Which Linux distribution that is more suitable for intel and why?

What does Intel's entry into Hadoop distribution mean for Cloudera and MapR?

• Intel Corporation announced the availability of Intel® Distribution for Apache Hadoop software (on 26 Feb 2013). Will this announcement affect startups such as MapR, Cloudera or HortonWorks in anyway? Below is a snippet from http://newsroom.intel.com/community/intel_newsroom/blog/2013/02/26/intel-aims-to-enrich-lives-by-unlocking-the-power-of-big NEWS HIGHLIGHTS The Intel® Distribution for Apache Hadoop software is built from the silicon up to deliver industry-leading performance and security to better access and use big data. New software offering now available with more than 20 partners supporting the launch. Intel is committed to strengthening the Apache Hadoop framework by contributing platform enhancements into open source in collaboration with the developer community. Intel is investing research and capital in disruptive technologies to spur innovation and enable the industry to build analytic solutions based on Apache Hadoop software. Partners supporting the launch include 1degreenorth, AMAX, Cisco, Colfax Corporation, Cray, Datameer, Dell, En Pointe, Flytxt, Hadapt, HStreaming, Infosys, LucidWorks, MarkLogic, NextBio, Pentaho, Persistent Systems, RainStor, Red Hat, Revolution Analytics, SAP, SAS, Savvis, a CenturyLink company, Silicon Mechanics, Simba Technologies, SoftNet Solutions, SuperMicro Computer, Inc., Tableau Software, Teradata, T-Systems, Wipro and Zettaset. Also see http://gigaom.com/2013/02/26/cloudera-who-intel-announces-its-own-hadoop-distribution/

• Answer:

  I was asked to answer.  Take this as mostly some random musings on the topic. From an outside perspective, Intel is targeting specific markets that the other folks are mostly ignoring. Intel appears to be trying to brand themselves as the security and embedded leaders.  This is a brilliant move.  Niche markets tend to be more profitable on a per unit basis and getting marked as focused in something like security is an easy money maker.  (Especially if you do it early enough!) This is compounded by the fact that most security consultants are completely confused by the Apache Hadoop community or the nuances involved.  That said, I'm concerned that Intel doesn't really understand that (data) labeling serves a different purpose than ACLs given some of the things I've read, but I'm hoping that's ultimately my own confusion on what they've said publicly. Meanwhile, MapR still has one of the weakest security models in the community.  We're several years into maprfs and it still relies upon an NFS implementation that lacks even rudimentary RPCSEC support.   That's a significant blocker for anyone serious about security.  (To be fair, almost all of the "We're replacing HDFS" plays lack a security model.) Given they are also a binary implementation, that mostly eliminates them from the embedded space as well without some serious work. Cloudera and Hortonworks have similar profiles and are busy trying to be everything to everyone.  They likely won't notice other than the usual in-fighting at the Apache level that is bound to happen.  It really depends upon how much pull Intel has with its customer base to get into the areas they aren't specifically making a play against.  Many C-level execs love "one throat to choke" solutions.  So that could work very well in their favor if they can either bully their way into the door or prove enough expertise that VARs and OEMs will feel comfortable making an Intel distribution choice #1 when they also buy CPUs, mobos, etc  from them.