Sailesh Krishnamurthy







A brief introduction

I completed my doctoral studies in Computer Science in the University of California, Berkeley in 2006.  Subsequently, I co-founded Truviso (formerly  Amalgamated Insight) along with Mike Franklin and Mike Trigg.

The Berkeley days

At Berkeley, I was associated with the Database Group and I worked on the Telegraph project with Prof. Mike Franklin and Prof. Joe Hellerstein.  We built TelegraphCQ, a new database system for processing continuous queries over potentially unbounded data streams. This was implemented using (and disembowelling)  the PostgreSQL code base. PostgreSQL is an open source database system that is a descendant of the Postgres project at Berkeley.

The gory details 

 My Ph.D. focused on the problem of query scalability  in a stream query processor (e.g., TelegraphCQ).  In particular, I showed how to build a stream query processor which could share the processing of multiple concurrent queries of various kinds (e.g., joins, aggregates) that are simllar, but different.  The unifying theme of my work is that my techniques all support on-the-fly sharing, i.e, the property where queries can be shared even when they are added and removed at will. I argue that this model of shared query processing is vital in a practical query processor.

Ph.D. Dissertation

Title: Shared Query Processing in Data Streaming Systems