Microsoft Meets the Demands of the Cloud
Intel® FPGAs Kushagra Vaid explains how Microsoft achieves flexibility and scalability with Intel® Xeon® Scalable processors, from AI to big data analytics. Kushagra Vaid, GM & Distinguished Engineer, Azure Hardware Infrastructure, Microsoft. One of the challenges we have in the cloud is that we have to accommodate a variety of different types of workloads: anything from email, through databases,... through machine learning. So all these workloads are extremely demanding. They require extreme compute processing, and network processing and storage processing. You have to have a base platform that can scale with the demands of these workloads. Intel® Xeon® Scalable processor allows us to meet these demanding requirements. The higher goal counts, the increased memory bandwidth, the new instructions such as AVX-512 all come together to provide that new capability for these workloads. The Microsoft Project Olympus* platform, which is based on Intel® Xeon® Scalable processors and Intel® FPGAs, provides us with a degree of flexibility and performance that helps us to meet the emerging demands of the cloud. The Intel® FPGAs that we are using in Azure help create a configurable cloud that can scale with the demands of these workloads. We see a trend in the data center space. AI and big data analytics are coming close together to power the next generation of enterprise applications. The Intel® Xeon® processors when combined with the Intel® FPGAs are an extremely powerful combination to solve the challenges we see with AI. Microsoft has built the industry’s largest deployment of FPGAs for accelerating detailed networks. Intel accelerators such as Intel® Xeon Phi™, Intel® FPGAs and Intel® Nervana™, along with their optimized set of libraries for the applications, all provide a new level of productivity and performance to make deploying insightful applications in an easier and more powerful way than ever before. And the combination of all these new technologies is what allows us to offer flexibility and scalability for these demanding applications in the cloud.