T. S. Mohan

Cloud Computing researchers and enthusiasts have been always found it challenging to understand the internals of typical commercial cloud offerings due to its abstractional opacity. Many times they tend to become speculative and conjectural about how Amazon or Google or Microsoft operate their proprietary cloud data centers. In this talk, we examine the open source options available for one to setup private clouds in enterprises – essentially Eucalyptus (and say Ubuntu EC) or Hadoop as well as their interoperability with their variants on Amazon Web Services. While configuring, deploying and managing cloud applications for performance on these private or hybrid cloud tend to be difficult, there are significant useful learnings that we can obtain from the world of Grid computing and the large scale scientific application enablement on them. In this talk, we examine the commonalities, issues and challenges in leveraging these grid learnings in the effort of augmenting private and hybrid clouds. We also examine where and how Google Apps or Azure fit in this landscape.