless ad hoc networks,” IEEE/ACM Transactions on Networking, vol. 28, pp. 182-195, February 2020.
- A. Ganesan, “On some distributed scheduling algorithms for wireless networks with hypergraph interference models,” submitted for publication in October 2019, revised in July 2020.
- A. Ganesan, “Distributed algorithms for QoS in wireless ad hoc networks under the primary interference model,” Proceedings of the International Conference on Wireless Communications, Signal Processing, and Networking (WiSPNET), Chennai, India, August 2020.
- A. Ganesan, “Performance analysis of a distributed algorithm for admission control in wireless networks under the 2-hop interference model,” Proceedings of the 22nd International Conference on Distributed Computing and Networking (ICDCN), accepted/to appear, Nara, Japan, January 2021, ACM.