We’re extremely proud to collaborate with Amazon Web Services (AWS) and continue to deliver ground-breaking technology to their customers. For over a decade the engineering relationship between Intel and AWS has provided tailored solutions for the most demanding workloads. This week’s announcement of the new Amazon EC2 instances for large, in-memory applications and workloads demanding high network throughput, both powered by 2nd Generation Intel® Xeon® Scalable, is another milestone born of our close collaboration with AWS.
Double the Available Memory
The previous EC2 High Memory instances offered 6, 9 and 12 tebibytes (TiB) of memory. The bare metal instances launched this week extend these options to an 18 and 24 TiB of memory and are powered by eight Intel® Xeon® Platinum 8280L processors. The large memory capacity is optimal for customers running demanding in-memory databases such as SAP HANA.
Each instance also has 28 Gbps of dedicated Elastic Block Store (EBS) bandwidth which enables high-speed loading of data into memory. According to Jeff Barr’s announcement on the AWS blog this week, his team loaded 9 TB of data in just 45 minutes, an effective rate of 3.4 gigabytes per second. Look here for the detailed specs and an overview of the scale-up and scale-out options that are possible when using these new instances to run SAP HANA.
Faster Machine Learning
Also announced this week is the availability of Amazon EC2 instances that utilize up to 100 Gbps of network bandwidth. The new M5n, M5dn, R5n, R5dn instances run on a custom 2nd Generation Intel® Xeon® Scalable processor and provide support for the new Intel® Deep Learning Boost instructions and our latest Advanced Vector Extensions (AVX-512). These new instruction sets are unique to Xeon Scalable and improve performance across different workloads. These new instances with network bandwidth ranging from 25 Gbps to 100 Gbps are designed for AI/ML, HPC, analytics and Big Data workloads, that can benefit from the improved packet rate performance.
Intel and AWS engineers collaborate on difficult technical questions every day, and it’s really rewarding when their work pays off with cloud solutions that help our end-users solve their biggest business challenges.
To learn more about these instances and all of Amazon’s Elastic Compute offerings visit https://aws.amazon.com/ec2/instance-types/. Visit Intel.com/CSP to learn more about Intel’s solutions for Cloud Service Providers.