UGREEN NAS Speed Test: Read, Write, and Network

A network-only benchmark shows how fast data can move between your computer and UGREEN NAS without testing the storage pool. A shared-folder benchmark then measures the complete path through SMB, the NAS processor, RAID, and the installed drives. Comparing those results tells you which layer is limiting performance before you change hardware or settings.

If you already know that normal file transfers are slow and need symptom-by-symptom troubleshooting, use our guide to fix slow UGREEN NAS transfer speeds. This guide focuses specifically on measuring performance and establishing a repeatable baseline.

How Do You Test the Network Without the Drives?

The clearest starting point is a local OpenSpeedTest container. It sends test data between a browser and the NAS without benchmarking the HDD or SSD storage pool.

Install OpenSpeedTest on UGREEN NAS

This method requires a UGREEN NAS model that supports Docker. The DH2300 cannot host the container because it does not support Docker.

1. Install Docker from the UGREEN App Center.
2. Open Docker.
3. Select Project and create a new project.
4. Name the project openspeedtest.
5. Paste the following Compose configuration:

services:
  openspeedtest:
    image: openspeedtest/latest:v2.0.6
    container_name: openspeedtest
    restart: unless-stopped
    ports:
      - "3000:3000"

The version is pinned so a future image update does not silently change the benchmark environment. The configuration follows the port and restart requirements published with the OpenSpeedTest Docker image.

6. Deploy the project.
7. Wait until the container status changes to Running.
8. From a wired computer on the same network, open:

http://NAS-IP:3000

Replace NAS-IP with the local IP address assigned to the NAS interface you want to test.

Do not forward port 3000 through the router or expose this container to the internet. It is intended to measure the local network.

Read the OpenSpeedTest Results

OpenSpeedTest displays the result from the browser’s point of view:

• Download: Data travels from the NAS container to the computer.
• Upload: Data travels from the computer to the NAS container.
• Ping: Round-trip latency between the computer and the container.
• Jitter: Variation in latency during the test.

Run three tests and record the median upload and download values. Large differences between directions can indicate a client adapter, processor, driver, or connection-path limitation, but this test does not identify the exact cause.

A browser-based test can also become limited by the browser or processor before a fast 10GbE connection reaches its ceiling. Advanced users can confirm high-speed results with iperf3, which is designed specifically to measure achievable IP-network bandwidth.

Once you have the network result, leave the connection unchanged. The next test adds SMB and the storage pool to the same path.

How Do You Benchmark a UGREEN NAS Shared Folder?

A mapped-share benchmark measures more than the drives. Its result includes:

• Client storage and processor
• Network connection
• SMB processing
• NAS processor
• Memory caching
• RAID calculations
• Filesystem
• HDD or SSD pool

That makes it the correct companion to the network-only result.

Test a UGREEN NAS Share on Windows

Use CrystalDiskMark against a mapped network drive.

1. Map the benchmark shared folder as a Windows drive.
2. Open CrystalDiskMark as a standard user.
3. Select the mapped NAS drive as the test destination.
4. Set the number of runs to 3.
5. Select a test size of at least 16GiB. Use 32GiB or 64GiB when the NAS has more memory and the shared folder has sufficient free space.
6. Run the sequential 1MiB, queue-depth-one test first.
7. Run the 4KiB random, queue-depth-one test only if your workload includes databases, application data, virtual machines, or large numbers of small files.
8. Record the read and write results separately.

CrystalDiskMark may hide mapped network drives when launched with administrator privileges, so open it normally. Its developer also notes that test size, fragmentation, controller, processor, and other variables affect the result.

The sequential queue-depth-one figure is the closest synthetic comparison to one large file moving through a single SMB connection. Higher queue depths measure a more parallel workload and may produce a larger number than ordinary file copying.

Test a UGREEN NAS Share on macOS

Use Blackmagic Disk Speed Test for a sequential benchmark.

1. Connect to the UGREEN NAS shared folder through SMB.
2. Open Blackmagic Disk Speed Test.
3. Select the settings icon.
4. Choose Select Target Drive.
5. Select the mounted NAS share.
6. Set the stress size to 5GB.
7. Start the test and allow at least three complete read-and-write cycles.
8. Stop the test and record the stable results.

Blackmagic Disk Speed Test is built around large sequential media workloads. It does not show how the storage pool handles thousands of small files.

Its maximum 5GB stress file can also fit inside memory on a higher-RAM NAS, so treat the result as short-duration sequential performance. Confirm sustained performance with a representative file from your normal workload.

Synthetic benchmarks make comparison easier, but they do not replace a practical transfer. Use the controlled real-file procedure in the slow-transfer guide when you need to confirm performance with actual data.

Which Improvement Matches the Measured Limit?

Once the tests identify the limiting layer, the improvement becomes much narrower.

The Network Test Reaches Its Ceiling

Higher throughput requires a faster end-to-end path, including the NAS port, computer adapter, and any switch between them.

A faster NAS port alone does not increase speed when the client or switch remains on the previous standard.

Sequential Storage Is Below the Network Result

When network-only performance is strong but sequential read or write speed is lower, examine the storage configuration.

Possible improvements include:

• Adding active drives where the enclosure and RAID configuration support it
• Choosing a RAID level that better matches the capacity, write-performance, and failure-tolerance requirements
• Moving active work to a compatible SSD storage pool
• Separating application data from a heavily used HDD archive

Review the supported paths in our guide to choosing the best RAID for UGREEN NAS before changing an existing pool. A RAID change is a storage decision, not a benchmark setting.

Small Random Access Is the Limit

An SSD storage pool provides low latency for virtual machines, databases, containers, and active project data.

SSD cache serves a different role. It is most useful when the same small blocks are accessed repeatedly. It offers limited value for a one-time sequential copy of a large video or backup archive. Our guide to using HDDs and SSDs together in UGREEN NAS explains when to use cache and when to create a separate SSD pool.

CPU or Memory Usage Reaches Its Limit

Watch Task Manager while the shared-folder benchmark runs. High CPU usage with unused network and storage capacity can point to SMB processing, security features, parity work, or another active application.

Memory Compression belongs only in a measured memory-pressure scenario. In UGOS Pro, it is available under Control Panel > Hardware & Power > General. Compression can reduce swapping when memory runs short, but it also consumes CPU resources. Enabling it on a lightly loaded NAS does not increase disk or network throughput.

Use the broader UGREEN NAS performance troubleshooting guide when the interface, applications, containers, and transfers slow down together.

The Existing Hardware Reaches Every Expected Limit

A confirmed ceiling is different from an unexplained slowdown. If the network, storage, and client all reach their expected limits and the workload still needs more performance, compare models by:

• Network-port speed
• Number of drive bays
• M.2 support
• Memory capacity
• Processor
• Thunderbolt support
• SSD storage-pool support

At that point, you can browse and filter UGREEN NAS models against a measured requirement instead of buying an upgrade based on a theoretical specification.

Change one component or setting at a time and repeat the same benchmark. If RAID, MTU, caching, and network hardware all change together, the new result cannot show which change mattered.

Frequently Asked Questions

Can the DH2300 run OpenSpeedTest?

No. The DH2300 does not support Docker, so it cannot host the OpenSpeedTest container. Use a mapped-share benchmark and a controlled real-file transfer to measure the complete storage path.

Why is OpenSpeedTest faster than CrystalDiskMark?

OpenSpeedTest measures the local network path without benchmarking the storage pool. CrystalDiskMark includes SMB, the NAS processor, RAID, filesystem, drives, and the client, so its result is limited by more components.

What CrystalDiskMark test size should I use for a NAS?

Start with at least 16GiB. Use 32GiB or 64GiB on a NAS with more memory when sufficient free space is available. A larger test reduces the chance that the result represents only temporary RAM caching.

Does SSD cache improve NAS benchmark results?

It can improve repeated access to frequently used small blocks. It normally provides little benefit to a first-time sequential transfer of a large file. Use an SSD storage pool when the workload needs consistently low latency or high random IOPS.