Managed Hosting, Bought the Way It Should Be: Providers Competing for You

Managed hosting means someone else owns the hardware and takes responsibility for keeping it running. The word managed covers an enormous range of arrangements, and the difference between them is precisely where outages become somebody's fault. This hub explains what to pin down.

What Managed Hosting Actually Covers

In managed hosting the provider owns the hardware and rents you its use, along with some agreed level of responsibility for keeping it operational. That responsibility is the product, and it varies enormously between providers using identical language.

At the lighter end, managed means the provider replaces failed hardware and keeps the network and power running, while everything above the operating system is yours. At the heavier end it includes patching, monitoring, backup, security hardening and application support.

Both are legitimately called managed hosting. The failure mode is assuming you bought the second and discovering during an incident that you bought the first, which is why the responsibility boundary deserves more scrutiny than the specification.

Where It Sits Between Colocation and Cloud

The three models differ mainly in who owns what and who is responsible when something breaks.

Drawing the Responsibility Line

Before comparing prices, write down who does what, and get the provider to confirm it. Ask specifically: who patches the operating system, and on what schedule. Who monitors the service, and does that mean the server responding or the application working. Who takes backups, where do they go, and who has ever tested restoring one.

Then ask the uncomfortable questions. If the application stops responding at three in the morning, who notices, and what do they do next. If a security vulnerability is announced, who decides whether it applies and who applies the fix. If data is lost, what is the recovery point and how long does recovery take.

Providers answer these readily when asked directly. The problems arise when nobody asks and both parties assume the other is covering it, which is the most common cause of the outage that surprises everyone.

Backup Is Not Backup Until It Has Been Restored

Almost every managed hosting agreement mentions backup, and mentioning it is not the same as being able to recover. Three questions turn it into something real.

How far back can you go, and at what granularity. A nightly snapshot retained for a week is a very different product from point-in-time recovery over months, and both get called backup. Where do the copies live: a backup on the same hardware protects against deletion but not against the hardware failing, and a backup in the same facility protects against neither a fire nor a flood.

And who performs a restore, how long it takes, and when it was last tested. A restore nobody has ever attempted is a plan rather than a capability. Ask for the result of the most recent test, and if the answer is that there has not been one, that is the answer.

When Managed Hosting Beats the Alternatives

Managed hosting suits organisations with steady workloads and a small internal team, which describes a great many businesses. You get hardware you do not have to buy, replace or capacity-plan, and a predictable monthly cost rather than a variable one.

It beats colocation when you would rather not own or maintain hardware and do not need the control that ownership brings. It beats cloud when demand is predictable enough that paying by the hour costs more than committing, and when you would rather have a provider responsible for the platform than manage it yourself.

It is the wrong choice when demand varies sharply, since you are committing to fixed capacity, and when you need deep control of the platform, since that control is exactly what you have delegated. Be honest about which of those describes you before comparing quotes.

How the Managed Hosting Marketplace Works

Because managed means different things to different providers, quotes are only comparable when the responsibility boundary is specified up front.

• Describe the workload and the boundaryWhat runs, and which responsibilities you want the provider to hold.
• Providers quote the same scopeSo differences reflect service rather than a different reading of managed.
• Compare responsibilities and responsePatching, monitoring, backup and escalation, not just the monthly fee.

What to Check Before You Sign

Nearly every dispute in managed hosting traces back to a responsibility neither side wrote down.

• Who patches, and how quickly? Operating system and application, with a stated schedule for urgent fixes.
• What exactly is monitored? Server responding is not the same as application working. Confirm which.
• Where do backups live? Same hardware, same facility or offsite. Each protects against different things.
• When was a restore last tested? An untested backup is a plan, not a capability.
• What is the escalation path out of hours? Who is contacted, how, and what response time is committed.
• Is the hardware dedicated or shared? It affects performance consistency and sometimes compliance obligations.
• What happens at the end of the term? How you get your data out, in what format, and at what cost.
• Who holds administrative access? And whether you retain enough to act independently in an incident.

Managed hosting, answered