Back

Speaker "Yuval Degani" Details Back

 

Topic

Accelerated Spark on Azure: Seamless and Scalable Hardware Offloads in the Cloud

Abstract

Microsoft Azure’s advanced compute and network infrastructure allows Spark to run in the cloud without compromising on performance. With the growing arsenal of hardware offloads available on cloud VMs, owning and maintaining bleeding edge hardware is no longer a prerequisite for accelerated compute. In this talk, we will demonstrate how hardware accelerations in Azure can be utilized to speed-up Spark jobs seamlessly, with the aid of RDMA (Remote Direct Memory Access) support in the VM. We will demonstrate use cases of benchmarks and real-world applications, that achieve impressive performance improvements with minimal configuration.

Profile

Yuval Degani is a Senior Manager of Engineering in Mellanox Technologies, leading a team of Software Engineers based in the San Francisco Bay Area. His team's focus is introducing new network acceleration technologies to Big Data and Machine Learning frameworks. Before that, Yuval was a developer, an architect and later a team lead in the areas of low-level kernel development for cutting edge high-performance network devices. Yuval holds a BSc in Computer Science from the Technion Institute of Technology, Israel.