AWS re:Invent: New Amazon EC2 Instances, Private 5G Services, and More

At AWS re:Invent 2021 in Las Vegas, AWS has announced the release of three new Amazon EC2 C7g instances powered by new AWS Graviton3 processors. Next to that, AWS has introduced its Private 5G services, as well as a Mainframe Modernization service, to name a few of the most eye-catching announcements at this AWS event.

New C7g instances powered by this new processor, Graviton3, would provide up to 25% better performance for compute-intensive workloads over current generation C6g instances powered by AWS Graviton2 processors. It would enable AWS clients to not only significantly improve the performance, but also the cost and energy efficiency of their workloads running on Amazon Elastic Compute Cloud (Amazon EC2).

C7g instances are the first in the cloud to use DDR5 memory, which delivers 50% more memory bandwidth than AWS Graviton2-based instances, allowing memory-intensive workloads like scientific computing to run faster. In comparison to AWS Graviton2-based instances, C7g instances provide 20% more networking bandwidth. Elastic Fabric Adapter (EFA) is a feature of C7g instances that allows programs to interact directly with network interface cards, resulting in reduced and more consistent latency for applications that require parallel processing at scale, such as HPC and video encoding. C7g instances are now accessible in preview.

AWS Trn1 Instances

In Amazon EC2, new Trn1 instances powered by AWS Trainium processors would offer a fitting pricing and training time for most machine learning models. For I/O-intensive applications running on Amazon EC2, new storage-optimized Im4gn/Is4gen/I4i instances based on AWS Nitro SSDs (solid-state drives) would offer the highest storage performance. These instances mark the introduction of new Amazon EC2 instances based on AWS-designed processors, which will enable clients to power their most mission-critical applications.

When compared to the newest P4d instances, AWS’s Trn1 instances powered by AWS Trainium processors would offer the greatest pricing performance and the fastest machine learning model training on Amazon EC2. According to AWS, it allows for savings up to 40% on deep learning model training. Trn1 instances have an EFA networking bandwidth of 800 Gbps (2x that of the newest EC2 GPU-based instances) and interact with Amazon FSx for Lustre high-performance storage, allowing clients to start Trn1 instances with EC2 UltraClusters support. Developers may use EC2 UltraClusters to scale machine learning training to 10,000+ Trainium accelerators coupled by petabit-scale networking, offering clients on-demand access to supercomputing-class performance to reduce training time from months to days for even the most difficult models.

AWS Nitro SSDs

Photo David Brown, Vice President, Amazon EC2 at AWS
“With our investments in AWS-designed chips, customers have realized huge price performance benefits for some of today’s most business-critical workloads,” said David Brown, Vice President, Amazon EC2 at AWS.

Im4gn/Is4gen/I4i instances featuring new AWS Nitro SSDs would deliver the highest storage performance for I/O intensive-workloads. Im4gn/Is4gen/I4i instances provide up to 30 TB of NVMe storage using AWS-designed AWS Nitro SSDs, with up to 60% lower I/O latency and 75% reduced latency variability compared to prior generation I3 instances, allowing for better application performance. Through improvements in the storage stack, hypervisor, and hardware, AWS Nitro SSDs are intimately linked with the AWS Nitro System. Customers benefit from increased functionality since AWS manages both the hardware and firmware of the AWS Nitro SSDs. SSD upgrades are supplied more quickly than with commercial SSDs.

Im4gn instances use AWS Graviton2 processors. According to AWS, they offer up to 40% better pricing performance and a 44% cheaper cost per TB of storage than I3 instances. Is4gen instances also employ AWS Graviton2 processors. In comparison to I3en instances, according to AWS, they offer up to 15% cheaper cost per TB of storage and up to 48% greater compute performance.

“With our investments in AWS-designed chips, customers have realized huge price performance benefits for some of today’s most business-critical workloads. These customers have asked us to continue pushing the envelope with each new EC2 instance generation,” said David Brown, Vice President, Amazon EC2 at AWS. “AWS’s continued innovation means customers are now getting brand new, game changing instances to run their most important workloads with significantly better price performance than anywhere else.”

Introducing AWS Private 5G

AWS Private 5G is a new managed service that enables businesses to set up and grow private 5G mobile networks in their buildings in days rather than months. Clients may specify where they want to build a mobile network and the network capacity required for their devices with just a few clicks in the AWS console. AWS will then deliver and maintain the small cell radio units, servers, 5G core and radio access network (RAN) software, and subscriber identity modules (SIM cards) needed to set up a private 5G network and connect devices.

AWS Private 5G would simplify network configuration and deployment, while expanding capacity on demand to serve more devices and network traffic. AWS Private 5G has no setup fees or per-device expenses. Clients would simply pay for the network capacity and throughput they need.

AWS Private 5G interacts with AWS Identity and Access Management (IAM), giving network managers complete control over which resources mobile devices on their private mobile networks have access to. Clients may use AWS Private 5G to establish small networks with fewer devices, assess network demands once up and running, and grow their private mobile network as they add more devices by using AWS’s elasticity and pay-as-you-go pricing. Clients can easily establish and operate private 5G mobile networks with AWS Private 5G without having to worry about purchasing, scaling, or maintaining equipment, or about needing to add extra capacity when adding more devices.

“Many of our customers want to leverage the power of 5G to establish their own private networks on premises, but they tell us that the current approaches make it time-consuming, difficult, and expensive to set up and deploy private networks,” added David Brown. “With AWS Private 5G, we’re extending hybrid infrastructure to customers’ 5G networks to make it simple, quick, and inexpensive to set up a private 5G network. Customers can start small and scale on-demand, pay as they go, and monitor and manage their network from the AWS console.”

AWS Mainframe Modernization

At AWS re:Invent 2021 in Las Vegas, AWS has also introduced its Mainframe Modernization service. It’s a new service that enables clients to transfer mainframe and legacy workloads to the cloud more quickly and easily, allowing them to benefit from AWS’ “agility, elasticity, and cost savings.” Clients may rework their mainframe workloads to operate on AWS using AWS Mainframe Modernization, which transforms ancient programs into contemporary Java-based cloud services. Customers may also leave their applications as-is and re-platform their workloads to AWS with little changes by reusing existing code.

AWS Mainframe Modernization includes a runtime environment that offers the compute, memory, and storage needed to execute both refactored and re-platformed programs while also handling capacity provisioning, security, load balancing, auto-scaling, and application health monitoring. AWS Mainframe Modernization includes all of the essential development, testing, and deployment tools to automate the transfer from mainframe and legacy environments to AWS.

AWS Mainframe Modernization may be used by MSPs and systems integrators (SIs) to assist corporate migration teams in assessing and analyzing migration readiness and planning migration initiatives. AWS Mainframe Modernization has no upfront expenses, and clients just pay for the computing resources they use.

“More and more customers are modernizing their applications by moving them to the cloud to take advantage of the elasticity and agility of AWS and to reinvent customer experiences at a lower cost. But, for most enterprises, migrating their mainframe applications to the cloud is uncharted territory,” said William Platt, General Manager of Migration Services at AWS. “With today’s launch of AWS Mainframe Modernization, customers and systems integrators can now more quickly modernize their legacy mainframe workloads in a predictable way and get rid of much of the complexity and manual work involved in migrations. AWS Mainframe Modernization provides automation in discovery, analysis, migration, and testing phases, and it gives customers and systems integrators a managed runtime environment to deploy their re-platformed and refactored mainframe workloads for their customers, which makes modernizing applications faster and easier.”


Im4gn instances Is4gen instances are already available, while C7g instances and Trn1 instances are available now in preview.

AWS Mainframe Modernization is available in preview today in US East (N. Virginia), US West (Oregon), Asia Pacific (Sydney), EU (Frankfurt), and South America (São Paulo) with availability in additional AWS Regions in the coming months.