VoltDB looks to gain ground in crowded in-memory database market
VoltDB is hoping to raise its profile in the burgeoning in-memory database market with a new release that offers improved analytic processing, better scalability and other features.
Version 4.0 of VoltDB has radically increased throughput for analytic queries, the company said Wednesday.
“We’ve removed some of the transactional overhead when running many kinds of global read queries, including ad-hoc SQL,” VoltDB vice president of engineering John Piekos wrote in a blog post. “These queries are still reading a live, fully serializable view of committed data, but they’re now up to 50x faster.”
Customers who upgrade to 4.0 will also gain the ability to more easily add nodes to running clusters, according to VoltDB.
“Since all topology changes and data movement are transactional and durable, your data is protected while the cluster is expanding or rebalancing,” Piekos wrote. “We’ve also carefully engineered this feature so that most customers will see no impact to their workload during expansion.”
The updates align VoltDB with competing in-memory databases that target both transactional and analytic workloads, such as SAP’s HANA.
VoltDB was co-founded in 2009 by MIT professor Michael Stonebraker, who is also known for his work creating the Ingres and Postgres databases.
The company offers a community edition of VoltDB under the GNU Affero General Public License Version 3, but it omits a number of features found in the commercial version.
In-memory databases provide major performance improvements by storing data in RAM, rather than reading and writing it to disks. However, they also typically provide some means to commit data to disk or another storage medium, allowing the database to be brought back up in the event of a failure.
In VoltDB’s case, users can make the system take automatic “snapshots” of the database which can then be used for recovery. VoltDB also offers a feature it calls K-safety, wherein database partitions are duplicated so the system can keep running even if a node goes down.
While the in-memory database market is growing quickly, there are many different types of products, meaning customers should choose a platform carefully, Gartner analysts Donald Feinberg and Roxane Edjlali wrote in a new report.
“The main business value offered by an IMDBMS is that it enables new business opportunities that would not have been possible previously,” the report states. “One example is SMS discount coupons sent to customers based on their preferences when they are in, or close to, a shop. To support such use cases, both the transactional data and the analytics need to be available in real time. The promise of in-memory DBMSs is to enable hybrid transactional and analytical processing platform (HTAP) use cases without having to move the data from one to the other.”
SAP has been working to reorient its entire product catalog around HANA, both by porting existing applications to it and creating an array of associated platform services and development tools.
Oracle has sold the TimesTen in-memory database for some time, but is expected to release a new in-memory option for its flagship database sometime this year. The upgrade will deliver an “ungodly” boost in speed, according to Oracle CEO Larry Ellison .
IBM and Microsoft have also been working on in-memory database technologies.
VoltDB is just one of many smaller vendors hoping to steal business away from those bigger players. Others include Aerospike, MemSQL and Kognitio.