Welcome to the Distributed System and Services (DSS) laboratory, where a dedicated team of researchers and experts are actively engaged in addressing the theoretical and implementational challenges of integrating autonomous distributed systems, such as IoT, edge, and cloud computing, and blockchain technology, to provide transparent, reliable and customized services to the customers.
Through our cutting-edge research, we aim to contribute to the academic community, industry, and society at large, fostering innovation and shaping the future of distributed systems and services. Our lab is currently focusing on the key challenges and pushing the boundaries of these domains to drive innovation and create real-world impact. Towards this end, the DSS lab focuses on various research problems, as discussed below.
Modern user centric applications rely on the data generated by the sensors, Internet-of-Things (IoT) or smart devices to enhance its quality-of-services. However, for the effective utilization of the data, these applications often need to transmit it to remote servers or the cloud for processing and perform analytics on the same. However, the transmission of large volumes of data necessitates substantial network bandwidth. To tackle the issue, we explore non-uniform sampling aimed at generating a concise representations of the data over time while ensuring maximal reconstruction accuracy of the signal. This approach is tailored to meet the specific requirements of the diverse applications, facilitating meaningful interpretation of the data. The applications of sampling can be extended to diverse fields such as medical condition detection, climate analysis, speech and image processing, and human activity monitoring. By strategically sampling the data, we aim to strike a balance between data size reduction and preserving crucial information for analysis and decision-making processes.
The DSS lab also focuses on the issues related to offloading of data generated by the IoT sensors to the cloud or edge computing environments. Its primary objective is to enhance the performance of smart devices, in terms of latency and/or energy by executing the computation intensive tasks on the edge and cloud computing platforms. This necessitates the understanding of the unique capabilities and limitations of both cloud and edge computing paradigms to address the research challenges like discerning the optimal destination for offloading the data, i.e., whether to offload it to the cloud or the edge network, and how to execute the data in the edge environment to minimize the execution delays. Ultimately, the goal is to enable the smart devices with the ability to make informed offloading decisions autonomously, adapting to changing network conditions and application requirements on the fly.
In recent years, decentralized blockchain technology has emerged as a platform to eliminate centralized control in the decision-making process to enhance the trust, transparency, security, and privacy of the application. However, this technology is at its inception and has various limitations that are generally referred to as the blockchain trilemma of balancing the decentralization, scalability, and security of the applications. Towards this end, our lab focuses on improving various architectural challenges of the platform, such as on-chain data search via smart contracts. In addition, our lab also focuses on the challenges of integrating layer-2 scalability solutions and studying their effectiveness for large-scale applications.
Software-defined networking (SDN) revolutionizes network management by decoupling the control plane from the data plane, providing a centralized and programmable approach to network administration. In comparison the traditional networks, it increases the flexibility and customizability of the underlay infrastructure. Nonetheless, researchers are working on several challenges for realizing the potentials of the platform. Currently, in the DSS lab, we consider the load balancing problem of the SDN controllers, using switch migrations and fractional switch migrations. The former involves reassigning a switch's connection from one controller to another, improving network efficiency. Whereas, the fractional switch migration allows the migration of a fraction of the switch's flow from one controller to multiple controllers. This adaptive load balancing in SDN can accommodate changing traffic patterns and diverse application requirements. Overall, SDN transforms network management and load balancing into more agile and automated processes, allowing organizations to adapt to the dynamic demands of modern network environments.
Distributed Systems and Services