Research
I broadly work in the area of wireless networks, applying techniques from stochastic processes, optimal control, optimization and game theory, for modeling and performance analysis. My Ph.D work addressed the problem of relay selection for packet forwarding in sleep-wake cycling wireless sensor networks (WSNs). During my post-doctoral studies I worked on the problem of designing adaptive caching and request assignment strategies for content distribution networks (CDNs). During this time I also investigated the problem of modeling mobile-data traffic, resulting in the development of a traffic simulator for generating synthetic mobile-data users. During my stay at IIT Madras, I worked on the topic of characterizing coverage and percolation properties of relay-assisted wireless networks. My current research focus is on the topic of network economics where I am interested in subscription and pricing problems. Scroll down for an illustrative description of some of my work.
Participation-Dependent Social-Learning
With the advent of online subscription platforms that encourage users to review the quality of the corresponding service availed by them, the subscription decisions of the prospective customers depend heavily on such reviews. While the reviews of the participated-users can help other users in making informed decisions, learning from such reviews can also assist the service providers in devising attractive subscription policies that can eventually maximize their overall revenue. Thus, the aspect of social learning that arise in these markets can yield an interesting economic scenario where the decisions of the users and the providers are coupled via the reviews given by the earlier subscribers. In this work we are interested in investigating the effect of such interactions on the subscription policies of the service provider.
Pricing in Heterogeneous Networks
We are in the midst of a very challenging communication era that is witnessing an ever-increasing demand for mobile data. One promising solution to meet this demand is the proposal of mobile data offloading where the mobile operators are allowed to offload some of their users onto small-cell service providers (e.g., public WiFi operators, femto-cell operators, road-side unit operators, etc.), thus balancing the mobile-traffic load among the heterogeneous networks. Enabling mobile data offloading leads to some important economic implications that need to be addressed in order to realize the full potential of the mobile data offloading technology. Such economic implications arise primarily because the heterogeneous networks, in general, are owned by competing private operators. As a result, the small-cell service providers (SSPs) will be reluctant to provide the offloading services unless the mobile network operators (MNOs) provide attractive monetary incentives to the SSPs, thus compensating the SSPs for the additional CAPEX and OPEX that they may incur. In this work we are interested in studying the aspects of these economic interactions using the theory of network pricing.
Mobile Data Offloading
Mobile Data Offloading is a promising solution to meet the future increase in demand for mobile data. The proposal involves offloading cellular operators' traffic onto small-cell network, e.g., WiFi. Such data-offloading scenarios yield interesting network utility maximization formulations. We seek to solve such formulations using techniques from optimization and game theory.
Next-Generation Wireless Networks
Femto-cellular Network
Wireless Sensor Networks
These futuristic networks comprise two types of nodes − sinks and relays. Sink nodes are connected to an infrastructure backhaul, while relay nodes are used to extend the coverage region by providing multi-hop connectivity to the sink nodes. Typical examples include femto-cellular networks and wireless sensor network. In such networks, our objective is to characterize coverage and percolation properties. Specifically,
Coverage is a microscopic property that determines whether communication can be established with a sink node from a given location. We use techniques from queueing theory and renewal theory for this purpose.
Percolation is a macroscopic property that determines whether communication can be established between distant sink nodes in the network. Techniques from continuum percolation are used for evaluating percolation probability.
Caching in Content Distribution Networks
The potential availability of storage space at cellular and femtocell base-stations (BSs) raises the following question: How should one optimize performance through both load balancing and content replication when requests can be sent to several such BSs? We address this question by introducing novel optimization models where bandwidth and content availability are represented via cost functions. By scaling the arrival rates and content chunking we obtain a limiting regime that can be investigated using techniques from mean-field analysis. We develop algorithms that are optimal and stable in the above limiting regime.
Competitive Relay Selection
The problem of relay selection arises while forwarding an alarm packet in a sleep-wake cycling wireless sensor networks. The relay selection problem comprises one or more forwarding nodes and a collection of relay nodes that are switching ON (waking-up) sequentially in time; the forwarding-nodes' objectives are to choose a relay node so as to optimize a linear combination of their respective delay and "reward" offered by the chosen relay; the reward, for instance, could be a function of the progress made by the packet and the power required to get the packet across. We model such scenarios using techniques from stochastic games. The solution is characterized in terms of Nash equilibrium policies.
