Amazon Web Services (AWS) has announced a new Amazon Elastic Compute Cloud (Amazon EC2) instance family designed for memory-intensive applications such as in-memory analytics, databases, caching, and scientific computing.

The so-called High Memory Cluster instances are based on two Intel Xeon E5-2670 processors, two 120 GB solid state drives of instance storage, high bandwidth networking, and 244 GiB of RAM.

Amazon claims that, for distributed memory-intensive applications, High Memory Cluster instances are the most cost-effective Amazon EC2 instance and provide a lower cost per GiB of memory than all other instances. Prices are $3.500 per hour for Linux/UNIX, and $3.831 per hour for Windows usage.

“Memory-intensive workloads such as real-time applications used by healthcare providers, social networking companies and advertising technology providers require large amounts of memory to maintain high-performance,” said Peter DeSantis, Vice President of Compute Services, AWS.

“We designed the High Memory Cluster instances specifically for these workloads, and this is the third instance family (along with High Storage instances and High I/O instances) launched in the past six months designed specifically for high performance applications.”

Last year, SAP made its HANA in-memory database available on Amazon Web Services at an hourly rate. Sapan Panigrahi, vice president of HANA Cloud at SAP said that the new High Memory Cluster instances would allow even larger workloads to run on HANA in the AWS cloud .

“By leveraging the strength of SAP HANA in-memory computing, with the agility and low cost of the AWS cloud, a new generation of solutions are now possible,” said Panigrahi.

Customers can launch Linux and Windows High Memory Cluster instances using the AWS Management Console, the AWS and Amazon EC2 Command Line Interfaces, AWS SDKs, and third-party libraries.

High-memory cluster instances are currently available in the US East (North Virginia) region, and will be made available in other AWS regions in the coming months.