Real time analytics on data that spans 100s of GBs
Submitted by Kaushik Paranjape (@kaushik-paranjape) on Thursday, 23 May 2013
Section: Storage and Databases Technical level: Intermediate
Analytics of stats data is a frequently faced problem by organizations that have online businesses. Doing real time analytics on this data is very important to take right business decisions. Sokrati’s solution for this problem uses sharded columnar databases as data warehouse. Using this technology query results are returned in milli-seconds on data-sets that span 100s of GBs.
Data processing, storage and analysis at a high scale is a hard problem to solve. I will be discussing evolution of sokratis architecture in this talk. The talk will also touch upon various design alternatives that were discussed along with pros and cons of each. Will deep dive into the current architecture that uses columnar database. The talk would conclude with a live demo of the system.
Kaushik Paranjape is a Co-Founding Geek at Sokrati. Sokrati is the leading Search Marketing platform in India. He is the architect behind sokrati’s robust technology. Some of his interest areas include big data, NoSQL, distributed systems.