Taming Big Data – A Recap of the O’Reilly Strata Conference

O’Reilly Strata, a conference dedicated to data science, held its second meeting of the year from September 22 – 23 in New York City.  The conference drew close to 500 attendees, including all the major technology providers in the space as well as user organizations from various industries that are dealing with Big Data problems. Platform Computing was one of the sponsors of the conference, and our introduction of the newly released Platform MapReduce 1.5 generated wide interest among conference participants.

Platform team greets booth visitors 

My takeaways from the conference are following:
1)   Data science is hot. The Strata conference lured participants from various industries. We met startups working on web analytics, energy companies trying to analyze their log files, banks of looking for use cases and solutions to help analyze their large data sets fast, and of course, Web 2.0 companies who are already at the forefront of tackling Big Data problems exploring  better solutions than what they are currently using. The enthusiasm around cracking Big Data   has never been higher. 
2)   Big Data market is still young. Many discussions we had at the show revealed that majority of the main stream organizations (excluding Web2.0 companies like Google, Yahoo, Facebook) are still at the early stages of the adoption where they are either exploring various technologies on the market, figuring out the proper applications built upon new technologies such as Hadoop and MapReduce, or in the midst of building small pilot projects. Comments such as “We are thinking about moving certain applications to Hadoop,” or “We only have one Hadoop project running at this time” were often heard at the conference.  A majority of the organizations that have large amounts of data are just beginning to tap into Big Data and looking into proper use cases.. The FAQ these days is “What to do with my data?” and there is no simple answer to that. 
3)   There’s a shortage of skilled resources. While Hadoop and MapReduce appear to be promising approaches to access and analyze Big Data, they are also new to developers, and the learning curve is rather steep. Adding to the length of the learning cycle is the fact that development tools in the ecosystem are still yet to mature. The reality is that there is lack of production quality applications for main stream user, most codes today are developed in-house and still being tested in R&D labs.  
4)   The market is fragmented. Various tools have been developed to fulfill the ecosystem while the mainstream market is still catching up on the basics.. We believe the gap will close and the market will eventually hit the tipping point as more applications become available. But for now, the path for Big Data remains wildly unpredictable. 
Big Data is here to stay. However, “Making Data Work”, the slogan at O’Reilley Strata conference, is no easy task. Companies dealing with  large amounts of data have a lot on their plates today --  disruptive technologies,  new application development,  understanding the meaning of the  new discoveries and their business impact, just to name a few. Needless to say, a well built-out ecosystem is critical to support all the efforts taking place in the market.    At Platform Computing, we are committed to not only providing the best solution in the ecosystem through leveraging our proven technology, but also working toward  bringing a viable,  end-to-end solution to the market.      


Post a Comment