Pliops said that it was looking to address the unprecedented scalability challenges raised by the cloud data explosion, the increasing data requirements of artificial intelligence (AI) and machine learning (ML) applications, the recent shift of work, education and entertainment online – and the end of Moore’s Law, in as cost efficiently a manner as possible.
“As the success of GPUs and AI processors has shown, there is a pressing need for specialised solutions that accelerate the pace of workflows,” noted Steve Fingerhut, president and chief business officer for Pliops. “Because data growth is absolutely booming at a time when storage and data processing is slowing, now is the time to take this approach and apply it to mainstream workloads like unstructured and structured databases, analytics and virtually any application that uses flash storage.”
A simple-to-deploy PCIe card that delivers more than 10x performance benefits for databases, server and storage infrastructures, the Pliops Storage Processor represents a new spectrum of dedicated processors, according to the company.
“Data centre architects are working hard to keep up with the data explosion, which is only accelerating with the new work-from-anywhere phenomenon,” said Peter Zaitsev, co-founder and CEO, Percona. “Enterprises are accelerating the adoption of open-source software to address these challenges and are constantly looking for industry-standard innovations to achieve more efficient scaling. We benchmarked the Pliops Storage Processor’s ability to scale databases, using Percona Monitoring and Management to show graphically how the platform behaved. Our results show that the Pliops Storage Processor can reduce customer costs by delivering substantial increases in performance and space efficiency."
By significantly improving workload performance and optimising SSD usage, the Pliops Storage Processor reduces storage and processor costs, enables system consolidation, makes much lower cost QLC SSDs viable for mainstream workloads, and is universally applicable with zero application changes.
Applications accelerated by Pliops include MySQL, MariaDB, mongoDB, Redis, Oracle, Apache Spark, and Cassandra and this is achieved through an innovative architecture that overcomes the major inefficiencies of software-only storage engines via a dedicated hardware product.
“The way we are addressing storage and database-specific processing challenges is in the same vein of other workload-specific processors, like GPUs and TPUs,” explained Fingerhut. “The Pliops approach accelerates compute-intensive functions and eliminates bottlenecks – solving the dilemma of choosing cost at the expense of performance – and vice versa.”
Pliops’ Storage Processors are now sampling to select customers, with mass production on track for later this year.
