Posted On: Nov 23, 2021
Now you can use knowledge tiering for Amazon ElastiCache for Redis as a decrease value option to scale your clusters to as much as a whole bunch of terabytes of capability. Information tiering gives a brand new price-performance choice for Redis workloads by using lower-cost strong state drives (SSDs) in every cluster node along with storing knowledge in reminiscence. It’s excellent for workloads that entry as much as 20% of their total dataset repeatedly, and for functions that may tolerate further latency when accessing knowledge on SSD.
When utilizing clusters with knowledge tiering, ElastiCache is designed to robotically and transparently transfer the least lately used objects from reminiscence to regionally hooked up NVMe SSDs when accessible reminiscence capability is totally consumed. When an merchandise that strikes to SSD is subsequently accessed, ElastiCache strikes it again to reminiscence asynchronously earlier than serving the request. Assuming 500-byte String values, you possibly can count on a further 300µs latency on common for requests to knowledge saved on SSD in comparison with requests to knowledge in reminiscence.
ElastiCache knowledge tiering is out there when utilizing Redis model 6.2 and above on Graviton2-based R6gd nodes. R6gd nodes have almost 5x extra whole capability (reminiscence + SSD) and might help you obtain over 60% financial savings when working at most utilization in comparison with R6g nodes (reminiscence solely).
To get began utilizing ElastiCache knowledge tiering, create a brand new cluster utilizing one of many R6gd node sorts utilizing the AWS Administration Console for ElastiCache, the AWS CLI, or one of many SDKs. Information tiering on R6gd nodes is out there within the Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Tokyo), Europe (Frankfurt), Europe (Eire), US East (N. Virginia), US East (Ohio), US West (N. California), and US West (Oregon) Areas. For pricing, see Amazon ElastiCache pricing and for extra info, see the ElastiCache knowledge tiering documentation.