The best storage isn't the one your vendor sells you. It's the one that fits the workload.
The real question was never "which storage is best." It's which storage for which workload — and whether you can prove that choice on your own data before you tie up capital in a disk array. We design storage architecture where every workload runs on the technology that fits its I/O profile, and that choice is verified on your real workload — not a vendor benchmark.
Two questions run through everything here.
Whether every workload — message queue, database, DMS / archive, Kubernetes volume — runs on a storage tier that matches its I/O profile, latency and durability needs. Backed by a Suitability Matrix, not a vendor datasheet.
Whether the choice is validated by a proof-of-concept on your real workload with defined success criteria (FIO) — including the negative result: what not to use, and why. Not a vendor benchmark.
Every engagement starts by measuring both gaps and ends by closing them — a validated technology choice and a deployed, tiered storage layer at handover. Don't sink capital into storage that doesn't fit the workload.
Storage done right is three moves: map it, prove it, build it.
Map it
Workload profiling, technology evaluation, a vendor-neutral Suitability Matrix, capacity and growth planning, and a hardware BOM — the right technology matched to each workload before anything is bought.
Prove it
A proof-of-concept on your real workload with FIO success criteria — including the negative PoC. Plus storage performance engineering (LVM vs raw, message-queue I/O profiling) for the load you actually expect.
Build it
Ceph cluster design, tiered storage with erasure-coding optimization, enterprise tiering and DMS migration, GPFS / CNSA on OpenShift, and deployment. Handover is validated storage — the workload verified against its I/O criteria.
What we deliver
Four ways to engage
The same lifecycle as everywhere at G2F — assessment, build, operations, training — focused on one question: which storage for which workload, and can you prove it on your own data?
01
AssessStorage Platform Assessment & Technology Evaluation
The vendor will tell you which array to buy. We tell you which technology fits which workload — and what to prove before you commit capital.
- Workload profiling (I/O per message queue / database / DMS-archive / Kubernetes volume) and a Storage Suitability Matrix — what fits what, from real operations, not a vendor benchmark.
- A Fit + Prove gap, a hardware assessment and BOM, and capacity & growth planning with configuration variants and migration paths.
- Vendor-neutral, fixed scope.
02
BuildFrom design to running storage
A disk array full of capacity isn't an architecture. We build the layer where each workload sits on the tier that fits it — and it's proven before it goes live.
- Ceph cluster design (greenfield to enterprise, CRUSH, data-integrity layer), tiered storage and erasure-coding optimization.
- Enterprise tiering with DMS / FileNet migration, GPFS / CNSA integration on OpenShift, and message-queue / latency storage strategy.
- Ceph and LINSTOR deployment. Handover is validated storage — or, for design-only, an implementable BOM.
03
OperateStorage Advisory & Support
Storage doesn't stop at go-live — capacity grows, workloads shift, and the integration edges (Kubernetes, networking) are where things break.
- On-demand Ceph and storage consulting, and post-deployment integration troubleshooting (MTU / routing / CSI / PVC).
- Capacity and growth monitoring, and ongoing advisory from a single assessment through to multiple workstreams.
- Retainer or on-demand blocks. Database high-availability lives in Resilience & DR.
04
TrainCeph Hands-on & Storage Enablement
The goal is that your team can run and extend the storage platform without us in the room.
- A Ceph hands-on course — a dedicated environment per participant, a final test, a certificate, and post-course instructor support.
- Storage architecture enablement: design principles, tiering, erasure coding, capacity planning.
- Format transferable to Kubernetes and MaaS — usually alongside a build or assessment.
Vendor-neutral by design — open-source-first
We don't sell you a disk array, and we don't resell anyone's. We assess what your workloads actually need and design the storage layer on an open-source-first stack: Ceph for distributed storage from greenfield to enterprise, LINSTOR for Kubernetes persistent volumes. Where you've already invested in IBM, we deploy GPFS / IBM Storage Scale — a proprietary, licensed platform, used where it earns its place, not as a default. The Suitability Matrix stays vendor-agnostic; we evaluate SAN, Mayastor and LVM in the as-is, but we don't sell you a proprietary array. The proof is a PoC on your real workload — including the honest negative result: what not to buy, and why.
Delivered in production
Where we've done this
Storage we've assessed, designed, proven and deployed — for a regulated bank, a satellite-data provider, an enterprise integrator and Kubernetes platforms.
A five-technology Suitability Matrix, a two-phase PoC (including a valid negative result), GPFS / CNSA on OpenShift, and enterprise Ceph tiering across a three-datacentre topology.
A greenfield Ceph design as a service — control and data plane, tiered SSD + HDD, erasure-coded pools — with a hardware BOM that saved thousands of euro against the originally considered configuration.
Hands-on Ceph deployment on bare metal, on-demand storage consulting and post-deployment troubleshooting — and a Ceph training course with a test, certificate and pricing model.
Distributed LINSTOR storage for Kubernetes persistent volumes, with NFS integration.
Which storage, for which workload? That's the first thing we prove.
A Storage Platform Assessment maps your workloads to the technology that fits them, shows you the gaps, and gives you a hardware BOM and capacity plan — before you commit capital. Vendor-neutral, fixed scope. Most first conversations take 30 minutes. No pitch, no deck.