VoltDB Updates In-Memory Operational Database
By enhancing the familiarity of SQL with capabilities that allow organizations to accelerate development cycles, streaming analytics, and time-to-market, VoltDB hopes to shine a light on enterprises' once "dark data."
The company today released Version 5.6 of its in-memory operational database, which melds streaming analytics and transactions to support mission-critical, real-time applications. The latest version features updated capabilities for importing and exporting data streams and Active/Active database replication, according to VoltDB.
The developer wrote its database from the ground up, said Bruce Reading, president and CEO, in an interview with EnterpriseTech. The challenge: Helping enterprises create usable insight from the majority of data stored in inaccessible or unused silos of Hadoop, OLAP, or others, he said. Speed was a necessity, said Reading, and the solution must be optimized for in-memory deployment.
"If you can't do something with it the moment it appears – milliseconds or microseconds – you lose the optimal window for business impact," he said. "Most, not all, enterprises are struggling mightily and are stil in the dark about how to address business challenges. It's all about creating value before your data becomes dark."
In version 5.6, app developers can import various data streams from Kafka to various tables or stored procedures, according to VoltDB. They also can export data to multiple targets, such as Kafka and Hadoop, at once, the developer said. The database now supports Kafka v0.8.2 and synchronized export and allows users to export VoltDB data to Elasticsearch servers or clusters, enabling them to do full-text searches.
In addition, VoltDB allows the replication of distributed workloads, including geographically distributed workloads. The system compresses data to accelerate WAN transfers, and enterprises now can partition database replicas with rack-awareness to improve availability in the event of power, rack, or chassis failure, the developer said.
Typically, VoltDB customers fall into two categories, said Reading.
"Type one is a company that is serving their customers and has apps that are falling down because their legacy apps were never built to handle the amount of data that is now being generated. Basically the app chokes or bottlenecks. Those apps are always centered on operational interactions," he said. "The second type is a company that has been capturing data, is not getting value from it in terms of not being real-time – it's five minutes, 10 minutes, an hour – it's getting insights, but those insights are not actionable because it's too late. Think of the application around providing context to a mobile user: I'm not going to wait much more than a few seconds before I hang up and try again."
Speed was one reason Openet selected VoltDB, said Michael O'Sullivan, global vice president, engineering, at the independent supplier of real-time BSS (business support systems) to communication service providers, in a statement. "Its new low-latency, highly efficient Active/Active database replication offering provides a key capability needed for high-availability geo-distributed telco applications," he said. "VoltDB delivers real-time analytics and high-velocity transactional processing in one easy-to-use platform, enabling us to reduce server and memory requirements, meet tough SLAs and deliver greater ARPU to Openet's tier-one global customers."
The latest version of VoltDB is currently available for immediate download.
