With the IT world bracing for the Big Data avalanche – according to a new Gartner study, 42% of IT leaders stated they had invested in big data technology, or were planning to do so within a year – database technology, whether traditional RDBMS, i.e. Oracle, Not Only SQL/NoSQL (Couchbase) or alternatives like MySQL (ScaleBase), is under tremendous pressure to evolve to handle bigger, faster and more complex requirements. While NoSQL is the up-and-comer grabbing most of the headlines, bringing in over a billion dollars in revenues last year, that still leaves around $29 billion for the other DB vendors, like the MISO (Microsoft, IBM, SAP, Oracle) oligopoly. For a little better understanding of the database wars, we spoke with Couchbase and ScaleBase for their perspectives on some of the factors behind this fast-changing segment.
Headquartered in Silicon Valley, and the company behind the Couchbase open source project and Couchbase Server, the company’s flagship NoSQL document-oriented database, Couchbase says its hundreds of customers with production deployments include AOL, Cisco, LinkedIn, Orbitz and Salesforce.com. Across the continent, just outside of Boston, Newton-based ScaleBase is a MySQL vendor that dynamically scales out relational databases.
Much like Cisco’s many-clouds thesis, ScaleBase’s Paul Campaniello, VP Global Marketing, believes organizations need to take a hybrid approach to databases, i.e. one size will not fit all. NoSQL technologies are typically used in one of two scenarios: queries that require a very short response time; or storing data without a well-defined or a frequently modified schema. Other needs, like data backup, complex joins queries, consistent data storage – all are still being delivered by relational databases.
“We’re more of a NewSQL than a NoSQL,” he said. “We’re about scale and about transactions.”
NoSQL can be divided into four categories, said Couchbase CEO, Bob Wiederhold: wide column store, document store, key-value store, or graph store. MongoDB is the lead in the document segment, Cassandra in column, Neo in graph, and Couchbase in pure key-value, he said.
“We think the key and document segments are in the process of merging into one. Document is just one element of key value. We think that will be the biggest segment of the market.”
Wiederhold said column is a large niche, and graph is a small niche, and Couchbase competes with both of them, depending upon use cases.
Big Data is only one of the trends impacting the DB market, he said. There are also big users, and the move to the cloud. “Your apps need to support increasingly large number of users.” With the cloud, you need to support hundreds of thousands of users, and in some case 10s and 100s of millions of users and that drives scalability. “You need to be able to quickly and easily scale, and that’s a real problem with the relational database.”
The cloud introduces another element of scale, said Wiederhold. CRM systems used to only support internal needs, but with applications like salesforce.com, you’re supporting tens of millions of users.
So customers are asking for databases with easier scalability, higher performance and are easier to develop for. So on the operational side – as opposed to analytical – whether you’re interested in using MongoDB, Cassandra or Couchbase, that’s what you’re looking for.
Like Campaniello, Wiederhold doesn’t believe RDBMS vendors are going away any time soon. They do some things very well, but increasingly NoSQL will be the choice, especially for newer applications. “Where relational technologies were 95% of the $30 billion database market, in 10-15 years they will be a lot less.”