Buying Your First Business Server: What to Specify
Most businesses buy their first server when files scattered across desktops stop being manageable, or when an application requires one. It is a purchase people overthink in some directions and underthink in others.
Do you need a server at all?
Worth asking honestly, because the answer is sometimes no.
If the requirement is file sharing and basic collaboration for a small team, cloud services frequently do it without hardware, maintenance or backup responsibility.
A server earns its place where there is a specific reason it has to be local: an application that requires it, data volumes that make cloud impractical or expensive, unreliable internet where local access must keep working, regulatory requirements about where data sits, or equipment that has to talk to something on the premises.
Many small businesses end up with both — cloud for productivity, a local server for the line-of-business application. That is a sensible outcome rather than a compromise.
Tower or rack
If you have no rack, a tower server is usually right.
Towers tolerate ordinary office conditions, run quieter, and need no rack infrastructure, PDUs or structured cabling. Many have rack-mount conversion kits, so growing into a rack later is possible.
Rack servers make sense once you have a rack, or a comms room where equipment belongs. Be aware that 1U servers are loud — fine in a comms room, unpleasant in an office.
One thing to plan regardless: where it will live. A server under a desk collects dust, gets knocked, and has no physical security. A cupboard with no ventilation cooks it. Neither is dramatic on day one and both shorten hardware life.
What to specify
Memory is the component most often under-specified. It is the constraint that bites first, particularly if you run virtual machines. Buy more than you think you need, and check the population rules so you leave room to expand — our memory guide covers this. ECC is standard on servers and worth understanding: it corrects errors that would otherwise cause silent data corruption.
Processor matters less than people assume for a first server. A modest processor handles file serving, a domain controller and a typical line-of-business application comfortably. If you plan to run several virtual machines, more cores help — but check per-core licensing costs before buying cores you will pay for repeatedly.
Storage needs two decisions: capacity and redundancy.
For capacity, size backwards from what you actually need usable, because RAID overhead and unit conversion both reduce it considerably. Our capacity guide covers the arithmetic.
For redundancy, a mirrored pair for the operating system and a separate array for data is the standard arrangement. Our RAID guide covers levels.
The things people skip
Four items that are not optional and frequently get cut from a first-server budget.
Backup. RAID is not backup — it protects against a drive failing, not deletion, corruption, ransomware, fire or theft. A server with no tested backup is one mistake from total loss. Budget for it as part of the server, not afterwards. Our backup guide covers the approach.
A UPS. It protects against the electrical events that happen constantly, not just outages, and it allows a clean shutdown rather than an abrupt power loss. Include shutdown software — a UPS without it provides runtime and nothing else. See our UPS guide.
Spares. At minimum, one spare drive matching what is fitted. A failure with no spare means running degraded until delivery.
Remote management. Every server has a management controller, but remote console is frequently a licensed feature. Without it you can see that the server failed to boot and cannot see why. For a business with no on-site IT staff, that licence is the difference between a remote fix and a call-out. See our management guide.
New or refurbished
A fair question for a first purchase.
Refurbished enterprise hardware gives considerably more machine for the money — more memory, more drive bays, better build than an entry-level new server at the same price. Enterprise equipment is built for years of continuous operation.
The considerations: vendor support may not be available on secondary-market hardware, and the platform will be older, which affects how long it stays supported.
A common sensible position is refurbished for a first server where budget is tight and the workload is modest, with spares held rather than a support contract. Our refurbished process page covers testing and warranty.
Whichever you choose, ask about the parts availability for the platform. That determines how long it stays maintainable, which matters more than the support end date.
A first-server checklist
Confirm you actually need local hardware rather than cloud. Decide where it will physically live, with ventilation and some security. Specify memory generously. Size storage backwards from usable capacity. Plan RAID for the data and a mirror for the OS. Budget backup, UPS with shutdown software, one spare drive, and remote management access. And establish who fixes it when it breaks, because that is the question a first server most often exposes.
Send us what the server needs to do and we will specify it. The quote page explains what to include.
Common questions
Do I need a server or will cloud do?
For file sharing and collaboration alone, cloud frequently does it without hardware or maintenance. A server earns its place where something must be local — a specific application, data volumes, unreliable internet, regulatory requirements, or on-premises equipment it must talk to.
Tower or rack for a first server?
Tower, if you have no rack. Towers tolerate office conditions, run quieter and need no rack infrastructure, and many have conversion kits for later. Plan where it will live — under a desk collects dust and an unventilated cupboard cooks it.
What is most often under-specified?
Memory. It is the constraint that bites first, particularly with virtual machines. Processor matters less than people assume for file serving and a typical line-of-business application.
What gets cut from the budget and should not be?
Backup, a UPS with shutdown software, one spare drive, and remote console access. RAID is not backup, and without remote console you can see a server failed to boot but not why — which for a business with no on-site IT means a call-out.
Is refurbished sensible for a first server?
Frequently yes — it gives more machine for the money than an entry-level new server, and enterprise hardware is built for continuous operation. Vendor support may not be available, so hold spares instead, and ask about parts availability for the platform.
Tell us what the server needs to do and we will specify it, including the backup and spares that should come with it.
