The product, which comes from the company's Amazon Web Services (AWS) group, aims to improve the performance of web applications by allowing the user to more quickly retrieve information by relying on an in-memory caching system, rather than slower disk-based databases.
The service itself helps with the management, monitoring and operation of in-memory cache environments.
Amazon says the service is best for read-heavy application workloads such as social networking, gaming, media sharing and Q&A portals. It's also good for compute-intensive workloads, such as a recommendation engine.
A few points:
- It's protocol-compliant with any existing Memcached environments.
- Nodes have pre-configured parameters and settings; additional control can be had using Cache Parameter Groups.
- Automatically replaces Cache Nodes in the event of network partitioning, host hardware or software failure. Replaced Cache Nodes carry the same DNS name as the failed Cache Nodes, avoiding the need for client-side refresh of end-point lists.
- Includes Amazon CloudWatch metrics for Memcached deployments, e.g. memory/compute capacity utilization, cache hits, cache misses and number of cache connections.
- Includes automatic software patching.
- You can scale memory resources within the AWS Management Console or with an API call.
Pricing is based on the size of the Cache Nodes used; Amazon starts at $0.095 per hour for a "small" cache (1.3 GB memory, 1 ECU) and scales from there ($0.38 per hour for a "large"; $0.76 per hour for "extra large").
It will be available first in the U.S. East (Virginia) region of AWS and will arrive in other regions "in the coming months."