This article was machine-translated from French and is being reviewed. Prices are in Algerian dinars excl. VAT.
For an SME, choosing between a server or a NAS depends on what the machine must do. A NAS is a network storage enclosure: it suits file sharing and backups. A server is an enterprise-class computer that runs your applications: business software with a database, Active Directory directory, virtual machines. Many SMEs need both. In any case, neither replaces a backup: RAID protects against disk failure, not deletion, ransomware or fire. The store statement of 06/11/2026 contains neither NAS nor server: prices are given on quote.
In short
- NAS: network storage with its disks and an integrated system, easy to administer. It serves files, backup and, depending on the model, a few services.
- Server: enterprise-class hardware (ECC memory, redundant power supplies, remote administration depending on the model) to run your applications continuously. It requires a competent administrator and licenses.
- In both cases: a UPS sized in watts, a RAID suitable for disk capacity, and backup on another medium, tested by restoration.
NAS or server: technical comparison
| Criterion | NAS | Server |
|---|---|---|
| Main role | File sharing, backups, light services | Business applications, databases, directory, virtualization |
| System | Manufacturer-specific system (DSM for Synology, QTS for QNAP) | Windows Server, Linux or hypervisor |
| Memory | Depending on the model, with or without ECC | ECC common on professional models |
| Power supply | Often single | Redundancy possible, depending on the model |
| Remote administration | NAS web interface | Dedicated admin controller, depending on the brand |
| Licenses | NAS system included, paid options per manufacturer | Operating system, plus user access to license |
| Evolution | Limited by the number of bays and the processor | Processor, memory, disks and expansion cards depending on the chassis |
The ECC memory detects and corrects single-bit errors, limiting silent corruption on a machine that stays on for months.
When to choose a NAS
Choose a NAS if:
- you share files between multiple workstations (accounting, commercial documents, scans);
- you want a central backup target for PCs and the server;
- you do not have a full‑time IT administrator: the NAS web interface is designed for this case;
- you do not run heavy business applications.
A NAS has limits. It is a single point: if the unit fails, no one can access the files. Some models run containers or virtual machines (Synology and QNAP, depending on CPU and memory), but that is not their main role. Also check it can join the company directory to apply per‑user access rights: this is common, but verify on the model sheet.
When to choose a server
Choose a server if:
- you use a client‑server business software. Example: Santymed is a locally installed solution, whose database resides on the institution’s server (FAQ of the software vendor, consulted 06/11/2026). The vendor’s requirements are not published: request them before any purchase, see Install Santymed: server, network, backup and UPS ;
- you want a centralized directory (Active Directory): one account per user, group rights, policies applied to workstations. This does not depend on a minimum number of stations;
- you need virtualization;
- you need availability: hot‑replaceable disks, redundant power supplies, manufacturer support;
- your team, or a contractor, can administer Windows Server or Linux.
A server is licensed. With Windows Server, the Standard and Datacenter editions are licensed per CPU core, with a minimum of 8 core licenses per processor and 16 per server. Each user or device that connects also requires an access license (CAL). The Standard edition covers two virtual machines, the Datacenter unlimited when all cores are licensed. The Essentials edition, limited to 25 users and 50 devices, does not require CALs. License prices are requested from the reseller: we do not publish them here.
RAID: what it does and what it does not
RAID groups multiple disks into a single volume to distribute or duplicate data. Here are the common levels, for disks of equal capacity C and n disks.
| Level | Principle | Minimum disks | Usable capacity | Disk failures tolerated |
|---|---|---|---|---|
| RAID 0 | Data distribution, no redundancy | 2 | n × C | None: a failure loses all volume |
| RAID 1 | Mirror | 2 | C (for 2 disks) | 1 of 2 |
| RAID 5 | Distribution with one parity | 3 | (n − 1) × C | 1 |
| RAID 6 | Distribution with two parities | 4 | (n − 2) × C | 2, whatever they are |
| RAID 10 | Distributed mirrors | 4 | (n ÷ 2) × C | 1 per mirrored pair; losing both disks of a pair destroys the volume |
Example with 4 disks of 4 TB: RAID 0 gives 16 TB, RAID 5 gives 3 × 4 = 12 TB, RAID 6 and RAID 10 each give 2 × 4 = 8 TB. A disk sold as 4 TB shows about 3.6 TB in Windows, because the manufacturer uses powers of 10 and the system uses powers of 2: the actual capacity is always less than the sum of the labels.
RAID is not a backup. It protects against disk failure. It does not protect against an accidentally deleted file, ransomware that encrypts the volume, software corruption, theft, a surge or fire: the error is immediately replicated on all disks. With large disks, reconstruction after a failure takes many hours. During that time, a RAID 5 tolerates no other failure, hence the preference for RAID 6 or RAID 10 on large capacities.
Disks: CMR, SMR and usage
A CMR disk writes tracks side‑by‑side. An SMR disk overlaps them like tiles: it stores more data, but its writes are slower and, during a RAID reconstruction, it can slow down significantly or be excluded from the volume. In 2020, several WD Red models of 2 to 6 TB (EFAX references) turned out to be SMR, leading Western Digital to distinguish the Red Plus and Red Pro ranges in CMR. Before buying, note the exact disk reference and check the manufacturer’s datasheet to see if it is CMR or SMR, as well as in the NAS manufacturer’s compatibility list.
Use disks marketed for NAS by their manufacturer. The so‑called “surveillance” disks, designed for continuous video stream writes, meet a different load profile: see How many days of recording with which hard disk for this use.
Size the storage
- Measure the current data, folder by folder, with a disk‑space analysis tool.
- Estimate growth from the previous year, without applying a standard rate.
- Calculate usable capacity with the chosen RAID level formula, to cover data and growth over the expected usage period, leaving free space.
- Plan backup on another medium and in another location: it must not be stored on the same volume.
- Keep racks free for expansion.
Example calculation with fictitious values: 6 TB of data today, plus 1 TB per year for 5 years, i.e. 6 + 5 × 1 = 11 TB to plan for. In RAID 5 with 4 disks, (4 − 1) × C ≥ 11 gives C ≥ 3.67 TB: 4‑TB disks suffice (12 TB usable). In RAID 6 with 4 disks, (4 − 2) × C ≥ 11 gives C ≥ 5.5 TB: 6‑TB disks are needed (12 TB usable). The extra protection costs in disk capacity.
Installation and maintenance
A technical visit is useful before purchase to:
- evaluate existing data (volume, file types);
- record applications, users and their rights;
- design redundancy (RAID, external backup);
- test restoration: a backup never restored is not a backup.
AsloTech supplies the sized NAS or server, installs it, configures backups and provides a maintenance contract with remote support. Contract details are in Maintenance contract: what it must contain. Disk warranty depends on the manufacturer: ask for its duration before buying.
Also plan regular monitoring: disk status alerts sent by e‑mail, NAS or server system updates (which do not always apply automatically), and RAID health checks. Do not connect a NAS or server directly to the internet: remote access must go through a VPN or configured firewall, with named accounts. See Why a firewall in SMEs and Enterprise VPN: connecting sites.
Combined solution: server and NAS
For an SME with a business application and files to share, the two machines share the workload:
- a server or mini‑PC dedicated to the application and its database;
- a NAS dedicated to files, also receiving the server backup;
- a Gigabit network link at least between the two, on a proper switch;
- a off‑site copy, because a NAS in the same location as the server faces the same fire risk;
This setup separates roles: if the NAS fails, the application keeps running, and vice versa.
Backup: the 3-2-1 rule
Apply the 3-2-1 rule: 3 copies of the data, on 2 different media, including 1 off‑site (another location, portable disk or online storage). Instantaneous snapshots from a NAS help revert after an error or ransomware, but they stay on the same chassis: they do not replace the off‑site copy. Details are in Backup: the 3-2-1 rule explained.
Power outages and UPS
A NAS or server must be protected by a UPS sized according to the total wattage consumption (server, NAS, switch, router), with a margin, and not by a pre‑set VA value: see UPS: how to choose the power. A sudden outage can corrupt disks and databases. Connect the UPS to the machine (USB or network) to trigger automatic shutdown before the battery is depleted. If the server has two power supplies, plug them into separate sources when possible. In a rack, see how to organize a patch rack.
Frequently asked questions
What is the real cost of a NAS or server over 5 years?
The purchase price is only part of the cost. Add: the case or server, disks (and their replacement in case of failure), the UPS and its batteries, licenses (system, user access, options), external backup storage, and maintenance interventions. The price of each line is requested on a quote; the sum over the expected usage period gives the true cost.
Can the manager’s PC be used as a server?
It is not recommended. The Windows 10 and 11 license limits the number of devices that can access a workstation for file and printer sharing to 20. A desktop PC usually has no ECC memory, no redundant power supply, no remote administration, and it restarts for updates: everyone loses access. For simple file sharing, a NAS is more suitable: it is designed for continuous operation.
What is virtualization?
Virtualization allows running several isolated systems (virtual machines) on a single physical machine, thanks to a hypervisor. Hyper‑V, included with Windows Server, and Proxmox VE, free software, are two examples. It can separate, for example, the business application server from a test machine. The host must provide enough memory and cores for all virtual machines, and each Windows VM must be licensed. A NAS does not replace a server for this use, even if some models host a few machines.
NAS or cloud to store your files?
A NAS keeps data in your premises, accessible even if the internet goes down, but the hardware, its maintenance and protection are at your expense. A online service avoids buying hardware, but depends on the connection and a recurring subscription, and it hands data to a third party. For an SME, the combination is most robust: a local NAS and an encrypted copy sent to online storage as off‑site backup.
Is RAID sufficient as a backup?
No. RAID keeps the service running when a disk fails, but it immediately replicates any deletion, error or ransomware across all disks. A backup is an independent copy on another medium that you have tested by restoring it.
To dimension your infrastructure during a technical visit, discover our professional IT hardware solution. For network protection, see our network security and firewall solution and our IT maintenance solution. For clinic management software, see our clinic management software solution. Also read: UPS: how to choose the power and Renewing a company's IT fleet.