Clustron Zaris is a distributed, in-memory store for .NET with drop-in ASP.NET Core caching. The same client runs in-process for dev and clustered for production — partitioned and replicated (RF2), highly available, and self-hosted.
Zaris implements the interface ASP.NET Core already caches through — register it in two lines of Program.cs and your caching code does not change, only what is behind it.
An in-process L1 over the distributed L2, with tag invalidation and stampede protection — the speed of local memory with the consistency of a shared store.
Beyond the cache interfaces, an async .NET client with the full store — the same API in-process and against a cluster.
Data is sharded across nodes and replicated at RF2 — a primary and a replica on different nodes — so capacity grows as you add nodes and a node loss is survivable.
Kill a node mid-traffic and its replicas promote instantly — clients never stop, and no acked writes are lost.
Add nodes while the cluster serves — Zaris regenerates the partition map and migrates data automatically, with no downtime.
Hashes, lists, sets, and sorted sets as server-side types, plus per-key time-to-live so cached entries age out on their own.
Cluster-wide publish/subscribe and a durable, replicated event log with consumer groups — coordination that usually needs a second system.
Distributed locks, transactions, watches, atomic counters and leases — use one store for cached data and the logic around it.
Run the store in-process — no separate server, no infrastructure — ideal for local dev and tests.
Point the same client code at a real remote cluster — the move is a configuration change, not a rewrite.
A platform-agnostic connection string reaches every deployment — in-process, Docker, Kubernetes, or Windows — with automatic failover.
Install Zaris, add two lines to Program.cs, and cache the way ASP.NET Core already does.