Cloud development and internship

Cloud computing, though a common term now, was still a convoluted term in the late 2000s, as it itself found it’s first official parent in AWS in 2006. Out of the zeal to understand the working and offerings of this new paradigm, the Linux Users’ Group at my college started working on the creation of a cloud.

Using Eucalyptus and Euca2ools, we raised a 20 node cluster and later expanded it to 40 nodes. Using HPL Linpack Benchmark, we ran a matrix multiplication program which measured our cluster’s processing power to be 74 GFlops, when all the nodes were running.

During my internship under Dr. Rajiv Mishra, I created yet another Eucalyptus cloud where I replicated almost the entire work done by our team at my college. The project required MANET based data processing and the data itself was crowd-sourced which generally came from automobiles. As the incoming-data could be huge, we needed a Big Data framework for data processing. I followed up the cloud deployment setup with Hadoop installation on top of it and ran some Hadoop codes to test the end-to-end functionality.

In college:

  • People – VIT-LUG members in around 2009-2011
  • A small report – 1, 2, 3

At Internship: