Case Study: Nuance Communications achieves faster, cost-effective natural language processing with IBM Spectrum Scale

Driving real-time understanding of natural language with scalable, high-performance storage

Nuance Communications, Inc., a company specializing in natural language processing and speech recognition, faced the challenge of needing to rapidly and cost-effectively ingest, store, and analyze huge volumes of speech data to enable real-time understanding. To meet the performance and scalability demands of its high-performance computing (HPC) environment, it turned to vendor IBM for a storage solution.

The solution implemented was IBM Spectrum Scale, a software-defined storage system. This provided Nuance with fast, massively parallel access to billions of files and featured automatic policy-based storage tiering. The results included a 20 percent performance boost from tiering, non-disruptive scalability for its HPC grids, and optimized storage costs by allowing the company to run at over 80 percent utilization.

Nuance Communications, Inc

Bob Oesterlin

Senior Storage Engineer