Download PDF
Driving real-time understanding of natural language with scalable, high-performance storage
Technology Category
- Infrastructure as a Service (IaaS) - Cloud Storage Services
Applicable Industries
- Telecommunications
Applicable Functions
- Product Research & Development
Use Cases
- Speech Recognition
Services
- Cloud Planning, Design & Implementation Services
The Challenge
Nuance Communications, Inc. is a company that specializes in voice recognition, natural language understanding, reasoning, and systems integration. They aim to create technology that makes it easier for humans to interact with computers. To run live customer-interaction environments for its clients, and to refine the speech models that power them, Nuance relies on high-performance computing (HPC). They compare spoken words against a language model that is a statistical representation of all the data previously collected. As well as using HPC to enable real-time understanding of speech, they continually fine-tune their underlying language models to improve their speed and accuracy. To deliver on its promise of facilitating human-computer interactions, Nuance must ingest, store and analyze huge—and ever-growing— volumes of speech data in real time.
About The Customer
Nuance Communications, Inc. is a multinational computer software technology corporation that provides speech recognition, and artificial intelligence. Nuance creates advanced, self-optimizing natural speech-recognition and processing technologies that are helping to define the next generation of human-computer interaction: intelligent systems. With 12,000 employees across 39 countries, Nuance helps organizations deliver intelligent self-service solutions that enable employees, customers and partners to quickly and easily get what they need from contact centers, websites and mobile apps.
The Solution
Nuance uses IBM Spectrum Scale to give its HPC community fast, massively parallel access to billions of files, with automatic policy-based storage tiering that balances storage cost and performance. Spectrum Scale provides a hardware-agnostic, software-defined file storage landscape with a single global namespace. Nuance ingests large amounts of voice data every day, storing them first as Swift objects in a private OpenStack cloud. As required for HPC, the company then pulls data into the Spectrum Scale landscape—currently multi-petabyte. Data stays on SATA disks, while metadata is stored on SSDs—giving a significant performance boost for a small investment. Nuance is currently testing the use of flash storage as a new top tier. Nuance uses the Spectrum Scale policy engine heat map to automatically move the most frequently accessed files to the fastest disk systems, and to demote unused files to slower, less costly storage.
Operational Impact
Quantitative Benefit
Related Case Studies.
Case Study
Vodafone Hosted On AWS
Vodafone found that traffic for the applications peak during the four-month period when the international cricket season is at its height in Australia. During the 2011/2012 cricket season, 700,000 consumers downloaded the Cricket Live Australia application. Vodafone needed to be able to meet customer demand, but didn’t want to invest in additional resources that would be underutilized during cricket’s off-season.
Case Study
SKT, Construction of Smart Office Environment
SK T-Tower is the headquarters of SK Telecom. Inside the building, different types of mobile devices, such as laptops, smartphones and tablets, are in use, and with the increase in WLAN traffic and the use of quality multimedia data, the volume of wireless data sees an explosive growth. Users want limitless Internet access in various places in addition to designated areas.
