Aerospike, Inc. unleashed a broad set of new features for its flagship in-memory database that provides developers with more flexibility in how they build their applications and addresses several key enterprise requirements. The launch marks the biggest update for the platform since the company released the underlying code to the open-source community in June.

It’s now possible to deploy Aerospike using the Vagrant virtualization wrapper and Docker container, freedom that can potentially make it easier for users to scale their implementations from pilot to production. The addition of support for the container engine is especially significant since it makes the database more appealing for powering next-generation services implemented as multiple discrete instances.

To that end, Aerospike is also rolling out clients for several additional web development languages, including Python, PHP, Ruby and the buzzed-about Go from Google. That enables more kinds of applications to run against the platform, which the company is encouraging with new features specifically geared towards accommodating a greater variety of workloads.

One of these is the ability to optimize the management of data operations on an individual basis, functionality that avoids the trade-off between reliability and speed in services that handle different types of requests. For example, a developer can now have Aerospike verify that an online purchase is properly logged before issuing receipts while allowing less critical but time-sensitive actions such as personal messages to propagate through the system faster.

The new version also enables IT practitioners to make distinctions among workloads, with a new role-based authentication and authorization feature that make it possible to determine who can access what information. That degree of control is crucial in enterprise deployments.

Aerospike has also removed the need to adapt data when shuffling workloads back and forth between its platform and Hadoop. It also implements the company’s patented Indexed MapReduce capability to expose only the specific subsets of data in an Aerospike deployment needed for a particular Hadoop operating instead of scanning through the entire system. Moving information the other way has been sped up as well.

Rounding out the upgrade is a set of improvements under the hood that Aerospike said make the database more efficient. Aerospike can now read recently input information directly from memory, which the company says helped one customer to achieve a 40 percent reduction in flash storage requirements. Overall, the platform can now handle 2.5 million transactions per second on a single bare-metal server and using up six times fewer resources than the competing Apache Cassandra when running in the public cloud.