Aerospike Inc., the company leading the power of real-time data to meet the demands of the “Right Now Economy,” is announcing the latest iteration of its real-time, multi-model database, Aerospike Database 7. This major release focuses its power on unifying the storage model, ultimately enabling a wide variety of memory advancements that place resiliency, scale, and speed at the forefront.
The management of in-memory, hybrid memory, and all flash storage models can induce rather complex workflows that inhibit the true effectiveness of the technology. Despite adopting in-memory databases for its powerful performance, these solutions leave enterprises with a database that is “often brittle, slow to recover, and a hassle to manage and scale,” according to Lenley Hensarling, chief product officer at Aerospike.
Aerospike Database 7 delivers on a plethora of fronts—including the previously mentioned resiliency, scale, and speed, as well as ease of management and TCO—through its new unified storage format. Users are now afforded the flexibility to choose the right storage engine for different types of workloads, even in the same cluster, according to Aerospike.
The new single storage engine further enables warm restarts for high resiliency, putting data in shared memory instead of process memory (RAM). Restarts occur in mere seconds, not minutes or hours.
“If everything's all in memory and you need to do things like move through and update operating systems under nodes in our very distributed database solution, you can restart those nodes much faster,” explained Hensarling. “You can flush all of the information that's in-memory out to a file system or remote file system, and then re-start those nodes much faster.”
Warm restarts not only ease database management more generally but also assists enterprises with security policies that necessitate periodic refreshes of the OS underneath the database; warm restarts enable this to be done rapidly, easily, and efficiently.
Version 7 of Aerospike’s database also innovates in compression, offering LZ4, Snappy, or ZStandard compression algorithms to maximize memory efficiency. This allows users to obtain the same compression ratios regardless of storage engine, ultimately reducing DRAM hardware costs, according to the company.
With in-memory data now mirrored to persistence layers, Aerospike Database 7 enables in-memory persistence operations—such as defragging, garbage collection, tomb raiding, and re-balancing—to occur solely in-memory without needing to extensively leverage the drive. These in-memory persistence processes only interact with the drive to reflect any mirroring changes.
If users choose to not take advantage of the persistence layer, in-memory namespaces can be deployed with the Aerospike Shared Memory Tool (ASMT), persisting the namespace from shared memory to the filesystem post shutdown.
The single unified storage engine, along with Database 7’s other advances, not only serves a user request purpose, but a future functionality. Designed to support Aerospike’s 7.1 multi-record transaction capabilities, 7.0 works as a foundational piece in driving Aerospike’s ability to address a wider set of applications.
“This whole idea of doing it [the Aerospike database] at scale with high throughput, making it available to a broad set of companies, and doing it in a way that's most cost effective, has been what drives us, forever,” said Hensarling.
To learn more about Aerospike Database 7, please visit https://aerospike.com/.