Mapping Session results: Cloud databases

At our Structure:Data conference, GigaOM Research hosted a Mapping Session to assess the market conditions around cloud databases. Mapping Sessions are interactive break-outs facilitated by analysts that tap into the GigaOM community to test hypotheses on new technologies, business models, customer trends, etc. They often feed into and set up more detailed analysis we deliver through Sector RoadMap reports.

The GigaOM analyst panel sketched out current market conditions surrounding cloud-based database-as-a-service offerings:

  • Hype surrounding MongoDB, due to its popularity with developers and use of familiar programming tools (REST, Javascript). However, it’s garnering less enthusiasm from IT operations: it’s easy to get up and running, but not to manage.
  • Big suppliers like Amazon, Google, and Microsoft are placing multiple bets on different technology approaches.
  • Cloud database offerings are picking off the low-hanging fruit by leveraging classic cloud values – low start-up cost, scalability – but are not addressing mission-critical functions. They’re doing fine at analytics tasks, but haven’t moved into transactional roles.

So what new technologies, business issues, and buying patterns can suppliers harness to gain ground in this market?

Assessing market disruption

We’ve done Sector RoadMap reports on categories like the Platform as a Service market and SQL on Hadoop platforms. Sector RoadMaps are collaborative research efforts that match up competitors’ abilities to align with what we call Disruption Vectors, i.e., the key technology or market forces that drive emerging markets. Smart vendors can ride Disruption Vectors to gains in revenue or market share. Throughout the interactive session, the panelists and session participants described and assessed various potential disruption vectors. Then we took a poll of all the participants to see which ones had some consensus as the most important market forces, and which might have a secondary, or longer-term influence. The chart below illustrates the results:

Source: GigaOM Research Mapping Session, March 2013

Time to deployment. The overwhelming – if, perhaps, conservative – consensus of the session participants was that winners in cloud database would continue to leverage the classic “blank-as-a-service” advantage of quick-and-easy set-up.

Tools. Complementary visualization tools have enabled Hadoop to disrupt the business intelligence sector. What will be the tools that move cloud databases deeper into transactional environments? And which management tools will smooth integration and operations, doubling down on the ease-of deployment theme?

Security. Andrew Brust called it a “buzzkill,” but security and encryption processes are a requirement for many regulated industries and for most serious IT departments. David Linthicum pointed out that it’s possible to deliver security from a cloud database, but not without serious performance hits. There was debate over whether there can be such thing as an abstracted security model – as described at Structure:Data by the CIA’s technology chief – that would support multiple clouds or distributed cloud databases.

Cost. GigaOM research director Jo Maitland observed that total cost of ownership often wins the day, and that new offerings with dramatically lower price points can be hugely disruptive, even if they don’t necessarily translate to a winning strategy. If new players threaten the established pricing structures of Oracle or Teradata, for instance, those structures will either collapse…or lead to M&A activity.

“One database to rule them.” In his SQL-on-Hadoop Sector RoadMap, Joseph Turian described the potential of a Hadoop-like analytics database with the traditional ACID characteristics of an OLTP database. A wonderful thing, no doubt. But not something that’s on the near-term horizon. Right now, it’s not clear whether that transaction/analytics mix is best delivered by an abstracted common interface for connecting multiple databases, or by a single database with separate performance “tune-ables.”


We welcome your feedback on these disruptive trends, and on what might accelerate the success of cloud databases. Have we missed or mis-emphasized anything that you believe will be key to driving the sector over the next 12 to 24 months? Continue the discussion by leaving a comment below.

Mapping Session Panelists