---

I'm a first year Computer Science graduate student at University of Massachusetts, and my research interests are primarily in networks, distributed systems and security. My advisor is Arun Venkataramani.

I graduated with B.S. in Computer Science with Departmental Honors from the University of Washington in June 2010. My undergraduate research at UW focused on network manageability, trusted computing and network load balancing using OpenFlow switches. This work was advised by Colin Dixon, Tom Anderson and Arvind Krishnamurthy.

Research

• ETTM: A Scalable Fault Tolerant Network Manager (Paper)
  Colin Dixon, Hardeep Uppal, Vjekoslav Brajkovic, Dane Brandon, Thomas Anderson, Arvind Krishnamurthy
  USENIX Symposium on Networked Systems Design & Implementation (NSDI), 2011.
  
  UW CSE Industrial Affiliates - Madrona Prize runner-up, 2010 (Poster)
  
  
• Enabling Trusted Distributed Control with Remote Attestation (Undergraduate Thesis)
  Hardeep Uppal
  Undergraduate Thesis, University of Washington, Department of Computer Science and Engineering, June 2010.
  
  
• OpenFlow Based Load Balancing (Paper) (Slides)
  Hardeep Uppal and Dane Brandon
  Submitted as a project requirement for grad networking course at UW CSE.
  
  

Other Projects

• Text Based Driving Direction in Real Time (Slides) (Report)
  Hardeep Uppal, Dane Brandon and Tatiana Gershanovich
  Google Mentor: Barry Brumitt
  Professor: Ed Lazowska
  

  Abstract:

  In this project we have implemented a modified version of Dijkstra's algorithm in MapReduce to precompute the shortest paths to highways from all reachable points (road intersection) in the map. By parallelizing search at runtime into independent segments on subsections of the map, which are then combined into a final path, we can find a near distance-optimal path between any two points on a map in less than one second.

  
  
• Rendering Map of U.S. using MapReduce (screenshot1) (screenshot2) (screenshot3)
  Hardeep Uppal and Dane Brandon
  Professor: Ed Lazowska
  

  Abstract:

  This project uses Hadoop's MapReduce framework on a larger scale cluster to render map of transportation/geographic features of U.S. as a set of tiles using the TIGER dataser. We also generate geocode index from TIGER dataset for address lookups on a map.