Last week we were lucky enough to have some time to sit down with ODPi members Steve Jones, global vice president of Big Data at Capgemini, Mike Maciag, COO at Altiscale, Sean Keenan, co-founder and vice president of products at UNIFi Software, and Todd Moore, vice president Open Technology at IBM, to discuss the future of big data, cognitive computing and how ODPi can help drive innovation. Todd Moore also expanded on his thoughts from his recent blog post on ODPi.
Here is what they had to say:
1. What are the challenges customers face with Hadoop and how will ODPi solve them?
Maciag, Altiscale: ODPi makes it much easier to select a Hadoop platform, select applications that will run successfully on that platform, and switch platform providers in the future, if necessary. Hadoop is actually a large ecosystem of different software projects and each component gets updated and released on a different schedule. As a result, customers today have a tough time making a decision about which platform to pick and application vendors have a hard time making their apps work on multiple platforms. With the ODPi standard, customers have the confidence that they are picking a certified solution with an array of applications that have been made for it and will run successfully. And if you don’t like the support that you’re getting from your platform provider, you can switch with confidence, since your applications will still work. The existence of a standard will allow the Hadoop ecosystem and the applications that run on it to flourish, leading to greater business value for customers.
Jones, Capgemini: The objective of ODPi is to develop a stronger and more innovative ecosystem by providing vendors and companies a firmer and more assured base from which to work from. The challenges have been the poor interoperability and fragmented initiatives in the market and competing technical solutions. ODPi makes it easier to deploy big data solutions and data-driven applications for a whole range of use cases, because of the cross-compatibility and standardization of this common reference platform.
Moore, IBM: Compatibility has been a huge issue for customers, partners and application ISVs who write on-top of Hadoop. ODPi will allow developers and organizations to have confidence that a distribution will most likely be able to run the same applications with other distributions that run this same application.
ODPi will drive interoperability and compatibility by providing a common platform against which to certify applications. It also opens up choices for developers by enabling this Interoperability with different distributions within an organization.
The ODPi test framework and self-certification also aligns closely with the Apache Software Foundation by leveraging Apache BigTop for comprehensive packaging, testing, and configuration. More than half the code in the latest Big Top release originated in ODPi.
Keenan, UNIFi: ODPi is providing standards sorely needed within the Hadoop ecosystem. This will enable a more mature, solidified, and hardened stack as vendors continue to adopt this standard. It will remove the barriers of distribution and vendor lock in.
2. What do you see for ODPi in the future?
Maciag, Altiscale: ODPi provides reassurance to our customers that whatever they build on Altiscale will be something that they can use in other places and in other situations. It gives them confidence in using a vendor. It also allows them to run applications in a hybrid environment, for example, using Hortonworks on premises and Altiscale in the cloud. It also allows customers to easily transition from an on-premises environment to a fully cloud one. We are talking to many customers who want to make a transition to the cloud and they are running on a different Hadoop distro today. Since Altiscale is ODPi compliant, they are comfortable that they can easily make that transition.
Jones, Capgemini: The next question for ODPi is what other pieces need to be standardized but also how to move towards a greater degree of verification-based compliance. Certainly as our clients see the tangible proof of how ODPi can reduce complexity and costs, we can expect adoption to accelerate.
Moore, IBM: We expect ODPi to grow. There is significant innovation happening in the industry right now and we like that ODPi will be able to provide a solutions for these challenges. By working together with the other member companies, we can enable organizations to better innovate with Hadoop and other big date technologies thanks to standardization.
Keenan, UNIFi: ODPi has made great strides and is quickly maturing into a standard with a ground swell movement behind it. Our expectation and belief based on these early successes is that this program will continue to see additional vendors and platform providers get on board. Hadoop has won the battle as the Analytic platform of choice, now it is time to build the appropriate standards to accelerate innovation and solution development.
3. Can you share some ways ODPi benefits your customers?
Maciag, Altiscale: The members of ODPi strongly believe that customers and the value that they get from Hadoop should drive the industry, not any particular vendor. When customers win, the industry and vendors win, too. The release of the runtime spec shows that standards can be established, that the vendors can work together. In the future, I look forward to the expansion of the spec and you’ll start hearing more and more about the real world benefits that this provides.
Jones, Capgemini: ODPi helps customers by enabling an ecosystem of suppliers who can work across multiple Hadoop vendors. This enables procurement and IT departments to focus on supporting the business ambitions rather than sorting out the technical issues of interoperability and competing technologies.
Moore, IBM: Within the IBM ecosystem, big data plays a tremendous role. As we are beginning to dive deeper into the cognitive world, there has been an explosion of structured and unstructured data feeding into it. Hadoop, Big Data, machine learning, they all plays into how this cognitive world will evolve. So having a platform to count on, starting with a small, specific and consistent packaging model that lives within the ecosystem, is priceless. As people take advantage of this and deploy into their own cloud infrastructure, they will be able to test once and run everywhere. In this way, standardization is an enabler and is a huge deal as we move into the cognitive era. IBM is looking forward to distributing ODPi applications. This will be a key differentiator for IBM to serve customers.
Keenan, UNIFi: Our customers will benefit by reaping the rewards of a more mature data processing platform that they can build their future analytics initiatives on. They also benefit from a Unifi perspective as we’ll be able to spend more resources delivering features, capabilities, and solutions that solve their business problems versus spending cycles supporting all variants of services across a no standards-based processing platform.
4. What advice would you give to others looking to join ODPi?
Maciag, Altiscale: If you’re an application vendor looking to accelerate your business and find more customers, this is the place to be. ODPi certification takes friction out of the sales process and provides peace of mind in a rapidly changing Hadoop ecosystem. By joining ODPi, you get to help influence the standard and know sooner about upcoming developments, so that you can take advantage of them more quickly. The membership of companies today really represents leading thinkers and makers in Hadoop, and so it’s a great group to join.
Jones, Capgemini: Get involved, look at what you want ODPi to do next in practical terms, and join with the ambition of helping to make that happen.
Moore, IBM: You do not have to be a member to join. Goals are self-explanatory, organizations can self-certify and everything can be found on Github. If you are a company, ODPi is a great community and has made significant progress, quickly. Please join us and anticipate in this evolution. It is certainly something Hadoop has needed for years and years.
Keenan, UNIFi: We would encourage participation. This consortium is good for the community and market as a whole and will only create additional value as more members contribute to the program.
5. Why did Capgemini join ODPi?
Jones, Capgemini: Capgemini joined ODPi because we believe in open standards ability to drive innovation. From our involvement with Open Group, to more technical and sector focused standards effort, we’ve consistently found that standards help drive innovation as well of course bringing the benefits of streamlining and cost efficiency.