Aerospike has made available an update to its multi-modal distributed database that can now support online transaction processing (OLTP) applications.
Company CTO Srini Srinivasan said the Database 8 edition of the Aerospike real-time distributed database guarantees strict serializability of transactions that comply with rules for atomicity, consistency, isolation and durability rules that are needed to process transactions concurrently.
Previously, the Aerospike database could run millions of transactions per second (TPS) involving single-record requests with sub-millisecond latency for single-record requests. Aerospike Database 8 expands data consistency guarantees with strict serializability to multi-object transactions requiring low latency, said Srinivasan.
At its core, the Aerospike database runs in memory to provide organizations with a NoSQL alternative to a traditional relational database. That approach provides IT teams with a schemaless multi-model database capable of supporting key-value, document, vector and graph-based applications.
The ACID capabilities now extend the capabilities of the Aerospike database into the realm of OLTP applications to enable IT teams to further rationalize the number of databases they might otherwise need to deploy and manage, said Srinivasan. Exiting Aerospike customers include Adobe, Airtel, Criteo, DBS Bank, Experian, PayPal, Snap and Sony Interactive Entertainment.
There is, of course, no shortage of database options, but organizations typically make a tradeoff between cost and performance. Providers of databases optimized for specific types of data typically argue that this approach enables IT teams to more effectively run specific classes of applications at scale. Advocates of multimodal databases generally contend that the total cost of IT is much lower when IT teams opt to rely on a database platform capable of supporting multiple classes of applications, many of which may require access to multiple types of data.
Aerospike, in contrast, is making a case for a multi-modal database running in-memory that is fast enough to support the needs of most applications running in a production environment.
It’s not clear how many classes of databases the average IT organization is now supporting, but as the types of data being incorporated into applications continues to expand, there is little doubt IT is becoming both more complex and costly to manage. Each IT organization will need to determine for itself to what degree to rely on multimodal versus single-purpose databases.
In many cases, the decision regarding which type of databases to use is initially made by the application developer, who then hands off an application to an IT service management (ITSM) team to maintain in a production environment. It’s not uncommon for those teams, however, to reduce costs by switching the underlying database initially used to one that is a corporate standard.
Regardless of the type of database employed, IT leaders would be well-advised to ensure that the next time business leaders sign off on a new initiative, they actually understand the financial implications of managing data in a modern era where with each passing day there seems to be yet another type of data platform being added to an IT environment.
