Data Solutions

A NAS is a storage appliance on your own network: drives in a chassis running their own operating system, sharing files to whoever is authorised and pulling scheduled backups off servers and PCs. It earns its place beside cloud backup because the restore is local. Recovering two terabytes from a NAS over the LAN is a lunch break; the same restore dragged back down an office internet line can run into days. The usual answer is both — the NAS for the fast restore, the cloud for the copy that survives the building.

RAID spreads data across several drives so the array survives one or more of them dying. RAID 1 mirrors a pair and leaves you half the raw capacity. RAID 5 survives one failure and costs one drive, but on today’s large disks a rebuild runs long enough that a second failure during it is a genuine risk. RAID 6 survives two failures and costs two drives — it is what we specify on anything built from big disks. RAID 10 mirrors and stripes for speed at half the raw capacity. Synology’s SHR and the QNAP equivalent do the same job while letting you mix drive sizes and grow later.

No, and this is the single most expensive misunderstanding in storage. RAID protects against a drive failing. It does nothing about ransomware, a deleted folder, a corrupted database, a fire or a theft — every one of those is written faithfully to all the drives at once. A RAID array is how the NAS stays up; the backup is the second copy somewhere else, which is why the cloud half of this page exists.

Bays decide capacity and how much of it redundancy takes back. Two bays in RAID 1 give you half the raw total; four in RAID 5 give three quarters; eight in RAID 6 give three quarters and survive two failures rather than one. Desktop units run from two to eight bays, rackmount units twelve to twenty-four, and enterprise chassis sixty or more. Buy a bay or two spare: adding drives to an existing array is far cheaper than replacing the unit because it filled up.

As of 2026, NAS-rated drives reach 30 TB — Seagate’s IronWolf Pro leads the NAS tier — and enterprise drives reach 32 TB through the channel. Larger ones exist: Seagate has announced 36 TB HAMR drives, but those go to cloud providers and hyperscalers before they reach resellers. Whatever the headline number, check the vendor’s own compatibility list first. Both QNAP and Synology publish tested-drive lists, and support cases go more smoothly when the drives are on them.

Far more than most businesses need. An eight-bay desktop unit filled with 30 TB drives is 240 TB raw. Rackmount units go further, and enterprise models scale with expansion shelves — Synology’s SA6400 takes up to 108 drives across eight expansion units for roughly 1.9 PB, and the HD6500 starts at 60 bays and reaches 300 drives fully expanded. In practice the ceiling is budget, rack space and power, not the platform.

Less than the drive labels add up to, for two separate reasons. Redundancy takes its cut first: eight 20 TB drives in RAID 6 is 160 TB raw and about 120 TB usable. Then the labels themselves are decimal while the operating system counts in binary, so a "20 TB" drive shows as roughly 18.2 TiB. Budget for about 70–75% of raw on a RAID 6 array, and leave headroom on top — snapshots need room, and any filesystem slows down near full.

Because a backup in the same building shares the building’s risks. A fire, a flood, a theft or a ransomware run that reaches the NAS over the network takes the original and the copy together. The off-site copy is the one that is still there afterwards. It is also the copy that answers an insurer or an auditor asking what happens if the office is unavailable.

Three copies of the data, on two different kinds of media, with one of them off site. It is the baseline every serious backup design still starts from. QNAP and others now promote 3-2-1-1-0: the extra 1 is a copy that is offline or immutable, and the 0 is zero errors — meaning the backups are actually verified rather than assumed. A backup nobody has ever restored from is a hypothesis, not a backup.

An immutable backup cannot be altered or deleted until its retention period expires — not by an administrator, and not by malware holding that administrator’s credentials. It matters because modern ransomware goes looking for the backups first: encrypt or delete those and the ransom becomes the only option. Both platforms we deploy support it — Acronis through immutable storage, Veeam through immutability on its repositories.

That depends on what you are restoring and how fast the line is, which is exactly why we pair the two. Whole-server recovery from cloud storage is bounded by your internet connection, so a large restore can take many hours. Veeam’s Instant Recovery sidesteps some of that by running a workload directly from the backup while it restores in the background. For anything large the local NAS copy is still the fast path, and the cloud copy is the fallback.

Yes — in transit and at rest. Both platforms encrypt before the data leaves your network, so what sits in the cloud is unreadable without your key. Key custody is a decision to make deliberately at design time: hold the keys yourself and nobody else can read the data, including the vendor and including us, but a lost key means a lost backup. We set this out in writing before the first job runs.

Yes. Microsoft operates the service and protects its own infrastructure; the data in your tenant is your responsibility, and its retention windows are short. A mailbox deleted past the retention period, a SharePoint library a script wrecked, or an account an attacker cleared out are all recoverable only from a backup you took. Both Acronis and Veeam back up Microsoft 365 specifically, and Synology and QNAP can pull a copy down to the NAS as well.

Three things: how much data you keep, how long you keep it, and how often you have to read it back. Storage is usually the smallest surprise — retention is what quietly grows a bill, because every extra month of history is another full set of changes held. Worth asking any provider about retrieval and egress charges before signing: some storage tiers are cheap to write to and expensive to restore from, which is the wrong shape for a backup.