Give the NAS a deadline before choosing its ports
A network-attached storage device, or NAS, puts files on a computer reached through your home network. Before buying one, establish whether that network can deliver the job you expect: overnight backups, photo imports or files opened from several rooms. An internet speed test cannot settle this question because a local transfer follows a different route.
This workshop produces a readiness record using two computers, Ethernet adapters where needed, two known-good patch cables and access to your router's device information. One computer temporarily occupies the planned NAS connection. If you have only one computer, complete the route inventory but mark performance unverified; a successful web download is not a substitute.
Write one acceptance target. In a hypothetical household, the target might be copying 30 GB of camera files within ten minutes. Using decimal units, 30,000 MB divided by 600 seconds requires an average of 50 MB/s, or 400 Mbps, before allowing for overhead and competing traffic. These are planning calculations, not measured results. Set a separate deadline for the much larger first backup.

Trace the route from the desk to the shelf
Choose a ventilated NAS location with power and an Ethernet connection on the household router's LAN side. Sketch the actual route: laptop, access point or mesh unit, its wired or wireless backhaul, switch and NAS. Include wall sockets and USB docks. A wireless mesh hop remains part of the route even when the NAS itself uses Ethernet.
Beside each wired link, record both port specifications and the negotiated rate if available. Distinguish advertised maximum capability from the current connection. A spare port may be slower than its neighbour. Seagate's NAS guidance identifies intervening hardware as a performance constraint; apply that principle to the path carrying this transfer, rather than assuming one old device slows every connection in the house.
A 1GbE segment has a raw ceiling of 125 MB/s; 100 Mbps means 12.5 MB/s, with usable file throughput below either figure. For the hypothetical ten-minute import, a 100 Mbps segment is a definite failure. A gigabit route is only a candidate. A 2.5GbE NAS cannot remove a gigabit limit elsewhere along the same path.
Make a wired reference connection
Place the stand-in computer at the NAS shelf and connect it to the intended LAN port. Connect the other computer by Ethernet, initially through the same switch or router. Keep the router providing addresses normally. Temporarily turn off Wi-Fi on these computers so the reference test uses the cables; note that setting for restoration.
On Windows, open PowerShell and run this read-only inventory, based on Microsoft's Get-NetAdapter documentation:
Get-NetAdapter -Physical | Format-Table Name, Status, LinkSpeed
Identify the connected Ethernet adapter by name and status. For a gigabit-capable pair, expect the active link to report 1 Gbps, not 100 Mbps. This reads the adjacent connection, not every link to the other computer. Check both endpoints and any switch-port status available through your equipment's documented controls. Other operating systems need their own documented adapter-status method.
If a capable link negotiates at 100 Mbps, reseat it, substitute a known-good short cable, then try another compatible LAN port. Bypass a wall run or dock one component at a time. Success is the expected rate remaining stable; repeated disconnections or a persistent lower rate leave that path unresolved. Keep automatic speed negotiation rather than forcing a faster value.
Rehearse the route without copying files
Install iperf3 on both computers from a trusted source appropriate to their operating systems. Check the installed help against ESnet's command reference. Use the stand-in's current LAN IPv4 address from the router's device information. The address below is hypothetical; replace it with that computer's address.
- On the stand-in, start the server:
iperf3 -s. - On the client, run
iperf3 -c 192.168.1.50 -t 30. This sends toward the future NAS location. - Run
iperf3 -c 192.168.1.50 -t 30 -Rto measure the reverse direction. - Save the final receiver summaries with their units and connection layout. Repeat each direction three times, then repeat from the actual Wi-Fi desk with the client's Ethernet disconnected and Wi-Fi restored.
The server normally listens on TCP port 5201. If necessary, permit that application only on your trusted local network, using the operating system's firewall documentation. Keep the firewall enabled and do not forward a router port. Stop the server with Control-C and remove any temporary permission afterward.
A connection refusal or timeout means no throughput result exists: check the address, running server, firewall permission and whether either computer is on guest Wi-Fi. Keep guests isolated; move the test client to the trusted household network. This test measures network transport between two computers, not NAS disk performance.
Judge the slowest useful result, then add ordinary household activity
Compare the lowest repeatable result in each direction with your target, keeping Mbps and MB/s separate. A result barely above the calculated requirement leaves little room for file-sharing overhead. For the hypothetical 400 Mbps requirement, repeated 430 Mbps transport results would be marginal; repeated 700 Mbps results would offer more headroom, without promising a 50 MB/s file copy. These numbers illustrate decisions, not a tested network.
Run another check during ordinary evening activity and record what else was happening. Ask household members before starting a throughput test because it can occupy shared capacity. If the quiet wired test is adequate but the upstairs desk fails, investigate that desk's wireless route. Improving NAS ports alone will not repair it. The existing wired mesh backhaul guide covers that separate upgrade.
Shared links also matter when two people copy files simultaneously. Add their required rates when both transfers traverse the same link, and allow headroom. Alternatively, schedule large backups separately. A single-client pass does not establish multi-client capacity.
Reserve a final gate for the actual NAS
Before purchase, record network readiness as provisional. Once the NAS is installed and its storage is healthy, create a temporary share with a named user who can read and write it. Use the NAS maker's current SMB instructions. On a Mac, Apple documents Finder > Go > Connect to Server for entering the server address and signing in. Confirm the expected share opens; discovery alone does not establish access.
Copy a disposable local file of at least 5 GB to the share and back into a different local folder. Time each direction, record size and elapsed time, and compare original and returned file hashes with your operating system's documented checksum tool. Then try a representative folder of small files. Keep originals throughout. A completed copy with matching hashes establishes more than a progress bar; it still does not test backup recovery.
Synology's transfer troubleshooting recommends a minimum 5 GB test file and distinguishes client, network and storage conditions. If transport passed but file copying disappoints, inspect client disk load, NAS storage health and background work before buying faster networking. Initial indexing or array maintenance can make an early comparison misleading. Leave MTU at its existing default during this workshop.
Leave a verdict someone can act on
Save the route sketch, link rates, dated transport summaries and outstanding file-copy gate. Mark the network ready provisionally when the intended rooms and normal load leave useful headroom; ready with a restriction when imports need Ethernet or backups need scheduling; or repair first when links disconnect, negotiate too slowly or cannot reach the trusted LAN.
Restore Wi-Fi and temporary firewall settings, remove test shares and files when finished, and retain the working cable layout. Leave remote access as a separate project. A reachable NAS with sufficient local throughput has passed a network check; protecting its contents still needs a household backup plan.
Sources, versions & scope
The steps describe the documented controls and expected checks; they are not a claim that every device or software release has been physically tested. Platform and version differences are stated with the procedure.
Knowligent Editors
First published · Updated
Documentation-led digital-life guidance, with AI-assisted research and drafting. Examples are identified, source links are provided, and device-testing limits are stated.
Editorial standards

