On Thursday Oracle announced another bad quarter. Oracle blamed sales execution for the miss last quarter, this time it blamed the economy. I’m not so sure. The accelerating shift away from relational databases to new technologies like NoSQL and Hadoop is likely having an increasing effect and contributing, at least in part, to the disappointing sales results.
Eighteen months ago Marc Andreesen of the Andreesen Horowitz venture capital firm noted there wasn’t a single company in his portfolio that was using Oracle as its database and that this would be a strategic problem for Oracle down the road. Many scoffed at the comment. But since then I’ve made a habit of asking other VCs what they’re seeing – and not a single one has had a different experience than what Marc described. I'm not surprised. We rarely see Oracle in competitive engagements unless we’re replacing them.
VCs invest in innovative start-ups that have no ties to legacy technologies and are comfortable with new technologies such as NoSQL. Because these startups are creating the next generation of applications and services, their use of technology, and the products they build with it, provide a glimpse into the future of application development and delivery overall. Marc and other VCs saw the data management choices of their portfolio companies as signs of vulnerability and a leading indicator of future problems for Oracle.
New applications being built today – and not just by startups – are very different from those of the past. They are cloud-based, highly scalable, highly distributed, and increasingly mobile applications. They are data-centric and process increasingly large amounts of unstructured and semi-structured data. They need to support 100s of thousands, and sometimes millions, of simultaneous users who access these apps 24x7x365. And the pace of innovation in most of the markets where these applications play requires an agile development environment that enables developers to rapidly deliver new data-centric features to their customers.
The requirements of today’s applications are generally better aligned with the capabilities that NoSQL databases deliver. A flexible data model that is better suited to unstructured and semi structured data. Easy, dynamic, and reliable horizontal scalability. Consistently higher performance. 24×365 availability. And agility that results from simpler and easier to write code.
NoSQL and Couchbase adoption initially took off with “green field” opportunities at startups and Internet companies – that’s where it was most common to find new app dev projects where emerging technologies were embraced and legacy technology was not a part of the equation. But it’s not just “green fielders” anymore. Over the last 12 months there has been a big increase in replacement of existing data management technologies. If you developed a web or mobile app from around 1995 to 2002 you didn’t have a choice but to develop it with Oracle (or DB2 or SQLServer). Now, companies are increasingly seeing the benefits of NoSQL and switching. Scalability, performance, availability, ease of development and much lower cost are the major catalysts for change. Likewise, if you developed a web or mobile app between around 2002 and 2010 you likely developed it using MySQL (or PostgreSQL). We’re also seeing a big increase in data layer replacements of these applications as well.
As NoSQL feature sets continue to evolve rapidly and as broader ecosystems develop, the trend toward NoSQL will only accelerate. Over the next 24 months I expect Oracle’s database sales to come under increasing pressure. NoSQL will be one of the catalysts, but it’s just one of a number of technologies having similar effects. The early indicators are clear, and it’s just a matter of time before we officially designate these shifts as a major technology disruption.