AWS on Compute Power Trip: Adding 5 New EC2 Types Amazon Web Services (AWS), added five new instance types this week to its Elastic Compute Cloud. (EC2), along with other enhancements that will increase its flexibility in its compute offerings. These announcements were made Tuesday at the company’s 2020 reInvent conference. This year’s event was held online for three weeks due to COVID-19 restrictions. The five new EC2 instances, each offering different compute, storage and cost benefits, add to the already extensive range of AWS compute options. The company announced that AWS already offers more compute instances types than any other cloud provider. Its instances are based on the fastest processors of Intel, cost-optimized instances using AMD processors, NVIDIA’s most powerful GPU instances, instances that offer up to 400 Gbps network performance, and the only Arm-based instances available in the cloud offering 40% better price performance thanks to AWS-designed AWS Graviton2 CPUs. These are the new EC2 instances:

• C6gn is powered by Arm-based AWS graviton2 processor. It is designed for network-intensive workloads like data modeling and high-performance computing. However, it has better packet processing capabilities than other instances. According to AWS, C6gn allows users to “consolidate their workloads onto smaller instances or smaller instance size, and reduce infrastructure cost.”
• G4adm, powered with AMD GPUs: These instances are designed for graphics-intensive applications and can render high-quality images for virtual worksstations, game streaming, and other jobs. However, they offer a “45% better price performance than NVIDIA-powered instances (namely G4dn).
• M5zn powered by Intel Xeon Scalable CPUs: Designed for high compute performance and high network throughput but with less memory. This allows users to avoid spending more computing resources than they actually need to perform complex calculations and analyze real-time data for their financial, analytical, and gaming workloads.
• D3/D3en, powered with Intel Xeon Scalable CPUs: Storage-intensive workloads that require high network speeds. These instances offer better storage and network performance than the D2 instance.
• R5b, powered with Intel Xeon Scalable CPUs: Designed to handle large relational database workloads that use Elastic Block Store storage (EBS), R5b promises a triple “instance-to EBS” performance improvement over the current R5 instance. AWS stated that this will significantly increase performance for large database workloads that process large amounts of data in memory.

AWS also announced other compute news. The company said it is working on smaller versions of its AWS Outposts product, which will be released sometime in 2021. AWS Outposts allows organizations to run the full AWS cloud from their premises using server racks that AWS delivers and installs at the location they choose. AWS Outposts currently comes in a 42-rack unit measuring 80 inches high, 24 inches wide, and 48 inches deep. AWS is currently developing two sizes of server racks: a 1.75-inch single-rack server and a 3.5 inch two-rack server. These new forms consume significantly less power and bandwidth than the existing AWS Outposts product. AWS is expanding its “Local Zones” footprint. Although they are not fully-fledged AWS regions, local zones can extend AWS’ capabilities to densely populated areas like Los Angeles. They are designed for applications that are sensitive to latency and throughput and require close proximity to end users than the existing AWS regions. AWS launched Local Zones in Houston and Boston, along with another dozen opening next year at re:Invent.