Engineering at the filesystem and kernel level
Our engineers work alongside your team on the code that matters most: kernel modules, storage engines, filesystem features and the tools to run them in production.
What we offer
Embedded team
If you're building a filesystem or storage product, our engineers join your team and work with you from planning to release.
Systems and kernel programming
Kernel modules and storage engines that stay fast and correct across new Linux kernel versions.
Custom eBPF tools
Tools designed for your system, so your team can keep finding new insights.
Systems we work on at source-code level
- LustreDev · install · tuning
- CephDeploy & tune
- pNFS / NFSDeploy & tune
- ZFSBackend storage
- RaftHA control planes
- K8s CSIBlock volumes
- io_uringStorage engines
- eBPFLive debugging
- LMCacheKV cache for LLMs
What we've built
Distributed block storage, from NVMe to Kubernetes
Our engineers have worked on every layer of a distributed block platform, from the Raft-replicated metadata control plane to native Linux block devices and Kubernetes volumes.
- Kubernetes
- Go CSI driver with the full gRPC surface. The block-serving process runs as a host-level systemd unit, so it survives CSI pod restarts without losing in-flight I/O.
- Kernel
- Linux ublk integration exposes distributed storage as native block devices, shipped as a DKMS driver.
- Control plane
- Raft-replicated, RocksDB-backed metadata service with leader election, leases and a stateless proxy tier.
- Data protection
- Chunk-level replication and erasure coding, background GC and compaction, S3-backed cold tiering.
- Storage engine
- Single-threaded io_uring reactor with a custom coroutine model and a hand-written assembly context switch.
- Devices
- Raw NVMe admin commands issued directly through io_uring passthrough.
Formal software modelling with TLA+
Protocol-critical logic is modelled and model-checked with TLC, with internal guides for translating C into formal models.
Deterministic system tests
An Ansible framework stands up multi-node control-plane clusters and runs hundreds of parametrised failure scenarios, with fault injection built into the storage engine itself.
Integrity under failure
Kill a storage node mid-write, let chunk recovery run, re-validate every block checksum. Checked automatically in CI.
Storage slower than your compute?
Send us your topology, workload and a benchmark run. We will get back to you.