The shift to the cloud for big data is on.
In fact, global spending on big data solutions via cloud subscriptions will grow almost 7.5 times faster than on-premise subscriptions. Furthermore, public cloud was the No. 1 technology priority for big data, according to Forrester data.
Needless to say, cloud is the hottest market for big data technology.
Public cloud innovation will bury on-premise options
Public cloud plus big data is creating exponential effects, just like Moore's law happened when silicon chips met transistors. My recent research projects that public cloud prices for big data storage and processing will be cut in half every few years, while processing power doubles.
- Your cost estimates for big data in the public cloud are likely wrong. Public cloud skeptics typically project costs as the same or slightly increasing over five or 10 years. When you recognize costs are going down on an exponential curve, things look different.
- The cost of upgrading big data software will look worse and worse. That Hadoop data lake is going to need upgrading before long. Your data scientists already want multiple versions of Spark. And this is just the beginning of your on-premise headaches. It's only going to get worse over time as AWS, Google, and Microsoft pursue a serverless strategy.
- The pace of public cloud innovation is the final killer. Technologies like artificial intelligence (AI) and quantum computing will mostly likely be consumed through public cloud. Quantum vendor D-Wave and Google plan to let customers use quantum computing in TensorFlow, just as an example. Firms who are not ready will be hard pressed to stay competitive.
Insight PaaS have advantages over other cloud solutions
To help our clients in their data journey to the cloud, I completed a Forrester Wave evaluation of eight big data vendors -- 1010Data, Amazon Web Services (AWS), Databricks, GoodData, Google, IBM, Microsoft, and Qubole.
What does this eclectic mix of vendors have in common? They all offer insight PaaS. Forrester defines an insight PaaS as an integrated set of data management, analytics, and insight application development and management components, offered as a platform the enterprise does not own or control.
Insight PaaS offer advantages over other big data solutions that advertise cloud deployment options. First, all these vendors have gone beyond their roots to offer insight application development platforms. They also offer some big, exponential advantages.
For example, my analysis found that insight PaaS are better than on-premises at:
- Managing and accessing large, complex data sets. For example, Google's BigQuery lets developers query petabytes in milliseconds. And you only need to make a few decisions about schema design and cluster size.
- Updating and evolving applications that deliver insight at the moment of action. GoodData's insight application life-cycle management features let customers build and optimize solutions without extensive code debugging or query tuning.
- Updating and upgrading complex technology. Databricks makes the current and older versions of Spark accessible by URL at the same time. Existing applications can keep using older versions until they are ready to migrate. Then they simply point to the new version -- no server upgrade required.
How did vendors rank in Forrester's insight PaaS Forrester Wave?
Although I can't give away the entire Forrester Wave, I can say that our evaluation had three strong performers that offer unique advantages -- and that Google was our only leader. While Google has consistently scored below AWS and Microsoft in many of Forrester's other Wave evaluations, it is a clear leader in insight PaaS. Google offers a PaaS-first platform that features a combination of machine learning, big data, and application development tools.
Congrats to Google.
By Brian Hopkins, vice president and principal analyst at Forrester
For a full list of rankings, check out Forrester's insight platforms-as-a-service Q3 2017 Wave -- subscription required.