Parallel & Distributed Storage Deployment
We deploy, integrate and optimise Lustre, Ceph and pNFS on ZFS or ldiskfs, matching the design and tuning to your workloads and hardware, from first deployment through to production.
Talk to usWhat we offer
New deployment
We design the architecture, then install the MDS/OSS servers and set up HA failover and LNet networks, on ZFS or ldiskfs.
Expansion or migration
Expand or migrate an existing system. You get it installed and benchmarked, with a runbook.
Deployment automation
We automate MDS/OSS installs, HA failover and LNet setup, so every node, and every rebuild, comes up exactly the same.
Tuning and benchmarking
We tune striping, RPC, cache and network settings for your workload and prove the gains with obdfilter-survey, IOR and IO500, with a before and after report.
Monitoring and handover
We hand over Grafana dashboards, a runbook and a tuning report, so your team can run the system.
Lustre support
We stay on after go-live to diagnose problems and re-tune the system as your workloads and capacity change.
How a deployment runs
Assess
We turn your workloads, hardware and capacity targets into an architecture.
Deploy
MDS/OSS install, HA failover, LNet networks and routing.
Tune
Striping, RPC, cache and network parameters set for your workload.
Benchmark
Raw, network and filesystem-level benchmarks, through to a full IO500 run.
Hand over
We hand over Grafana dashboards, a runbook and a tuning report.
Other solutions
eBPF observability
Find the bottleneck in live systems without redeploying, and get custom eBPF tools built for your stack.
Storage for AI
Keep GPUs busy with KV-cache offload, GPUDirect Storage, and data and checkpoint I/O tuned for your cluster.
Storage slower than your compute?
Send us your topology, workload and a benchmark run. We will get back to you.