Oracle today announced availability of new bare metal Oracle Cloud Infrastructure compute instances to help enterprises run various performance sensitive workloads, such as artificial intelligence (AI) or engineering simulations, in the cloud. These new instances are part of Oracle’s new “Clustered Network,” which gives customers access to a low latency and high bandwidth RDMA network. Oracle now provides large organizations with a complete set of solutions for any high performance computing (HPC) workload, enabling businesses to capitalize on the benefits of modern cloud computing while enjoying performance comparable to on-premises compute clusters at a cost that makes sense for their businesses.
Countless enterprises have considered a transition to the cloud for years for their legacy HPC workloads, but were unable to do so due to the lack of high performant cloud offerings or economics that would make it too expensive to run in the cloud. Organizations have struggled to find a cloud that can support new workload requirements while realizing cost efficiencies and flexibility that meet business goals and overall technology vision. With today’s news, organizations finally have a low-cost path to extend their on-premises HPC workloads to the cloud without sacrificing performance.
“HPC has been underserved in the cloud due to lack of high performance networking (RDMA) and unappealing price/performance. We’ve listened to our customers and over the last few years Oracle has focused on improving high performance bare-metal offerings, such as Clustered Networking, to provide on-premise customers with the options they need to extend their HPC workloads to the cloud,” said Vinay Kumar, vice president, Product Management & Strategy, Oracle Cloud Infrastructure. “Our growing collaboration with trusted vendors helps Oracle Cloud Infrastructure continue to expand and offer the best performance at the lowest cost for the workloads that customers really need to extend into the cloud.”
With Oracle’s Clustered Network offering at the data center level, Oracle has opened up an entirely new set of use-cases and workloads that enterprises can harness the power of cloud computing for, ranging from car crash simulations in automotive and DNA sequencing in healthcare to reservoir simulation for oil exploration. Organizations can now deploy additional use-cases previously out of their reach, such as using data from their Oracle Database and cutting-edge NVIDIA(R) Tesla(R) GPUs to run Neural Network AI training and instantly adding value to their data.
The new offering is powered by high-frequency Intel® Xeon® Scalable processors and Mellanox’s high performance network interface controllers. For more than 25 years, Oracle and Intel have worked closely to bring innovative, scalable, and secure enterprise-class solutions to its customers. This new addition completes Oracle’s comprehensive set of infrastructure solutions for the full range of both CPU- and NVIDIA GPU-based HPC workloads, providing customers with streamlined and lower cost access to specialized offerings in the cloud. Oracle Cloud Infrastructure now provides a complete range of HPC workloads, including the only bare metal Infrastructure-as-a-Service (IaaS) offering on the market with RDMA.
These new HPC instances with Clustered Networking are offered across Oracle regions in the US and EU at $0.075 per core hour, a 49 percent pay-as-you-go savings compared to other cloud providers on the market.
Additionally, these capabilities will also be expanded for Oracle’s recently announced support for NVIDIA HGX-2 platform on Oracle Cloud Infrastructure. Customers will be able to take advantage of clustered networking as part of the next generation of NVIDIA Tensor Core GPUs, opening up large scale GPU workloads such as AI and deep learning.
“HPC is critical for more use cases, complex workloads, and data-intensive computing than ever before. Access to HPC capabilities in the cloud enables users to do more by extending into the public cloud while providing new HPC and AI users a platform to develop and test new classes of HPC and AI algorithms,” said Lisa Davis, vice president of Intel’s Data Center Group and General Manager of Digital Transformation and Scale Solutions. “We are working with Oracle to enable leading HPC offerings that take advantage of the advanced performance, efficiency and scale delivered by Intel® Xeon® Scalable processors across HPC and AI workloads, all while benefiting from the agility and flexibility of Oracle’s next generation, enterprise-grade cloud offering.”
“As organizations look to ensure they stay ahead of the competition, they are looking for more efficient services to enable higher performing workloads. This requires fast data communication between CPUs, GPUs and storage, in the cloud,” said Michael Kagan, CTO, Mellanox Technologies. “Over the past 10 years we have provided advanced RDMA enabled networking solutions to Oracle for a variety of its products and are pleased to extend this to Oracle Cloud Infrastructure to help maximize performance and efficiency in the cloud.”
“With the massive explosion of data, you need larger clusters of GPUs to process and train the data being produced. HGX-2 on Oracle is the perfect fit for this problem,” said Ian Buck, vice president of Accelerated Computing, NVIDIA. “HGX-2 on Oracle combines cutting-edge V100 Tensor Core GPUs and NVSwitch along with ability to scale across multiple HGX-2s using Clustered Networking to solve the biggest computing challenges.”