New AWS EC2 Instances Target Data-Intensive Cloud Workloads

Amazon Web Services (AWS) is expanding its cloud offerings to enterprise customers with a new Bare Metal cloud instance for the Amazon Elastic Compute Cloud (EC2) and a new H1 Storage Optimized instance designed for high-demand computing workloads.

The new AWS EC2 instances were unveiled at the company's AWS re:Invent conference in Las Vegas, which runs through Dec. 1.

The Amazon EC2 Bare Metal instances, which are available in preview form immediately, are the first AWS EC2 instances that will allow customer applications to run directly on the underlying hardware while still providing access to all the elasticity, scalability, and security of the cloud, according to Amazon. The Bare Metal cloud services, which are part of the company's i3 instance family, are designed for non-virtualized workloads and give applications direct access to Intel Xeon E5-2686 v4 processors, 512 GiB (gibibytes) of memory, 36 hyperthreaded cores and 15.2TB of local, SSD-based NVME storage. The i3 Bare Metal instances include the flexibility and capabilities of AWS EC2 instances, including support for attaching Amazon Elastic Block Store (Amazon EBS) volumes, the ability to use Amazon Virtual Private Cloud (Amazon VPC) and security group settings, and the ability to use Elastic IP addresses or Elastic Load Balancers.

The Bare Metal instances, which are built on top of the AWS Nitro System for high performance and high availability through hardware offload cards, offer secure protection of the hardware and system software with custom AWS Nitro security processors that are built onto the motherboard. Applications can be deployed on Bare Metal instances in minutes and can be scaled up and down in seconds, like other AWS instances.

The new Storage Optimized H1 instances are designed for data-intensive workloads such as MapReduce, distributed file systems, network file systems, log or data processing, and big data clusters, according to Amazon. In addition, AWS also launched the next generation of General Purpose instances, M5, which have up to 50 percent more vCPUs, 50 percent more memory, and 25 percent more network bandwidth than previous generation M4 instances. The H1 instances are available immediately to customers.

The H1 instances are powered by 2.3 GHz Intel Xeon E5 2686 v4 (Broadwell) processors and provide up to 64 vCPUs and 256 GiB of DRAM. The H1 instances also use up to 16TB of magnetic storage and 25Gbps of network bandwidth per instance, making them suitable for processing very large data sets.

"AWS continues to expand and enhance what was already the cloud’s broadest and most capable compute service," Matt Garman, vice president of AWS Compute Services, said in a statement. "Most of our customers have diverse computing needs, and they've told us having the right instance for the right workload really matters."

Also unveiled at AWS re:Invent conference is AWS PrivateLink, a new managed service which allows customers to access third-part SaaS applications from their Virtual Private Cloud without exposing it to the public internet, and Amazon GuardDuty, a managed intelligent threat detection service which helps customers protect their AWS accounts and workloads by continuously monitoring account activity for malicious or unauthorized behavior.