dW Interviews: Crowdsource Big Data creativity at IOD Unconference
allenkane 270002Y2HJ Visits (2012)
Information on Demand 2012 Unconference (October 21-25, Las Vegas; Unconference on the 25th), part of the IOD conference that reflects the attendees' voices because the IT professionals at the conference choose the topics.
So what topics should we watch for?
Ubiquitous data and storage. Kobielus: Petabytes in the palm of your hand. Storage cost will be negligible, maybe less that US$1 per PB by the end of 2020. Data hoarding on your phone; security for the easily stealable devices.... O'Grady: Most people are already accessing petabytes of data ... for example, Google Maps.
Deployment models. Kobielus: Not one optimal deployment model for Big Data ... centralized, scalable, 3-tier architecture, PureData ... all components play together as if they are one big happy system.... O'Grady: Look at what people are trying to do with data ... what does it look like, where does it sit ... the challenges to correlating disparate data sets means using custom approaches depending on the data sets and outcome.
Relational vs. noSQL storage. O'Grady: Nothing is replacing the relational database ... procuring someone with skills in things like MapReduce often means not being able to use the superior solution for a problem.
Crowdsourcing Big Data. Kobielus: The notion of open source market for data science expertise is coming into its own.... O'Grady: [Bill] Joy's Law ... no matter who you are, the smartest people work for someone else. For example, the Netflix prize. The return for organizations that can leverage outside expertise is there ... the success depends on how and where you execute the outsourcing. [See also Kaggle: Making data science a sport.] 22:51
MP3 Download | Transcript