Couchbase is cracking open its NoSQL document store’s architecture by allowing core database services to run on different server parts.
The firm announced Monday Couchbase Server 4.0, featuring an architecture it's developed called Multi-Dimensional Scaling.
Multi-Dimensional Scaling will let you isolate query, index and data services in Couchbase, allowing users to optimise according to workload. Couchbase Server 4.0 is due to appear in the second half of this year.
Couchbase chief executive Bob Wiederhold in a statement said Multi-Dimensional Scaling meant the server could support a broader set of applications.
"This improves performance, reduces hardware costs, and enables enterprises to support a much broader set of applications with a single database: Couchbase Server,” Wiederhold said.
Couchbase made its point by juxtapositioning the news against NoSQL rivals MongoDB and Cassandra and the one whose lunch crumbs they all want to eat – Oracle.
Customers include Tesco, Amadeus and BetFair.
Other features planned for Server 4.0 are SQL for Documents, to attract database heads used to working with the dominant relational architecture that NoSQL once opposed.
Also coming is ForestDB, a storage engine built to work with multi-core processors and Solid-State drives (SSDs).
Couchbase, however, is leading on Multidimensional Scaling in Server 4.0, meaning it would no longer run as a monolithic whole, with all workloads treated equally running through the same parts of all nodes.
Multi-Dimensional Scaling lets you speed up an application’s performance without taking the standard approach of indexing or bolting on more nodes.
The former makes applications more complicated and can hit performance. The latter means a bigger hardware bill, while adding more nodes can actually slow things down over the long run, so you’re no further ahead.
By cracking open the Couchbase architecture, workloads can now be allocated based on their unique needs. Queries that suck up large amounts of CPU resources can be allocated accordingly, while indexes can be directed towards high-performance SSDs. Meanwhile, data – which is less CPU intense – can be driven in the direction of memory.
Couchbase is the latest NoSQL player to try and up its ante in the enterprise. Rival MongoDB last month sought to make its NoSQL database more scalable with the introduction of a backend architecture and storage engine – WiredTiger. ®