AWS AmazonElastiCache: Document distributed KV caching feature in Valkey 9.1
Summary
Adds a feature bullet describing distributed KV caching on cache.i8ge.48xlarge node clusters with Valkey 9.1+, storing LLM inference KV cache over TCP or RDMA over EFA.
Security assessment
This is a pure feature announcement describing new functionality and data transport options; it makes no security claims or configuration guidance.
Evidence
+ * **Distributed KV caching** – Node-based clusters of the `cache.i8ge.48xlarge` node type, with Valkey 9.1 or later, store the key-value (KV) cache of LLM inference engines with the large object data type, over TCP or RDMA over Elastic Fabric Adapter (EFA). For more information, see [Distributed KV caching](./distributed-kv-caching.html).
Diff
diff --git a/AmazonElastiCache/latest/dg/engine-versions.md b/AmazonElastiCache/latest/dg/engine-versions.md index 635c31f..86ff341 100644 --- a//AmazonElastiCache/latest/dg/engine-versions.md +++ b//AmazonElastiCache/latest/dg/engine-versions.md @@ -22,0 +23,2 @@ Here are some of the new features introduced in Valkey 9.1 (compared to ElastiCa + * **Distributed KV caching** – Node-based clusters of the `cache.i8ge.48xlarge` node type, with Valkey 9.1 or later, store the key-value (KV) cache of LLM inference engines with the large object data type, over TCP or RDMA over Elastic Fabric Adapter (EFA). For more information, see [Distributed KV caching](./distributed-kv-caching.html). +