Teradata's recent Appliance for Hadoop product announcement highlights big data's continued march toward enterprise readiness. If big data technology is to become an operationalized tool, as Ovum expects, prepackaged platforms or solutions will prepare the way. As a product, the appliance signals Teradata's acceptance of the importance of Hadoop implementations that are as open as possible, a shift away from the company's previous approach.
Teradata, Cloudera, and Hortonworks: flexibility to grow enterprise confidence
Announcing 3Q15 availability of the next-generation Teradata Appliance for Hadoop with Cloudera (Cloudera Enterprise 5.4) / Hortonworks (HDP 2.3) and support for Teradata's QueryGrid coming in 4Q15 marks a positive (and enterprise-friendly) step toward the mainstreaming of Hadoop as an accepted business tool. It is also reflective of the evolution of Hadoop from just "store it all" data lakes and specialized Internet-scale use cases to a set of more easily recognizable enterprise use cases – from cold storage (Ovum has seen a big uptick in inquiries on this subject) and high-performance, fast data workloads (IoT is coming, like it or not).
In my opinion, Hadoop has been too complex to deploy without a broad range of specialist skills (both analytical and technical) and has lacked business leadership's critical "I get how this makes stuff better" moment. Teradata, a company with epic brand recognition among data-driven enterprises, is among an exclusive set of vendors ideally positioned to make Hadoop big data technology enterprise-friendly; and, an appliance is the means to address some of that complexity in actually deploying it.
The appliance comes in three different flavors, which highlight the growing variety of use cases and workloads associated with Hadoop deployments. Balanced, Capacity, and Performance nodes offer a blend of different Intel CPU, RAM, and hard disk options that reflect use cases from real time / streaming through less-time-sensitive analytical use cases and cold storage. While it is possible to identify realistic examples for each of these options, with Teradata's RainStor product, the cold storage of sensitive or regulation-governed data (for example, from decommissioned banking applications) offers a clear-cut "enterprise-grade" solution to a growing problem for many financial institutions.
The addition, for example, of Cloudera's capabilities around data governance, security, and flexible cloud deployment options add further enterprise cred to the appliance. Teradata has also published its certification schedule for new Cloudera and Hortonworks general availability launches (+30, +60, and +90 days), with RainStor notably featuring in the +30-day schedule.
Overall, I view this announcement as strongly positive for the future of Hadoop. While not wholesale new, it further enhances Hadoop on two fronts: ease of deployment / manageability, and growing credibility in the enterprise, linked to better-defined use cases. I look forward to seeing case studies for this product: it's time for Hadoop to shake its overly technical image and join the toolbox of enterprise IT.
Tom Pringle, Practice Leader, IT – Information Management