If you’re an Enterprise IT manager, chances are that you are thinking about using Hadoop. You’re probably wondering “which distribution is best for my organization?” That’s when you end up sinking time into researching, weighing pros and cons, and delaying adoption. How do you figure out which of the components and tools you need, from which vendors and which version? Wouldn’t it be great if it were less complicated?
Pivotal thought so. Which is why they’re taking a chapter from their Cloud Foundry playbook (an open source enterprise class platform-as-a-service community) and leveraging open source and open collaboration for the Hadoop ecosystem. 15 firms, including Pivotal, GE, Hortonworks, IBM, Infosys, SAS, Altiscale zData Inc., and Capgemini are teaming up to form the Open Data Platform Initiative (ODPi) an initiative to bring standards to this great cacophony of confusion and make open source big data technologies more usable and broadly adopted.
The thinking here is that for accelerating innovation and adoption of a platform technology like Hadoop, an open source model is best, so long as it comes with some minimal organization around testing, standards, and versioning to ensure enterprise quality and readiness.
Shaun Connolly, VP of corporate strategy of Hortonworks said: “The Open Data Platform initiative will rally both enterprise end users and vendors around a well-defined common core platform against which big data solutions can be qualified.”
Not everyone is excited about ODPi, namely Cloudera, who is dismissing the whole thing as a mere marketing ploy. They and the more rabid among the open source fans also can’t resist seizing upon an opportunity to decry perceived infringement on consumer choice, asking what the ODPi Crowd has against freedom.
Their arguments sound like this: Shouldn’t customers be free to choose what they want in big data? What if Hagan Daz and Breyers decided to simplify their market? The Open Ice-cream Platform they’d call it. Everyone gets vanilla, nothing else. Is the saved shelf space really worth the loss of choice?
T o the ODPi-ers this view is so far off it’s earned a feature on the Nova series. An example that comes closer to this planet would involve licensing, and/or certifications. Doctors need to be certified in order to practice. It’s how we avoid getting diagnosed by people whose only credentials come from WebMD. Hadoop is currently an ecosystem with no baseline qualifications. If you go to the Apache website and look at the current stable release of Hadoop, and related projects like Hive , H-base, Pig, Zookeeper etc…you’ll find you can’t tie that stable release to any particular version from any vendor. ODPi will give vendors a common quality standard that they can be held to. When people see that a vendor is ODPi compliant, they’ll know that this means the vendor has a common core, quality code contributions and consistent distribution.
buy cheap viagra
But won’t working with a common core eliminate differentiation?
Don’t worry, the market will be growing from increased adoption. Instead of killing the ability to differentiate, the fact that companies will be building out of the same release level will give a baseline to differentiate from. After all, the MCAT (Medical College Acceptance Testing) hasn’t killed differentiation among doctors.
If you’re still thinking in terms of a product, you’re thinking wrong; this is about an entire ecosystem of open source projects.
Anjul Bhambhri, VP of Big Data and Streams at IBM describes ODPi as a “rich technology palette” for customers to work with. You can get core Hadoop from Vendor 1, a data wrangling solution from Vender 2, and a SQL product from Vendor 3, and they would all work together.
buy sildenafil citrate no prescription
An impressive new technology isn’t going to get anywhere without an ecosystem to allow it to thrive.
The big data market needs a “standard, predicatable, and mature Hadoop-based core platform for modern management solutions” according to Leo Spiegel, senior VP of corporate development and strategy at Pivotal. A large ecosystem of users and developers measuring from the same starting point will make a technology more stable and add more capabilities. Getting the badge of “ODPi Compliant” will finally cut down the duplication and take the guesswork out of sifting through what works and what doesn’t. For Hadoop this also means a greater interoperability across technologies, making it easier to adopt and use, in a shorter amount of time. For the rest of us it means rapid broad adoption and more Big Data use cases, which will result in better, faster, cheaper ways to glean insights from information.
To learn more about the ODPi and how you can get involved visit – opendataplatform.org