The Fifth Elephant 2013

Let’s talk about a multi-faceted system with unique data storage and processing challenges.
One of the core challenges in SEM is to setup and maintain a marketing campaign. A good way to do so is to know the inventory in and out and be smart about it.
With this system we aim at addressing such challenges in a scalable manner.

Outline

Yes, we ended up using three different data storage paradigms...certainly a maintenance hassle. What did we gain out of it?

This talk would present the evolution of the system and how it leverages the USPs of such variegated storage technologies to become more elegant than a trivial implementation.

What aspects of our data qualify the contemporary big data paradigm.
Where does all that data come from? Certainly not from an inventory.

• Redis: For simplifying the programming constructs and enhance the abstractness factors for processing-intense
  algorithms.
• Mongodb: Significance of dynamic collection-naming in multithreaded processing, achieving inherent lock-free-ness vital for performance in processing big data.
• Mysql: For the ACIDic wisdom of relational aspects of the data.

Speaker bio

Have been working on designing, implementing and testing scalable data-heavy systems in Sokrati since one and half years. Have been working on FOSS for about four years. Personal interests align with big data analytics in terms of machine learning. Core expertise is in JAVA. Completed my graduation from PICT, Pune.