One engine across every machine you run it on.
A cluster is several machines running Semurg as one. Your data and the work on it are spread across every disk and every core of every node, so adding a machine adds its disks, cores and memory to everything the engine does. Start with one node; add nodes as the data grows.
What a cluster gives you
Room across machines
A store is laid out over the disks of every node, not kept on one machine with the others as spares.
Every machine at work
A write, a read, a graph walk or a scan runs on every node's disks and cores at once. The performance page measures exactly that, on a two-node cluster, at every release.
The same doors
Your applications and agents use the same HTTP, MCP, WebSocket and binary doors as on a single node; nothing in your code changes when the cluster grows.
One node, then more nodes
- The first node is free, forever, for real work as well as for trying it, and it needs no licence key: prove Semurg on your own data before you talk to anyone.
- A key exists only to join a further node to a cluster. Talk to us about the cluster you need: we size it with you and issue the key for each node directly.
- A node is a node: its licence does not change with its cores, memory or disks. The pricing page has the price per node.
- Or rent it. Semurg Cloud runs nodes on servers we provision for you, and a further server joins your cluster the same way.
Measured as a cluster
When we measure the engine, we record what every machine was doing while it answered, and publish that beside the number rather than instead of it.
Every disk
Each disk's throughput, against what that disk can physically do, so a number from an idle disk is visible as one.
Every core and thread
How many were actually working, judged over the whole run rather than at its busiest instant.
Every node
The same figures per node and for the cluster together, so one busy machine beside an idle one cannot read as a full one.
The published figures come from one cluster of two nodes with 16 CPU cores, 32 threads and four NVMe disks between them; the performance page names the release and the machine beside every figure. A number published without that record is not a result we would ask you to believe.
When a machine is lost
Running more than one node is meant to protect your data when a whole machine, a disk or a site is lost. Those protections are being tested now. The durability page lists each test across nodes and its result, and this page will say what a cluster survives only once those tests pass on a published release.
Until then, plan for each node as you would for one copy: a node keeps what it acknowledges on its own disk, and a stopped or killed Semurg process loses nothing it acknowledged (tested with kill -9 on published releases). Semurg runs on, and recommends, datacenter NVMe with power-loss protection.