NAS file copies are slow but internet speed is fast: test the local path https://wiretraceuk.co.uk/guides/nas-file-copy-slow-internet-speed-fast/ Publisher: WireTrace UK Published: 2026-11-20 Updated: 2026-11-21 Local network performance NAS file copies are slow but internet speed is fast: test the local path Separate the laptop link, switch path, storage and file workload instead of treating an internet speed test as a NAS benchmark. WireTrace UK editorial team · Published 20 November 2026 · 6 min read · Updated 21 November 2026 In this guide Keep the two paths separate Make one repeatable file comparison Bypass one segment at a time Use a network-only test when it answers a remaining question Match the next action to the bottleneck Comparison record Sources and scope Start here. Compare the same large test file over a known-good wired path and the normal connection, while recording the NAS and client workload. Internet throughput measures a different destination and cannot prove the local storage path is healthy. The question: My broadband speed test is fast, but copying files to the NAS is slow. Is it the cable or the disks? Keep the two paths separate An internet test transfers data to an external server. A local NAS copy passes through your device, adapter, local link, switch, NAS interface, storage and file protocol. One path can be fast while the other has a bottleneck. Confirm you are actually using the local address or supported local connection rather than a cloud relay or remote-access route. Record read versus write direction, file type and size, device connection and any NAS activity. Many small files exercise metadata and storage latency differently from one large file. An encrypted or compressed workload may load a processor. A single number without these conditions cannot tell you whether the network or storage is limiting performance. [Image: A home storage device used for a local photo-backup comparison] A home storage device used for a local photo-backup comparison. AI-generated illustrative scene; not an installation or product-test photograph. Make one repeatable file comparison Use a non-sensitive test file large enough to observe a stable transfer, and copy it to a suitable temporary location with enough free space. Do not overwrite originals or use a production share where testing would disrupt work. Record the operating system's units: MB/s and Mb/s differ by a factor of eight before protocol overhead. For example, 90 MB/s is approximately 720 Mb/s. A 1 Gb/s link has a theoretical raw rate of 125 MB/s, but file throughput is lower because of framing, protocol and system overhead. A value below the raw ceiling does not by itself prove a fault. Repeat the same direction and account for caching rather than comparing a cached read with a first-time write. Bypass one segment at a time Connect the same laptop by a known-good short Ethernet lead to the intended switch or router. Check negotiated link rates in the device and NAS management views where available. The browser cannot read every port's negotiated speed. A USB dock, a 100 Mb/s port or a NAS connected through a wireless mesh backhaul can limit the path even when the broadband service is faster. If the wired comparison improves markedly, investigate the normal Wi-Fi or backhaul segment. If it remains slow, check NAS disk activity, rebuilds, backups, CPU and free-space warnings using the vendor's supported interface. Avoid disabling SMB signing, encryption, antivirus or other protections to chase a headline rate. Those settings have security and compatibility consequences. Use a network-only test when it answers a remaining question A tool such as iperf3 can compare network throughput between two authorised endpoints without the same file-storage workload. It needs suitable software at both ends and access permitted by the network administrator. A good network-only result with a slow file copy points towards storage, protocol or workload for further investigation; it does not prove which one is at fault. Keep any test service restricted to the trusted local network and stop it afterwards. Do not open an internet-facing port for this purpose. Choose an existing permitted diagnostic method if installing a service on the NAS is unsuitable. A browser-based internet test cannot substitute for this local endpoint-to-endpoint comparison. Match the next action to the bottleneck Evidence | Next check Only the normal Wi-Fi path is slow | Compare location, node and mesh backhaul. A port negotiates at an unexpected lower rate | Check its exact capability, patch leads and installed run. Network-only test is healthy; file copy is slow | Check storage state, protocol, processor load and file workload. Only small-file jobs are slow | Compare metadata and storage behaviour before replacing the cable. A multigigabit switch helps only when the relevant endpoints and rest of the path support it and the storage can use the added capacity. A continuity tester helps identify basic disconnected-cable faults but does not certify throughput. After a change, repeat the original copy direction and file set, then confirm normal office workloads still behave correctly. Keep a comparison record Use a non-sensitive test file and the same transfer direction for each comparison. Record bytes and elapsed time, then keep MB/s separate from Mb/s. A folder containing thousands of small files is a different workload from one large file. Note the negotiated rate at both ends, the switch path and whether either endpoint is on Wi-Fi. If permitted, run a local network-only test between suitable endpoints to remove storage from that comparison; an internet speed test cannot do this. Avoid changing NAS security, RAID or filesystem settings during a connectivity check. Record indexing, backup or rebuild activity that could compete for disk time. Repeat the original file workload after the network or storage branch is addressed. Fields in the worksheet: Client and connection; NAS connection; Negotiated link rate; Test file size and count; Transfer direction; Elapsed seconds; MB/s; Local network-only test; NAS disk task; Protocol or encryption; Next comparison. Download the blank CSV worksheet (https://wiretraceuk.co.uk/_wt/editorial-20261120-real/worksheets/nas-file-copy-slow-internet-speed-fast.csv) Questions that change the next step Is 10 MB/s evidence of a 100 Mb/s link? It is consistent with that scale but not proof. Check actual port negotiation and compare a known-good path; storage or other limits can produce a similar rate. Should I enable jumbo frames everywhere? Not as an initial fix. They require consistent supported configuration along the path and do not remove storage or low-rate-port bottlenecks. Continue with the evidence you have Check an unexpectedly slow USB Ethernet path (https://wiretraceuk.co.uk/guides/usb-ethernet-adapter-dock-slow/) Eight continuity lights but only 100 Mb/s (https://wiretraceuk.co.uk/guides/ethernet-continuity-passes-but-100mbps/) Confirm mesh backhaul (https://wiretraceuk.co.uk/guides/mesh-ethernet-backhaul-check/) Match a tool to the unresolved check → (https://wiretraceuk.co.uk/testers/#tool-options) Questions before you change anything Why are NAS copies slow when my internet speed is fast? A local copy follows the LAN path, storage system, file protocol and client workload. It may never use the provider connection, so an internet speed test cannot establish the NAS transfer limit. Is MB/s the same as Mbps? No. Compare megabytes per second with megabits per second; eight bits make a byte. Allow for protocol overhead and rate averaging before deciding that a displayed copy rate is unexpectedly low. How do I tell a disk limit from a network limit? Compare large-file transfer, small-file workload, NAS CPU use and link rate. Cached reads can differ greatly from sustained writes, so state which operation was measured. Could the switch or USB dock limit NAS speed? Check every Ethernet hop, adapter capability, USB connection and switch port. One 100 Mbps link can constrain a path made mostly of gigabit devices. Should I buy multi-gig Ethernet for the NAS? First establish disk throughput, NAS port support, client port support and useful transfer size. An upgrade helps only when the relevant network hop is the demonstrated bottleneck. Questions and source links reviewed 21 November 2026. The answers apply the limits described in this guide; they are not a test of your equipment. Sources and scope These steps combine the references below with a practical comparison sequence. Calculated examples are illustrative; no photograph or example is a claimed bench test. Follow the exact device documentation and site methodology (https://wiretraceuk.co.uk/methodology/). ESnet: iperf3 documentation (https://software.es.net/iperf/) Cloudflare: speed-test measurement scope (https://speed.cloudflare.com/about) Fluke Networks: cable-testing capabilities (https://www.flukenetworks.com/expertise/learn-about/cable-testing) [Image: Rack-mounted servers and network switching equipment] Rack-mounted servers and network switching equipment. Photo: LSST Project/NSF/AURA (https://commons.wikimedia.org/wiki/File:Installation_of_the_first_summit_network_equipment_May_2018_(rubin-2018-05-02-192510).jpg) · CC BY 4.0. Context photograph; not a measurement result for this guide. [Image: Local storage path for comparing file-copy performance — AI-generated scene illustration] Local storage path for comparing file-copy performance. AI-generated scene illustration; not a measured result or an exact wiring reference. Reader questions that informed this update: Reddit: reader discussion (https://www.reddit.com/r/HomeNetworking/comments/1pazsks/gigabit_internet_speed_capping_out_at_100mbps_at/) · Reddit: reader discussion (https://www.reddit.com/r/HomeNetworking/comments/13ijrwd/how_can_i_fix_my_double_natget_open_nat_in_games/). Individual reports help identify questions; answers above use the technical references in this guide. Explain your own result Replace the brackets with observations you actually made. Leave unknowns as “not checked”. This draft is written for this guide; it is not a quotation or an automatic diagnosis. A prompt for an assistant or a support conversation NAS [model/path] and client [connection/rates]. Copy [read/write] of [large file/small files] gives [MB/s], local network-only test [result], CPU/disk/rebuild [observations]. Help separate network, storage and protocol limits without changing security or overwriting production files. Copying stays on your device. Remove passwords, account numbers and private network details before sharing. Reader questions that informed this update: Reddit: reader discussion (https://www.reddit.com/r/HomeNetworking/comments/1pazsks/gigabit_internet_speed_capping_out_at_100mbps_at/) · Reddit: reader discussion (https://www.reddit.com/r/HomeNetworking/comments/13ijrwd/how_can_i_fix_my_double_natget_open_nat_in_games/). Individual reports help identify questions; answers above use the technical references in this guide. In this guide Keep the two paths separate Make one repeatable file comparison Bypass one segment at a time Use a network-only test when it answers a remaining question Match the next action to the bottleneck Comparison record Questions Sources and scope Need an earlier starting point? Test the connection, then compare the failing task and location. Open connection checks → (https://wiretraceuk.co.uk/troubleshooting/#diagnosis) Original guide: https://wiretraceuk.co.uk/guides/nas-file-copy-slow-internet-speed-fast/