Getty Images/iStockphoto
Latest Aerospike database update prioritizes performance
The vendor's latest platform update addresses performance at scale by adding multi-object ACID transactions to better support the needs of online transaction processing.
The latest Aerospike database update prioritizes performance efficiency.
Version 8 of the vendor's platform, generally available on Wednesday, adds support for online transaction processing (OLTP) applications with distributed ACID (atomicity, consistency, isolation and durability) transactions, also known as multi-object ACID transactions.
OLTP applications are frequently used in industries such as banking, e-commerce and healthcare, in which transactions are common. However, as transaction volumes increase, some data management systems fail to meet workload needs and their performance decreases as demand scales.
Distributed ACID transactions are a means of improving data management systems' performance.
Atomicity guarantees that all transactions succeed or fail completely. Consistency ensures that all data is stored accurately and uniformly. Isolation enables users to run workloads concurrently but separately so as not to cause overloads. Durability ensures that changes made by a transaction are permanent and won't be lost even if there's a system failure while a workload is running.
Given that multi-object ACID transactions aim to improve database efficiency, which in turn helps reduce the compute power required to run workloads and the costs associated with cloud consumption, Aerospike's update is significant, according to Stephen Catanzano, an analyst at Enterprise Strategy Group, now part of Omdia.
"The update … means [Aerospike] can guarantee the highest level of data consistency while maintaining high performance and scalability," he said.
However, while new for Aerospike, distributed ACID transactions are neither new nor is Aerospike the first database vendor to provide such capabilities. For example, competitors MongoDB and Couchbase added support for distributed ACID transactions more than seven years ago.
Based in Mountain View, Calif., Aerospike provides a multi-model database that supports document, key-value, graph and vector data types. In December, Aerospike upgraded its vector search and storage capabilities. Earlier in the year, the vendor raised $114 million in venture capital funding to finance development of capabilities that help customers develop generative AI tools.
New capabilities
Because of the volume of data needed to train AI models and applications, surging interest in generative AI development has illuminated the need for efficiency when running data workloads.
While Aerospike's update is not focused on AI but on transaction processing, the problems of performance and resulting costs at scale are similar. Just as AI requires large amounts of data to deliver accurate outputs, transactional data workloads now contain volumes beyond what traditional databases were built to handle, according to Srini Srinivasan, Aerospike's founder and CTO.
"Performance and efficiency have always been critical, but as applications scale, high-speed processing and strong consistency have become more pressing," he said.
To meet those needs, Aerospike and other databases need to evolve to provide both consistency and speed, Srinivasan continued.
"The ability to maintain speed while adding strong consistency ensures businesses can operate at scale without sacrificing data integrity," he said.
Focusing on performance and data consistency is not new for database vendors such as Aerospike. In testing, Aerospike's database has historically handled millions of single-record requests per second, whether volumes reach gigabytes or petabytes of data, according to the vendor.
But single-record requests -- a database operation that interacts with one specific record – are no longer enough because of the growing data needs of customers with large transaction workloads.
Multi-object ACID transactions -- transactions that simultaneously access or manipulate data related to multiple records -- meet those needs, according to Catanzano. Unlike single-record requests, multi-object ACID transactions provide the atomicity and data consistency across large data workloads that are now needed.
"Multi-object transactions are … critical for applications where multiple data points need to be updated all at once to prevent errors and maintain data integrity," Catanzano said.
In addition, multi-object ACID transactions are noteworthy because they enable transaction management at the database level, where data is easier to configure than at the application level, where transaction management otherwise takes place, he continued.
Atomicity and data consistency are particularly critical to industries such as finance, gaming and fantasy sports that have high transaction volumes, according to Srinivasan. It is imperative in such industries that account balances are correct and each transaction is accurately processed without losing any data.
"Distributed ACID transactions [enable] companies to build more sophisticated, reliable applications without burdening developers to manage consistency at the application level," Srinivasan said.
Regarding the impetus for adding distributed ACID transactions to support OLTP applications, customer demand played a significant part, according to Srinivasan
In particular, reducing complexity and susceptibility to errors by managing transactions at the database level rather than the application level pushed Aerospike to update its platform.
"It was driven by the needs of large-scale applications and customer demand across sectors," Srinivasan said. "[The update] makes it easier for organizations to scale efficiently and meet their operational needs by bringing this functionality directly into the database with performance intact."
Looking ahead
With Aerospike 8 now available, the vendor will continue to focus on improving the performance and scale of the database, according to Srinivasan. In addition, ensuring that cloud deployments perform properly across the major clouds -- AWS, Alibaba, Google Cloud and Microsoft Azure -- is a priority.
Lastly, as more enterprises invest in developing AI applications to detect fraud, provide recommendations and perform other tasks, Aerospike intends to add capabilities for powering AI and machine learning workloads, according to Srinivasan.
"These priorities highlight our commitment to scaling our technology and business for the next phase of growth," he said.
Catanzano, meanwhile, suggested that a focus on ease of use and expanding its multi-model database capabilities to include more data models would be logical next steps for Aerospike. In addition, more transparency around Aerospike's performance relative to its competitors could attract more customers.
"Providing clear benchmarks and comparisons against specific competitors would strengthen their market position," Catanzano said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.
