FPGA workload acceleration using Red Hat Enterprise Linux
Jon Masters, Red Hat
Field programmable gate arrays (FGPAs) and other programmable logic devices are rapidly gaining traction for emerging data analytics and processing workloads. Their power efficiency and performance advantages are numerous, but FPGAs have traditionally required extreme levels of technical knowledge to use. This is changing, thanks to industry standardization efforts. In this session, attendees will learn about FPGAs and how they can be managed at scale using emerging efforts within the larger open source community, including OpenStack and Kubernetes orchestration.
Modernizing your HPC applications to take full advantage of the open hybrid cloud
Steven Carter, Red Hat
Adopting cloud technologies and principles in the development and deployment of HPC applications can yield substantial benefits in portability and scientific throughput. This session covers the impediments to running existing HPC applications in the cloud and details how to avoid them.
Dynamic API back-end selection based on real-time health checks through 3Scale API management and OpenShift
Tom Corcoran, Red Hat
Veer Muchandi, Red Hat
Red Hat OpenShift is a game-changing containerization platform that runs your application workloads as containers at scale. 3Scale by Red Hat is an API gateway solution. While applications APIs can be managed centrally by 3Scale, occasionally an API back-end application may become unavailable. We will demonstrate the implementation of a circuit breaker pattern using a polling mechanism into the 3scale configured API gateway solution running on OpenShift. We will show how to configure the API gateway with primary, secondary, and even tertiary back-end servers to handle any given 3Scale configured API request. The polling mechanism is configured to test the health of each primary (and, if necessary, secondary and tertiary) back end every X miliseconds. This results in a recently proven-to-be-available back end serving each request. It improves application reliability and availability, as well as latency, as there are order-of-magnitude fewer waits for API timeouts.
Fast as lightning : NVMe and NVDIMM's Impact on your data center
Tom Coughlan, Red Hat
The future of storage is memory, and the future of memory is storage. The increasing availability of nonvolatile memory express (NVMe) devices for use across fabric infrastructures and the growth of persistent nonvolatile dual inline memory module (NVDIMM) memory is rapidly changing datacenter architecture. Learn what these technologies are, how you can leverage them, and what's available in Red Hat Enterprise Linux.