Technology Services
Compute, storage and facility operations for workloads that stay put.
The problem
Every cloud programme leaves a residue: the workloads that cannot move for licensing, latency, residency or simple economics. They tend to get less attention every year, which is exactly backwards, they are usually the ones that cannot fail.
The residue is not a rounding error. In most organisations that have run a cloud programme, what remains on premises is the set of systems with the least tolerance for downtime: the manufacturing line controller, the clinical system, the trading application, the thing with a licence tied to a physical socket. They stayed because moving them was hard, and hard usually means important.
What happens next is a slow withdrawal of attention. The talented people follow the interesting work, the refresh budget follows the strategy, and the estate gets one more year of support every year. Nothing fails immediately. What degrades is the margin for error: support contracts lapse to next-business-day on systems that need four hours, spare capacity gets consumed, and the person who knew the storage layout retires.
Backup is where this concentrates. Backup jobs report success, and reporting success is not the same as being able to restore. The gap between the two is only ever discovered at the worst moment, and it is usually not the backup that failed, it is that nobody had tested the restore path, the recovery documentation referenced a system that has changed, or the recovery point turned out to be twelve hours older than anyone believed.
Capacity planning stops being connected to the application roadmap. Storage gets bought when it runs out rather than when the roadmap said it would, which means it gets bought at list price under time pressure. Hardware lead times then turn a capacity problem into a project delay, and the lead time is discovered rather than planned for.
The economics deserve honesty in both directions. Some of this estate genuinely should move and has not, because nobody has done the arithmetic since the first migration. And some of it will be cheaper on premises for its entire remaining life, which is a legitimate answer that a supplier whose margin is in cloud services has no incentive to give.
Recognise any of these
If three or more of these are true, this page is about your estate.
What we build
Named platforms, not categories. If you run one of these, this is the conversation.
Sizing comes from measured utilisation over a representative period, not from the specification of what is being replaced. The like-for-like refresh is the most expensive habit in infrastructure: it carries forward a decision made against a workload that has since changed, usually in the direction of needing less compute and more storage throughput.
The test is a restore into an isolated environment, with the recovery time and recovery point measured rather than asserted. This is the first thing the assessment does, because it is the finding most likely to change what the client wants to spend money on. A green tick in a backup console is evidence that a job ran, and nothing more.
Capacity gets planned from the application roadmap for the next two years, so purchases happen on a forecast rather than at exhaustion. The saving is not just the price difference between planned and urgent procurement, it is the projects that do not slip because the capacity was already there.
Where the room is owned rather than leased, power, cooling and physical access get reviewed alongside the equipment. This is where the surprises live: a refresh that draws more per rack than the distribution supports, or a cooling design that was adequate at half the current density.
The link between what stayed and what moved gets designed rather than inherited, with the latency and bandwidth requirements of the specific integrations that cross it. Most hybrid performance complaints trace back to this boundary being an accident of migration sequencing.
How the engagement runs
Every workload, dependency, contract and kilowatt, in one inventory.
VSI delivers
Discovery from flow data and facilities records, the inventory later savings are measured against.
Move groups sequenced by dependency, with rollback per group.
VSI delivers
The migration runbook, rehearsed, including the failback nobody hopes to use.
Cutovers land in agreed windows with the business watching.
VSI delivers
Execution with per-group verification before the source is touched.
Hardware retired, contracts closed, space handed back.
VSI delivers
Sanitisation documented, support contracts terminated on schedule, the stage where the saving becomes real.
How it deploys
And what we need from you at each step. A timeline with no client obligations in it is a timeline that slips.
Utilisation and capacity assessment against the roadmap.
Refresh design, procurement and lead-time plan.
Install, migrate, test a restore, decommission.
How this goes wrong
Published because it is only writable by somebody who has had the failure. Each of these has happened on this kind of work, and each has a specific thing that prevents it.
Return on investment
Every lever names the mechanism and how it is measured against your own baseline, captured before the work starts. That is how the return stays a number your finance team can audit rather than a promise on a slide.
Racks, power and support contracts for equipment that virtualises onto a fraction of the hardware. The business case is written in your own colocation and maintenance invoices, and measured the same way after the consolidation wave completes.
Ageing hardware burns power twice, at the plug and in the cooling. Refresh economics are modelled from your metered draw where available and manufacturer plate data where not, and the measurement is the utility bill and the PUE trend.
The largest cost in a data-centre move is the outage nobody planned. Dependency-mapped move groups, rollback per group and rehearsed cutovers make the measurement simple: business interruption during the programme, target zero.
Post-consolidation estates keep paying support on hardware that left the building. Contract cleanup is a line-item recovery, measured on the renewals themselves.
Timing
Most of the value in this decision is in when, not whether. Find the row that matches your situation.
| Criterion | Verdict | Why |
|---|---|---|
| Support or maintenance contracts on this estate expire within two quarters | Now | The decision is forced at that date. Making it with measured utilisation beats making it from a renewal quote. |
| Nobody can name the last tested restore | Now | This is the cheapest high-value thing on the page and the one most likely to change your spending priorities. |
| A capacity constraint has delayed a project in the last year | Now | It will happen again, and planned procurement is materially cheaper than urgent procurement. |
| A major application on this estate is being replaced within two years | Soon | Do not refresh capacity for a workload that is leaving. Assess now so the refresh is scoped around the replacement. |
| Recent refresh, tested recovery, capacity headroom against the roadmap | It can wait | You are in good shape. Revisit at the next renewal, and keep testing the restore. |
For government estates the on-premises workloads are frequently the ones with the strongest reason to stay: a defined boundary, a physical control requirement, or a system whose accreditation would have to be redone. That makes the assessment more valuable rather than less, because the question stops being whether to move and becomes how to run a boundary properly for another five years. Media sanitisation and equipment disposal need documented chain of custody, which is a deliverable rather than an assumption. Where an existing accreditation covers the current configuration, we say plainly what a refresh would require re-examining, that is often the deciding factor and it is better known at the start.
Objections
Questions
Occasional updates across every VSI service, new capabilities, new offerings, and what changed. No sales sequence; leaving is one reply.
Used only for these updates, see the privacy statement.
What it costs
Assessment is fixed-price. Hardware is quoted at line-item level through VSI TechSource so you can see what you are paying for.
Book a 20-minute assessment
