Networks Lab

Introduction

Banking and financial operations rely heavily on network services. A minor disturbance in network configurations may lead to huge financial losses. Management of a large-scale network for a financial institution is a challenging task. The problem is magnified when there are different types of emerging financial applications (e.g. Crypto-currency, Cloud banking, API-driven information exchange, AR/VR-based customer service, etc.) demanding different types of inherent network support.

In the Networks Lab, we aim to look into the design, development, and management of different types of network architectures to achieve an evolutionary network design and a prototype infrastructure to test the behavior of newly developed financial applications. The long-term research vision is to develop a management-free and future-proof network architecture with enhanced reliability. Enhanced reliability comes from the ease of debugging the network infrastructure, which is required for any secured financial institution. The goal is to design a network management system that is auto-scalable, self-adjusting, and at the same time provides ease of debugging network issues.

Objectives

The Network Lab will focus on the following short-term and long-term objectives

Short-term
  • Dynamic Network Telemetry: The availability of low-cost smart devices, has been one of the motivating factors behind the rise of IoT infrastructure. There are several use-cases of IoT to aid banks and financial institutions. For example, smart devices can be used for the assessment of mortgages/debts. On the other hand, consumers are often attracted to banks, which can provide smarter support. All of these use cases depend on the reliability of the underlying IoT infrastructure, which requires frequent adjustments and configurations.

    On one hand, the self-adjusting and self-healing network management system can increase network reliability; on the other hand, it can reduce the cost of management. However, to implement self-management, monitoring of network events and detection of dynamic changes in the network is key. Such “telemetric” application deployment and management can become a bottleneck with the scale of the network. Further, the optimum benefits from these applications require a thorough analysis of resource demand and resource consumption. Apart from that, dynamic deployment of telemetric applications can save resource consumption but may introduce monitoring lapses. In this context, the Network Lab is exploring the trade-offs between resource consumption and monitoring lapses. A objective of the lab is to design and develop a framework to ease the task of network administrators.

  • Vulnerability Analysis using Log Level Information: Security threats have always been a major concern for banks and financial institutions. Even though it is not always possible to anticipate and prevent all the attacks or malicious activities on organisations, identification of the root cause behind a set of events is absolutely necessary to prevent further exploitation. However, modern applications are often deployed in the system in an intermingled manner, which makes system administrators’ task of finding the root causes behind any event very difficult. The problem becomes even more difficult when the system is built in a distributed fashion. One of the ways to detect the execution sequences is by going through the logs generated by the applications and the systems. However, identification of the log sequences generated due to a particular event is non-trivial. Therefore, we intend to develop a framework that can provide a framework to the system administrators to identify the root causes more efficiently. Further, we want to extend the framework to assess the security vulnerabilities of the entire system.
  • Common Network Application/Protocol Development Framework: The Internet is inherently a collection of heterogeneous devices and technologies. Design, deployment, verification, and debugging of an end-to-end application have been a major challenge since the beginning. The innovative techniques of software engineering cannot be applied directly to the network application/protocol development due to this heterogeneity. Recent advancements in SDN and Network Functions Virtualisation (NFV) have tried to instil Dev-Ops-like software design and deployment techniques into network development. However, verification and debugging of programs remains difficult, as the segregation of test and production environments requires significant efforts. This difficulty can be overcome by designing a common integrated development framework that can ease the task of the programmers. In the Network Lab, we are currently looking into the programmable data-plane architectures and orchestration frameworks to develop a distributed development and debugging interface.
Long-term

Conduct applied research on the following topics:

  • Design and development of future proof Internet Architecture and evolutionary protocols to assist banks and financial institutions to develop new products.
  • Development of integrated frameworks for debugging distributed applications.

This lab also intends to provide a realistic test environment where banks and financial institutions can rapidly prototype their distributed product development ideas.

Coordinator

Dr. Subhrendu Chattopadhyay

Assistant Professor

Research