On Founding the (Big Data) Foundry

(Contributed by Michael Hay, VP/CTO, Hitachi Data Systems)

Steve Hoover, PARCAs a start-up, an established research organization, an established networking devices plus IT infrastructure vendor, and an established IT infrastructure and devices vendor we observed that a variety of Bring Your Own X events that must transpire before launching a next generation solution.  At PARC, our ability to setup, experiment, and teardown computing infrastructure is the key driver to deliver breakthrough ideas. For example in the PARC and SAP real-time analytics exploration project, we spent 6 months on infrastructure setup before we could start doing data science. In addition very complex computing landscapes were needed to iterate machine-learning algorithms on very large volumes of data. Our lesson: Having PARC researchers, who are highly specialized in solving complex problems, use their precious time for infrastructure plumbing and data wrangling was and is wasteful.  In addition to these, data security and privacy issues can’t be ignored. If there is way to rapidly compose, use and tear down different computing landscapes on-demand, we could have conducted far more data science experiments resulting in more and higher quality outcomes.

Another example:  Boeing’s case, related to Edge on AWS, where they realized BDFQUOTE-SKRthat Amazon doesn’t have loads of architects, data plumbers, and data scientists on hand to answer your every question. Instead Boeing had to Bring Their Own Teams to build Edge. More generally, there are still largely unanswered questions about data privacy, sovereignty, and safety, which not only persist but also due to Black Swan events are gaining steam. Therefore, having reviewed publicly available literature and mashed it up with our own experiences we realized that well there had to be a better way.   And as a group Quantiply, PARC, Cisco Systems and Hitachi Data Systems set out to find a better way.

Lew Tucker, Cisco SystemsTherefore, these observed and experienced industry-wide challenges provided the catalyst to create the Big Data Foundry. As a group we seek to not only attack these challenges, but also explore and resolve new problem spaces. What we’re launching provides a place whereby it is feasible to quickly set up your software stacks, combine talents from the members (i.e PARC, Quantiply, HDS and Cisco), conclude your project and then pick the deployment model best meeting your requirements. In essence we’ve all heard the challenges from the market around Big Data/Advanced Analytics and the emerging space of the Internet of Things typified by questions like: “How can I buy a full and complete stack of software and hardware where this stuff just runs?” Well this is the kind of work that we we’re tackling, and our journey’s just beginning. In fact already we’ve unleashed the PARC researchers as well as selective startups onto Big Data Foundry. What we’ve witnessed are both processing time reductions and provisioning speed improvements, and the key is not only can you work this with the founding members, but you can buy the stack as well.

QUOTE-HDSAs we look ahead to the Internet of Things (IoT) revolution with an estimated 25 billion connected devices expected in the next 5 years we believe the Foundry has a role to play. Specifically, we think that there are profound challenges associated to new application and solution assemblies spanning from people to technology. A quick survey of items that come to the top of our mind include: Development costs, social engineering issues, access to new technologies, risks and concerns related to data governance, and scarcity of resources all represent obstacles on the path to realizing seeing true value from IoT. It is our intention to begin tackling many of these challenges in a pragmatic and executable fashion showing clear value along the way.

With the notion of reality and pragmatism we’d like to end this post with a photo of the team. Interestingly on this day we were concluding discussions about business models, value oriented use cases, how to connect with sales teams and how bind our works to generally available solutions. If you’re thinking about the pedigree of folks and organizations engaged it might be a bit surprising we debated these points. However because we want to make a difference we quickly pivoted towards these areas!

DSC01861 copy

Comments ( 0 )

    Leave A Comment

    Your email address will not be published. Required fields are marked *