Looking for Bachelor or Master projects, or a Master thesis? Below are some project ideas.
Looking for a Ph.D. semester project? Email George.

Our projects are about investigating new ways to build systems and analyze them.
Don’t worry if you don’t know some of the concepts we mention, as long as you’re willing to learn.

The following list is for Spring 2021.
If you’re interested, e-mail us at projects@dslab.org

  • one paragraph of motivation explaining which project(s) interest(s) you and why
  • your résumé (curriculum vitae)
  • grade transcript

Implement a fast driver for a 100 Gb/s network card

DSLAB has designed a new network card driver model, which is faster than existing drivers because it is less flexible. We have an implementation for a 10 Gb/s server network card, the Intel 82599. How fast could a driver in this model be using a 100 Gb/s card, such as the Mellanox ConnectX-5 or Intel E810?

In this project, you will:

  • Implement a driver for a 100Gb/s network card
  • Evaluate its performance compared an existing driver

You will learn:

  • Low-level network card driver details
  • New ideas in driver architecture

Quantify the overheads of modern Linux networking

Traditional network APIs in operating systems, such as “sockets”, were invented where the network was slow. To handle modern network speeds, such as 100 Gb/s Ethernet, some networking frameworks bypass the operating system, which is fast but loses all isolation. However, modern operating systems have added faster APIs, such as Windows’s “RIO” and Linux’s “AF_XDP”, which are designed to increase performance by avoiding most of the current overheads. In a previous project, we found that Linux’s “AF_XDP” networking API is faster than the older Linux APIs but still not as fast as kernel-bypassing framework such as DPDK. Is this a fundamental limit given the isolation guarantees provided by Linux? Or are there still unnecessary overheads?

In this project, you will:

  • Find the overheads in modern Linux networking
  • Determine which overheads could be eliminated at what cost

You will learn:

  • Advanced Linux networking

A 100% secure and reliable smart home

Homes are becoming increasingly “smarter”, where various aspects of home operations are automated and controlled intelligently: heating/cooling, lights, blinds, alarm system, video surveillance, interphone, door access, etc. The controls usually reside within a “hub”, which is accessible over the Internet. This poses serious security and reliability challenges, and there are plenty of hackers happy to take advantage of the opportunity. The goal of this project is to build a smart-home hub and then formally verify its reliability and security; see open-source examples here.

In this project, you will:

  • Take an existing smart-home hub and modify it (or build a verified one from scratch)
  • Formally verify the result hub

You will learn:

  • Lots of details about smart homes and related protocols and equipment
  • Formal verification techniques

Your own idea

Do you have a project idea related to DSLAB’s interests, such as dependability or practical formal verification?
Feel free to contact us!

In this project, you will:

  • Work on your own idea with our guidance
  • Discover something new, small or large

You will learn:

  • Something you don’t know already