Why hardware giants are Hadoop distro happy

Why hardware giants are Hadoop distro happy

Summary: EMC and Intel launch Hadoop distributions as they realize that big data is going to create a lot of hardware opportunities.

SHARE:

EMC has its own Hadoop distribution and now Intel has joined the cause. What's notable is that both EMC and Intel are primarily hardware vendors increasingly playing a software game. The only way to ride the big data wave is to play Hadoop.

Let's recap recent events:

  • EMC launched its Pivotal HD Hadoop distribution. The move melds EMC's Greenplum and storage intellectual property with Apache Hadoop.
  • Intel launched the Intel Distribution for Apache Hadoop. The name isn't exactly trendy, but the cause is notable. Intel is looking to build Hadoop from a silicon level to bolster computing speeds as well as security. With Xeon processors now supporting the Hadoop Distributed File System, Intel is claiming that processing time can be cut from four hours to 7 minutes.
  • Aside from contributing to what is becoming a Hadoop distribution glut, EMC and Intel are making moves that will matter to their futures. Here are the moving parts:
  • Hardware vendors see Hadoop as the next big thing that will drive compute.
  • The computing needs of big data will result in more hardware---servers, storage and networking.
  • Optimized Hadoop for EMC storage and Intel processors makes a lot of sense.
  • Hardware is ultimately a bad business as software begins to dominate the data center.

With that backdrop, the EMC and Intel Hadoop happy moves make a ton of sense.

pivotalHD
intchadoop022613

 

What's the fallout for the big data landscape?

  1. Oracle may need its own Hadoop distribution to bridge big data and its databases. Today, Oracle is a Cloudera partner. Perhaps Oracle buys Cloudera.
  2. Teradata and other data warehousing players have connectors to Hadoop. These vendors may need their own distributions, but would have trouble rising above the din.
  3. Informatica makes a living on data integrated. If big data is integrated at the hardware level there's less work to do.

The problem for IT buyers is obvious: What Hadoop distribution do you pick? Intel has a bevy of partners. EMC is going after Cloudera. IBM has its own distribution. Toss in Cloudera and Hortonworks and the selection borders on too much. For now, it's probably best to sit back and evaluate all of the players. Your infrastructure---as well as your intent to stick with it---may be a gating factor.

Topics: Big Data, Data Centers, EMC, Enterprise Software, Hardware, Processors, Storage

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

7 comments
Log in or register to join the discussion
  • Hadoop is legacy technology under new marketing

    It will be dead within two to three years and replaced by the next hype.
    jorwell
    • Hadoop is legacy technology under new marketing?

      I agree but like many of the "new" technologies that are old or legacy ideas, architecture, software... ("Cloud", iPhone, iPad, tablets ... etc comes to mind) when the "big boys" in the industry start implementing solutions or products using such legacy technologies, it's a sign that the technologies time has arrived.
      toomuchtime
      • Sorry I used the wrong word - I meant obsolete.

        Key value pairs, as last used on 1970s vintage mini-computers that didn't have enough memory to run ISAM.

        The thing is a bad joke really and a huge step backwards for data management.
        jorwell
  • Does the name "HADOOP" actually mean something?

    Just curious. I've not done much in the way of big data yet, but know that it is coming. The only thing that I can think of is that it means "High Availability Data ..." and then it gets lost for me.

    Thanks.
    MikeBlane
    • Never mind...

      I found the answer...
      MikeBlane
  • Hadoop and hardware

    Is this a surprise?

    Hadoop is great at pulling together the resources of large pools of processor, RAM, and disk deployed on smaller commodity servers.

    This is a great way to sell hardware.

    It's also a huge threat to companies like EMC, who sell storage area networks, which are not necessary with Hadoop deployments. It's either co-opt or die.

    Especially since they can add small pieces of software or integrate existing piece (ie: Greenplumb) and have an instantly sale-able branded product. Minimal investment, maximum exposure.

    Is this shocking to anyone?
    Marc Jellinek
  • Hadoop on Windows...

    Hortonworks has a version that will run on Windows currently in beta. We are only a few months away from this taking off as something to use the extra VM host resources. The numerous word count examples make this seem like a useless tool. Once you figure out how it can find missing relationships in your ERP and pull trend data out of all the documents in your enterprise portal, then it becomes something that will increase profits. Mention to a CIO that you can increase profits simply by using the extra cycles on the VM hosts and you have a winner.
    josh@...