Update: I just spoke with Mark Zuckerberg right after the event and asked him what the tech was involved in this new big search plan.  He told me that Facebook built their own system, called Unicorn, and it will have two strategies for this front end to handle the queries and natural language as well as a back-end, persistent storage system.

He would not mention the role Hadoop would play (or not); only that Facebook is building their own system.

Facebook today launched their search strategy called Graph Search currently in private beta according to Mark Zuckerberg. This is clearly a big direction for Facebook but one that is taking “baby steps”.

This will be free service for users and something that will augment their existing user experience.

Privacy is the big concern, and Facebook is walking a fine line there.  This, however, is a hard technical problem and Facebook is setting expectations very low calling this beta of the beta.

Translation:  We don’t want to awake Google about our land grab to drive billions of new revenue.

They then detailed the “Three Pillars to Facebook’s Product Platform,” as:
1. Finding topics what’s happening – News Feed.
2. Finding what happened – Timeline.
and 3. Finding people – Now Graph Search

Facebook CEO Mark Zuckerberg made sure to explain that this new product is privacy aware and that each object of content has own audience and that most content is not public.  Zuckerberg had trouble explaining it because this is a hard problem to solve and the press corps are not that savvy on Big Data.  Zuckerberg made sure to explain the difference between web search versus graph search.

It will work at the top of Facebook as a search box and is available for each users social graph – their connections.  Again it will be a beta version initially.

Facebook promoted four main use cases for this new graph search:people, photos, places, and interests. 

Tom Stocky and Lars Rasmussen gave a demo.  I am not super impressed with how they communicated this new product.  I do know that this is a hard problem to solve and most likely they are using Hadoop or equivalent NoSQL databases.

My Angle: Facebook is going to have a hard time to crack the code on “discovery” in social networks that will translate into a positive user experience “transaction”.  Social network “transactions” are so diverse that it’s extremely difficult to achieve.

I found the personalization option awkward and not elegant.  So this is search play not a push technology.

That being said traditional search is broken (I’ve been saying this for years) and the utility from users in social networks are transforming.  So I like the direction.  That being said Facebook seems to focused on locking users into their platform when the world doesn’t want one social network.  Facebook is like a “night club” it will eventually get boring and trying to keep people in one network is not what users want.

If Graph Search can create a “fresh” user experience then it might have a chance.  We certainly need more signal and less noise so Facebook’s mission and direction is noble and appreciated.  Time will see.

CNBC Coverage with Ben Parr

Prior Footage from Wikibon’s Jeff Kelly Before the Announcement

Here is Jeff Kelly Wikibon big data analyst commenting on what he was expecting from this announcements on this morning SiliconANGLE.tv Breaking Analysis program.   There is a big data angle on this direction for Facebook.  Go to Wikibon.org/bigdata for all the lastest technology research on the top trends in big data.