How to Build a Reliable Synology NAS Performance Baseline
A performance benchmark baseline gives you a clear picture of how your Synology NAS behaves before you add services, replace drives, enable SSD caching or change the network. Instead of relying on vague impressions such as “file transfers feel slow”, you record repeatable measurements for storage, networking, applications and system health.
This baseline becomes a reference point for troubleshooting and upgrades. It can reveal whether a slowdown comes from hard drives, a saturated link, encryption, indexing, Docker containers or an overloaded backup schedule. For Australian households and small businesses, it is also useful when comparing performance across NBN connections, remote offices in Perth or Melbourne, and local hardware purchased through Australian retailers.
Define The Testing Environment
Start by recording the hardware and software that may influence the results. Note the Synology model, DSM version, installed memory, drive models, RAID or SHR layout, storage capacity and the number of active volumes. Record whether the NAS uses 1GbE, 2.5GbE, 5GbE or 10GbE, and identify the switch, router and cabling between the test computer and NAS.
The client device matters just as much. A modern desktop with an NVMe SSD can expose NAS limits, while an older laptop with a slow internal drive may become the bottleneck. Use the same computer, network port and cable for every test. Wi-Fi should generally be excluded from the core baseline because interference and signal strength can vary significantly between rooms, especially in dense Sydney or Melbourne apartment buildings.
Record the testing conditions as well. Note the time, approximate room temperature, running services and whether Hyper Backup, Synology Drive, media indexing or surveillance recording is active. A NAS tested during a quiet Sunday morning may produce very different results from one handling several household backups during the evening.
Choose Useful Performance Metrics
A balanced NAS performance test should cover sequential throughput, small-file performance, latency and system resource use. Sequential reads and writes represent large media files, virtual machine images and backup archives. Small-file operations are more relevant to office documents, source code, photo libraries and applications that constantly open and update many files.
Measure both read and write activity. A RAID array may write more slowly than it reads, and a volume with SSD caching can show impressive repeated reads while offering little improvement for new data. Capture transfer speed in MB/s, operations per second where available, average latency and the duration of each run.
Also monitor CPU, memory, disk utilisation and network traffic in DSM Resource Monitor. A result of 110 MB/s on a 1GbE connection may be close to the practical network ceiling, while the same result on a 10GbE link could indicate a storage or configuration problem. For useful background on NAS hardware and storage features, the NAS storage guides provide additional product context.
Establish A Clean Baseline
Before testing, allow DSM and storage services to settle. Pause scheduled backups, cloud synchronisation, media conversion, antivirus scans and large downloads. Confirm that the volume is healthy in Storage Manager, check the storage pool status and review any recent drive warnings. Do not run demanding benchmarks on a degraded RAID or SHR array, since the results will be abnormal and the extra workload may increase risk.
Create a dedicated test folder rather than benchmarking an active work share. Use a mixture of files: a large multi-gigabyte video file, several hundred medium-sized documents and a large collection of small files. Keep the test data reproducible so you can repeat the same workload after a DSM update or hardware change.
Run each test at least three times and record the median rather than the single best result. The first run may be affected by cache warming, antivirus inspection or file indexing. Between runs, allow the NAS to return to a similar idle state. If your NAS serves files to a Brisbane office while the main system is in another location, test the local LAN and the remote connection separately.
| Test Area | Suggested Workload | Measurements To Record | What It Reveals |
|---|---|---|---|
| Large-file read | 10–50 GB video or archive | MB/s, duration, network rate | Sequential read and network limits |
| Large-file write | Copy the same data to a new folder | MB/s, duration, CPU use | Sequential write and RAID overhead |
| Small-file transfer | Thousands of documents or photos | Files per second, latency | Metadata and random I/O performance |
| Concurrent access | Two to four clients copying files | Aggregate MB/s, latency | Sharing capacity under load |
| Application load | Drive, Plex, containers or VMs active | CPU, RAM, disk utilisation | Service impact on responsiveness |
| Backup workload | Hyper Backup or USB backup job | Backup rate, duration, errors | Sustainable real-world throughput |
Test Network Throughput Separately
A fast NAS cannot overcome a slow or unstable network path. Test the connection between the client and NAS independently from file copying, using a suitable LAN throughput utility or a controlled large-file transfer. Check negotiated link speed on the NAS, switch and computer. A cable that has fallen back from 2.5GbE to 1GbE can make an upgrade appear ineffective.
Test wired performance first, then Wi-Fi as a separate scenario. A Wi-Fi 6 laptop may show excellent results close to the access point but much lower throughput through walls. For Australian homes using an NBN modem-router supplied by an internet provider, remember that internet speed and local NAS speed are different measurements. An NBN 1000 service does not make a local file copy faster if the NAS and computer are connected through 1GbE.
Internet-based testing deserves its own record. Uploading to a remote NAS, cloud destination or office in another city includes ISP routing, latency and provider congestion. Keep those results separate from the LAN baseline. A reference download from an external website, such as oil product information, should likewise be excluded from storage benchmarks because internet response time can distort the result.
Compare Realistic Workloads
Synthetic benchmarks are useful, but practical workloads often tell you more. Copy a large 4K video, open a project folder containing many small files, stream media while copying data and run a scheduled backup. These tests show how the NAS behaves when several services compete for storage resources.
For media servers, record whether Plex or Video Station is direct-playing or transcoding. Direct play mostly tests network delivery, while transcoding can consume substantial CPU or hardware acceleration resources. A NAS that performs well for file storage may struggle when it generates thumbnails, indexes a large photo library and transcodes video at the same time.
Include concurrent-user testing if the NAS supports a household, studio or small business. Two people editing files, a security camera writing recordings and a computer running Synology Drive can create a more meaningful workload than a single uninterrupted copy. If the NAS is used by staff in Adelaide or Canberra, run tests during a normal work pattern rather than assuming a single-user result represents the whole environment.
Interpret Results And Track Changes
Store benchmark results in a spreadsheet or plain-text log with the date, test file set, client device, network path and NAS workload. Include firmware changes, drive replacements, RAM upgrades and configuration changes. A baseline is valuable only when you can identify what changed between two measurements.
Look for patterns rather than isolated numbers. A drop in sequential speed alongside high disk utilisation may indicate a busy service, rebuilding array or failing drive. High CPU with modest disk activity may point to encryption, compression, antivirus scanning or video transcoding. Low throughput with low NAS utilisation often suggests a client, cable, switch or network negotiation issue.
SSD caching should be evaluated carefully. Run tests with cache disabled if possible, then repeat the same workloads after the cache has warmed. Compare small-file and repeated-read performance rather than relying on one headline transfer speed. Caching may improve frequently accessed data while leaving large sequential backups largely unchanged.
Practical Benchmarking Recommendations
Use these habits to make your Synology NAS performance baseline repeatable and useful:
- Test over wired Ethernet first, recording the negotiated link speed at every connection point.
- Run three or more repetitions and keep the median, highest and lowest results.
- Separate synthetic tests from real workloads such as backups, media serving and file collaboration.
- Monitor CPU, memory, disk activity, temperature and network traffic during every run.
- Keep a change log for DSM updates, drive replacements, cache settings and newly enabled packages.
- Repeat the baseline after major maintenance and before deciding that new hardware is necessary.
A sensible baseline also includes reliability observations. Record drive temperatures, S.M.A.R.T. warnings, storage pool status and unexpected disconnections. In regions affected by summer heat, including parts of Western Australia and Queensland, compare temperatures during warmer conditions and confirm that fans and ventilation remain clear. A performance result is less useful if the NAS cannot sustain it safely.
Use the baseline when planning upgrades. If a 1GbE link is already limiting large-file transfers, adding faster SSDs may produce little visible benefit until the network is upgraded. If small-file latency is the issue, SSD storage, additional memory or a different application layout may help more than replacing the entire NAS. Australian pricing, GST, shipping times and warranty support should be included in the upgrade decision, particularly when comparing local stock with overseas marketplace listings.
With a documented performance reference, every future change becomes easier to judge. Run the same workload after adding RAM, switching from HDDs to SSDs, enabling Btrfs features or moving services into containers, then compare the evidence rather than relying on perception. Start with a quiet, repeatable test, preserve the results, and use that record to keep your Synology NAS fast, stable and appropriately configured.