A cluster network is a group of high performance computing (HPC), GPU, or optimized bare metal instances that are connected with an ultra low-latency remote direct memory access (RDMA) network. Cluster networks with instance pools use instance pools to manage groups of identical instances.

Use cluster networks with instance pools when you want predictable capacity for a specific number of identical instances that are managed as a group.

If you want to manage instances in the RDMA network independently of each other or use different types of instances in the network group, use compute clusters instead. For details, see ComputeCluster.