Cognitive/AI applications operate across large data sets to detect, learn, decide and respond autonomously. Advances in CPU, GPU and FPGA co-processors are dramatically increasing performance while reducing the cost of Cognitive/AI computing. Unfortunately, data performance at scale has not kept pace with these compute advances and is fundamentally limiting application performance.
What's required is a shared data platform that delivers hot tier performance at Cloud scale and economics. Vexata's Accelerated Data Architecture achieves unparalleled throughput at ultra-low latency for the highly parallel, random, mixed read/write data patterns that characterize Cognitive/AI applications. This 3-stage software architecture delivers unmatched performance and efficiency across massive data sets without relying on caching or tiering.
VX-OS, at the core of Vexata's Accelerated Data Architecture consists of 3 primary architectural elements:
VX-OS runs as a distributed software architecture within the VX-100 and VX-Cloud hardware subsystems to deliver unmatched performance with ultra-low latency. Together, this powerful, efficient system supports highly parallel, random, mixed R/W data patterns that characterize Cognitive/AI applications. Vexata achieves this transformative performance across large data sets without relying on caching or tiering.