Cloud & Data Center
You asked about cloud and got quoted by the company that owns the data center.
Their answer was always going to be their own floor space.
Cloud & Data Center Services for Growing Businesses
We scope against the workloads you actually run — public cloud, private cloud, colocation, disaster recovery — then price it across competing providers with real SLAs. We have no house brand to protect.
Downtime Doesn't Announce Itself
Most outages trace back to something that was avoidable.
Cloud, Colocation and Disaster Recovery for Growing Companies
Most mid-market companies did not choose their current infrastructure so much as accumulate it. A server refresh here, a platform there, a backup product nobody has restored from in two years. It works until a lease expires, a support contract ends, or an auditor asks a question. We help you decide what should move, what should stay, and who should run it — then bring competing options from more than a hundred cloud and data center providers.
- Public & hybrid cloud
- Colocation & data center
- Backup & disaster recovery
- Desktop as a Service
- Direct cloud interconnect
- Migration support
- Server & workload migration planning
- Rack space, power & cross-connects
- Recovery time & recovery point planning
- Microsoft 365 & SaaS backup
- Hardware lease & refresh timing
When Your Server Room Reaches End of Life
Hardware refresh cycles force the question whether or not anyone is ready for it. A five-year-old host cluster comes off support, the UPS batteries need replacing, and suddenly there is a capital request on the table that nobody has an appetite for. That moment is the most common reason companies call us about cloud.
The reflexive answer — move everything to the public cloud — is right about as often as it is wrong. Some workloads get cheaper and more resilient. Others get more expensive and slower. The six factors below are what actually decide it, and they are worth working through before anyone quotes you a migration.
What actually needs to move
Rarely everything. Most environments split cleanly into workloads that belong in public cloud, workloads that are cheaper and simpler left on dedicated hardware, and a handful of legacy applications that will dictate the whole design if you let them.
Where latency really matters
Applications that chat constantly with a database, or that field staff hit over a mobile connection, behave very differently depending on where they sit. This is where regional facilities and direct interconnect earn their cost, and where a cheap distant region quietly does not.
The real cost of leaving it where it is
Not just the hardware quote. Power, cooling, floor space, support renewals, licensing, and the hours your team spends maintaining it. That total is what any cloud or colocation proposal should be compared against, and it is usually higher than the number people carry in their heads.
Backup that is genuinely separate
Backups sitting on the same infrastructure, under the same credentials, are not really backups. Separation of failure domain and separation of identity are what make a restore possible on the day something encrypts or deletes the primary copy.
Who operates it after go-live
Unmanaged infrastructure hands you a bill and a login. Managed offerings cost more and absorb patching, monitoring, and response. Which one is right depends entirely on the team you actually have, not the team the org chart implies.
The exit you cannot see yet
Egress charges, proprietary services, and contract terms determine what leaving costs three years from now. Worth reading before signing rather than after, because it is the part of the agreement that is hardest to renegotiate later.
How We Approach a Cloud or Data Center Decision
Infrastructure spending goes wrong in two familiar directions: postponing a refresh until something fails, or lifting everything into a hyperscaler and discovering the run rate three months later. The process below is built to prevent both.
Workload Review
What you run today, what it depends on, how it performs, and what the current environment genuinely costs once power, licensing, support, and staff time are counted honestly.
Placement & Sizing
Which workloads belong in public cloud, which are better on dedicated or colocated hardware, and which should stay put for now. Sized against real utilization rather than the specs someone bought three refreshes ago.
Competing Proposals
Two to three providers scoped to the same requirement, with committed performance, support response, and migration assistance stated plainly — plus the egress and renewal terms most quotes leave to the fine print.
Migration & Review
Phased cutover so you are not moving production, backup, and identity in the same weekend — and a review once real usage data exists, because the first bill is where sizing assumptions get tested.
What We Source: Cloud, Colocation, Backup and DR
Vendor-neutral across the infrastructure stack, from hyperscaler landing zones to a single cabinet in a regional facility. If a workload does not belong in the cloud, we will say so.
Public & Hybrid Cloud
Landing zones, migration assistance, and ongoing management across the major platforms and the regional providers that often price better for steady-state workloads.
Colocation & Data Center Space
Single cabinets through private cages, with power, cross-connects, and remote hands. Frequently the cheaper answer for hardware you already own and intend to keep running.
Backup & Disaster Recovery
Immutable backup, replication, and recovery-as-a-service — scoped to a recovery time and recovery point you have actually agreed on rather than inherited from a default setting.
Desktop as a Service
Hosted desktops for contractors, seasonal staff, regulated data, or an acquisition you need working on day one without shipping hardware.
Direct Cloud Interconnect
Private, dedicated connections into cloud platforms that bypass the public internet — steadier performance, and often cheaper than paying egress at internet rates.
Managed & Private Cloud
Dedicated environments with an operator behind them, for workloads with compliance, licensing, or performance constraints that make shared public infrastructure awkward.
Why the Cloud Bill Never Matches the Estimate
This is the single most common frustration we hear from companies already in the cloud. The migration was quoted against a tidy monthly figure, the actual invoice arrived meaningfully higher, and it has kept climbing since.
The causes are consistent. Instances sized against old on-premise specs rather than real utilization, and left running around the clock when the workload is used during business hours. Storage tiers that never get reviewed as data ages. Data transfer out of the platform, which is easy to overlook at design time and is frequently the line item that surprises people most. And orphaned resources — volumes, snapshots, and test environments nobody remembered to switch off.
The structural point is that cloud pricing rewards attention and punishes autopilot. On-premise hardware costs the same whether or not you optimize it. Cloud does not, which is an advantage only if somebody owns the review.
If you are already in the cloud and the number keeps moving, bring us a recent bill. We will separate what is sizing, what is architecture, and what is simply a commitment you could be getting a better rate on — and tell you what competing providers would charge for the same workload.
Disaster Recovery That Has Actually Been Tested
Almost every company we speak to has backups. Far fewer have a restore they have proven, and fewer still can say how long a full recovery would take or how much data they would lose getting there. Those two numbers — recovery time and recovery point — are the entire conversation, and they should be decided by the business rather than inherited from whatever the backup software defaults to.
The gap usually shows up in the same places. Backups that live on the same infrastructure or behind the same credentials as production, so a single compromise takes both. Restores tested at the file level but never as a full system. And recovery plans written against a hardware failure rather than the scenario that actually happens now, where the data is intact but encrypted and the pressure is to restore quickly and cleanly.
Fixing this is rarely about buying more storage. It is about isolating the copy, agreeing the recovery targets, and running a real restore on a schedule. We scope that against what you already own before recommending anything new — and when new capability is genuinely warranted, we bring competing options for it.
Colocation and Data Center Options Around St. Louis
Colocation gets overlooked because cloud absorbs the attention, but for hardware you already own and intend to keep, it is often the cheaper and simpler answer. Here is how the decision usually breaks down.
Metro facilities for latency-sensitive workloads
For applications your staff use constantly, or that talk to systems still sitting in your building, proximity matters. A metro facility keeps round-trip times low while getting the equipment out of a closet that was never designed to be a data center.
Regional facilities for disaster recovery
A recovery site should not share a power grid, a flood plain, or a weather event with production. Regional facilities a few hours out are usually the right distance — far enough to be genuinely independent, close enough that someone can drive there.
Carrier density and why it matters
The number of networks present in a building determines your connectivity options, your pricing leverage, and whether a genuine second path exists. A well-connected facility gives you competition on circuits for as long as you are there; a poorly connected one quietly does not.
Which facility fits depends on where your people are, what your applications talk to, and what your recovery targets are. We are not tied to any operator, so the shortlist gets built from your requirement rather than from who we happen to have space with.
Cloud and Colocation Questions We Hear Most
Should we move everything to the cloud?
Almost never. In most environments some workloads get cheaper and more resilient in public cloud, some are meaningfully cheaper on dedicated hardware in a colocation facility, and a few legacy applications are best left alone until they are replaced. A good design is usually a mix, and anyone recommending all-in without reviewing your workloads is selling rather than advising.
Is the cloud cheaper than buying servers?
Sometimes, though not always for the reasons people expect. Cloud converts a capital purchase into an operating expense and removes the refresh cycle, which has real financial value. But steady-state workloads that run continuously often cost more per month in a hyperscaler than the equivalent hardware in a colocation cabinet. Cloud wins clearly on variable and unpredictable demand. Dedicated hardware tends to win on flat, predictable load.
What is colocation, in plain terms?
You own the servers; someone else owns the building, the power, the cooling, and the network connections. You rent space by the cabinet, and it usually comes with redundant power, generator backup, physical security, and a choice of carriers. It is the middle ground between a closet in your office and renting compute from a cloud platform.
How long does a cloud migration take?
For a mid-sized company, a typical migration runs a few months from decision to completion, though it depends far more on application complexity than on data volume. Straightforward file and virtual server moves go quickly. Applications with database dependencies, licensing constraints, or an integration nobody fully documented are what extend timelines. The provider performs the migration; we scope it and hold them to the plan.
Do you actually perform the migration?
No. We are a broker. The provider you select does the migration, the engineering, and the ongoing operation. Our role is defining the requirement, running competing providers against it, negotiating the terms, and managing that provider through delivery so the project stays on track and somebody answers when it does not.
What does cloud and data center sourcing cost us?
Nothing. The provider you select compensates us once a contract is signed — a small monthly residual paid over the life of the agreement rather than a lump sum. Where one provider's compensation differs from another's, that difference never decides what we recommend. You get the comparison and the negotiation at no charge, and pricing through this channel is generally the same or better than going direct.
Our cloud bill keeps going up. Can you help?
Yes, and it is a common reason people call. Bring a recent invoice. Most overruns trace to a short list of causes — instances sized against old on-premise specs, resources running around the clock for a business-hours workload, storage tiers nobody reviews, data transfer charges, and orphaned volumes and test environments. We separate what is fixable by tuning from what is genuinely a pricing problem worth taking to market.
What is the difference between backup and disaster recovery?
Backup is a copy of your data. Disaster recovery is a tested plan for getting the business running again, including the infrastructure to run it on. Plenty of companies have solid backups and no realistic recovery plan, which is how a restore that should take hours ends up taking days.
How much data would we actually lose in an outage?
That is your recovery point objective, and it is set by how often your data is replicated or backed up. Nightly backups mean up to a day of work at risk. Continuous replication brings that down to minutes. The right answer is a business decision about what a day of lost work would cost you, not a technical default.
Do we need a data center facility near St. Louis?
For production workloads your staff hit constantly, proximity helps. For disaster recovery, distance is the point — a recovery site sharing your power grid or weather exposure is not really a recovery site. Many companies end up with a metro facility for production and something several hours away for recovery.
Can we keep the servers we just bought?
Usually, and often you should. Hardware with useful life left is frequently best moved into a colocation cabinet rather than written off, which gets it out of an office closet and onto proper power, cooling, and carrier connections without a capital write-down.
What happens to our data if we want to leave a provider?
It depends on terms most people do not read until they need them. Egress charges, notice periods, and how proprietary the services you built on are all determine what an exit costs. We raise these at contract time, when you still have leverage, rather than at renewal when you do not.
Does moving to the cloud make us more secure?
It changes what you are responsible for rather than removing the responsibility. The platform secures the underlying infrastructure; you remain responsible for identity, access, configuration, and your data. Most cloud incidents trace to misconfiguration or stolen credentials rather than to a platform breach.
We have an MSP already. Does this conflict?
Usually not. Plenty of capable MSPs do not source infrastructure competitively, or resell one platform because of a partnership rather than because it fits your workload. We are frequently brought in alongside an incumbent to run the selection, and they carry on doing what they do well.
Find Out What You Should Actually Be Running
A 30-minute conversation about what you run today, what is coming off support, and whether cloud, colocation, or staying put is the right answer — with competing quotes for whichever it turns out to be.
Book a CallGet Competing Quotes
Tell us what you're running and we'll bring competing bids from 200+ providers — at no cost to you.
No cost, no obligation. Gateway Telecom Advisors is paid by the provider you select.