The challenge.
A research environment depends on more than its compute nodes. Users need a way in, storage and interconnections need to work, and the physical infrastructure must support the complete environment.
Why Neurulus fit.
A research cluster is useful when people can reach it and its compute, storage and network work together. Neurulus fit because those requirements could be deployed and commissioned within one platform environment.
How the platform addressed it.
Neurulus deployed the research-computing platform through Uday Network Technologies, joining two HPC nodes, a GPU node and an access/login node. The delivery included 32 TB of SSD storage and interconnection switching.
Four workstations, converged switching, a managed rack, power backup and software supported the user and operating environment. Supply and commissioning brought the cluster and the infrastructure around it into the same deployment.
In the deployment
- Two HPC server nodes, one GPU node and one access/login node.
- 32 TB SSD storage with interconnection switching.
- Four workstations with UPS equipment.
- Converged switching, a managed rack, power backup and software.
- Supply and commissioning within the engagement scope.
The outcome.
The institution gained an integrated research environment with compute, user access, storage and supporting infrastructure brought together. The platform addressed how researchers reached the resources as well as the resources themselves.
Multiple computing resources
HPC and GPU nodes supplied distinct resources within the research environment.
A path from users to compute
An access/login node and workstations supported use of the cluster.
Infrastructure commissioned together
Storage, networking, rack and power support accompanied the compute deployment.