[Preview test] VPS.Town A4 Hong Kong new storage plan, performance / private network / international connectivity detailed test

I previously bought VPS.Town's B1. Back then I heard their large-storage plans offered good value, and this time I had a chance to join the internal beta, so of course I had to test it properly.

International connectivity is in section 5.

1. Configuration

Pricing: 2C4G20G is 29 CNY/month, 4C8G20G is 49 CNY/month; HDD data disk is a one-time purchase at 130 CNY/TB; SSD system disk is 0.1 CNY/GB/month.
A4 has already launched. See the provider post for more promotion details:https://www.nodeseek.com/post-514773-1

🔗 Quick links

2. Performance tests

2.1 Geekbench 6 30-round loop test

Test tool: Geekbench 6.5; only the GB6 total score and selected subtests were used for data processing

Under high load, performance fluctuations are fairly stable. There are 1-3 drops, but the magnitude is not especially large. The retail version is expected to be system-limited after rounds 9-10, while the performance jitter in this loop test appeared after round 10.

This is an internal beta environment. On the retail version, do not sustain high usage for more than 60 minutes; otherwise, according to the official FAQ, the CPU will be limited to 70% of its original performance. Malicious long-term sustained high usage may result in penalties or termination.

Test dataset CSV

2.2 YABS fio 100-round loop test

Test tool: YABS.sh fio section, 50/50 mixed read/write

2.2.1 HDD Data disk

• Bandwidth

• IOPS

• Data

+------------+--------+-----------+--------------------+----------+------------------+
| Block Size | Metric | Average MB/s | Std Dev MB/s | Average IOPS | Std Dev IOPS |
+------------+--------+-----------+--------------------+----------+------------------+
| 4k         | Read   |   16.5077 |  0.02019376        |  4099    | 10.00000         |
| 4k         | Write  |   16.5246 |  0.02133996        |  4099    | 10.00000         |
| 4k         | Total  |   33.0352 |  0.04137730        |  8199    | 10.00000         |
| 64k        | Read   |  265.5886 |  0.22039001        |  4100    |  0.00000         |
| 64k        | Write  |  266.9865 |  0.22211734        |  4100    |  0.00000         |
| 64k        | Total  |  532.5794 |  0.44312668        |  8300    |  0.00000         |
| 512k       | Read   | 2778.8288 |  5.37827754        |  5300    |  0.00000         |
| 512k       | Write  | 2929.0496 |  2.01673099        |  5509    | 28.76235         |
| 512k       | Total  | 5708.3904 |  7.33812542        | 10835    | 47.93725         |
| 1m         | Read   | 1500.9792 | 24.23138860        |  1397    | 22.27015         |
| 1m         | Write  | 1601.4336 | 26.35510521        |  1497    | 22.27015         |
| 1m         | Total  | 3108.3520 | 49.36476509        |  2897    | 48.10552         |
+------------+--------+-----------+--------------------+----------+------------------+

2.2.2 SSD System disk

• Bandwidth

• IOPS

• Data

+------------+--------+-----------+------------------+----------+------------------+
| Block Size | Metric | Average MB/s | Std Dev MB/s | Average IOPS | Std Dev IOPS |
+------------+--------+-----------+------------------+----------+------------------+
| 4k         | Read   |  172.1076 |   1.360296       | 42985    | 336.76266        |
| 4k         | Write  |  172.5614 |   1.363783       | 43090    | 337.39944        |
| 4k         | Total  |  344.6724 |   2.724133       | 86125    | 680.66738        |
| 64k        | Read   | 1687.8592 |  27.617336       | 25774    | 418.90188        |
| 64k        | Write  | 1696.4608 |  27.770318       | 25918    | 426.96486        |
| 64k        | Total  | 3388.9280 |  55.562658       | 51741    | 846.99051        |
| 512k       | Read   | 2536.9600 |  14.362204       |  4798    |  20.00000        |
| 512k       | Write  | 2672.8448 |  15.332738       |  5057    |  49.75699        |
| 512k       | Total  | 5215.6416 |  28.603208       |  9912    |  64.00758        |
| 1m         | Read   | 1500.3648 | 113.202594       |  1367    | 115.51816        |
| 1m         | Write  | 1601.2288 | 120.930792       |  1467    | 115.51816        |
| 1m         | Total  | 3105.6896 | 233.663434       |  2926    | 219.55879        |
+------------+--------+-----------+------------------+----------+------------------+

3. Nodequality

🎬IP quality

🌐Network quality

During the internal beta, the Lumen outbound route was not yet direct, so latency is higher. You can refer to recent NQ tests on the forum, such as [NodeQuality] VPS.town HKG.A4 freshly released

📍Return route

4. Internal network bandwidth test

Test tool: iPerf3
Test target: two VPS.Town A4 instances

Internal network traffic is not counted. It is not currently rate-limited and performs very well. After official launch, it will be limited to 5Gbps, while cross-host internal networking will also be enabled.

4.1 Single-thread (30 seconds per run)

Test records

4.2 Multi-thread (30 seconds per run, 8 threads)

Test records

5. International connectivity - iPerf3 TCP IPv4 Single-threadspeed test (10 seconds per run, 2025-11-19 21:00-21:30)

Test tool: iPerf3

Local bandwidth can reach 4Gbps+, and international connectivity is also still good.

For Mainland China nodes, you should focus on average/max upload (Mbps), because the mainland side is downloading resources from the remote server.

Test records

I have not yet measured a route to Europe that uses RETN (using RETN would not necessarily mean lower latency), but the route to Netcup Nuremberg looks similar in latency to another Hong Kong machine that does use RETN. The route to Misaka Moscow did not use RETN; if it had, latency would not be above 120ms.

Update: according to @Wenxin's reply, the route to Netcup Nuremberg should be using RETN.

VPS.Town to Netcup	VPS.Town to Misaka	RETN to Netcup	RETN to Misaka

6. Intra-Hong Kong connectivity - iPerf3 TCP IPv4 single-thread loop speed test (30 seconds per run, 10 tests in each direction, 2025-11-19 14:30-15:00)

Test tool: iPerf3
Download test: GreenCloud HKG.DC1, peak bandwidth 10Gbps
Upload test: GreenCloud HKG.DC2, peak bandwidth 10Gbps

As you can see, 300-second intra-Hong Kong download/upload tests can still maintain 4Gbps+.

300-second intra-Hong Kong download

Test records

300-second intra-Hong Kong upload

Test records

7. Speedtest Multi-threadspeed test

curl -sL nws.sh | bash

---------------------------------- nws.sh ---------------------------------
      A simple script to bench network performance using speedtest-cli     
---------------------------------------------------------------------------
 Speedtest.net (Region: GLOBAL)
---------------------------------------------------------------------------
 Location         Latency     Loss    DL Speed       UP Speed       Server      

 ISP: Suzhou Yesong Information Technology Co. 

 Nearest          0.39 ms     0.0%    4470.85 Mbps   4767.83 Mbps   fdcservers.net - Hong Kong 

 Bangalore, IN    69.37 ms    0.0%    3377.51 Mbps   1128.46 Mbps   Bharti Airtel Ltd - Bangalore 
 Chennai, IN      76.71 ms    0.0%    3190.55 Mbps   1284.32 Mbps   RailTel Corporation of India Ltd - Chennai 
 Mumbai, IN       103.97 ms   0.0%    4345.25 Mbps   3044.31 Mbps   Melbicom - Mumbai 

 Seattle, US      142.36 ms   N/A     922.61 Mbps    630.18 Mbps    Comcast - Seattle, WA 
 Los Angeles, US  147.30 ms   0.0%    3417.31 Mbps   343.14 Mbps    ReliableSite Hosting - Los Angeles, CA 
 Dallas, US       183.84 ms   0.0%    2427.70 Mbps   1801.37 Mbps   Hivelocity - Dallas, TX 
 Miami, US        200.81 ms   0.0%    2832.97 Mbps   1620.87 Mbps   Frontier - Miami, FL 
 New York, US     181.89 ms   0.0%    4487.40 Mbps   1840.18 Mbps   GSL Networks - New York, NY 
 Toronto, CA      197.43 ms   0.0%    937.22 Mbps    448.84 Mbps    Rogers - Toronto, ON 
 Mexico City, MX  203.64 ms   N/A     1014.94 Mbps   434.66 Mbps    INFINITUM - Ciudad de México 

 London, UK       185.19 ms   0.0%    4191.37 Mbps   1777.03 Mbps   VeloxServ Communications - London 
 Amsterdam, NL    174.10 ms   0.0%    4494.64 Mbps   1174.34 Mbps   31173 Services AB - Amsterdam 
 Paris, FR        186.58 ms   0.0%    4404.89 Mbps   1954.57 Mbps   Scaleway - Paris 
 Frankfurt, DE    180.02 ms   0.0%    2608.26 Mbps   511.20 Mbps    Clouvider Ltd - Frankfurt am Main 
 Warsaw, PL       249.16 ms   0.0%    3598.40 Mbps   1387.64 Mbps   Play - Warszawa 
 Bucharest, RO    212.93 ms   0.0%    3336.44 Mbps   1689.51 Mbps   Digi        
 Moscow, RU       164.92 ms   0.0%    4233.79 Mbps   1852.37 Mbps   Misaka Network, Inc. - Moscow 

 Jeddah, SA       138.92 ms   0.0%    4217.79 Mbps   2090.40 Mbps   Saudi Telecom Company 
 Dubai, AE        121.13 ms   N/A     4131.58 Mbps   3075.98 Mbps   e& UAE - Dubai 
 Istanbul, TR     220.44 ms   0.0%    3988.27 Mbps   1608.48 Mbps   Turkcell - Istanbul 
 Tehran, IR       248.74 ms   0.0%    1261.46 Mbps   349.34 Mbps    Irancell - Tehran 
 Cairo, EG        209.95 ms   0.0%    1078.43 Mbps   417.90 Mbps    Telecom Egypt - Cairo 

 Tokyo, JP        53.82 ms    67.8%   2229.15 Mbps   796.44 Mbps    fdcservers.net - Tokyo 
 Shanghai, CU-CN  290.92 ms   N/A     77.75 Mbps     858.35 Mbps    China Unicom 5G - Shanghai 
 Hong Kong, CN    0.61 ms     0.0%    4471.26 Mbps   1197.08 Mbps   Misaka Network, Inc. - Hong Kong 
 Singapore, SG    39.54 ms    0.0%    3809.42 Mbps   2528.53 Mbps   ViewQwest - Singapore 
 Jakarta, ID      105.07 ms   3.0%    1379.82 Mbps   584.94 Mbps    PT Solnet Indonesia - Jakarta 

 Sydney, AU       129.41 ms   0.0%    4388.97 Mbps   2515.04 Mbps   Aussie Broadband - Sydney 
---------------------------------------------------------------------------
 Avg DL Speed       : 3080.21 Mbps
 Avg UL Speed       : 1507.35 Mbps

 Total DL Data      : 120.50 GB
 Total UL Data      : 56.43 GB
 Total Data         : 176.93 GB
---------------------------------------------------------------------------
 Duration           : 14 min 35 sec
 System Time        : 18/11/2025 - 07:50:11 UTC
 Total Script Runs  : 508987
---------------------------------------------------------------------------
 Result             : https://result.nws.sh/r/1763451294_D29DHH_GLOBAL.txt
---------------------------------------------------------------------------

8. Summary

If A4 launches at the expected price, it will be a very cost-effective choice: large port capacity, cheap storage, and good performance. China Mobile can use Lumen direct routing, which is a plus for a large-storage plan. However, after official launch, the actual China Mobile direct-route effect will still depend on region and congestion.

9. About the tests

”VPS.Town A4 2C4G40G” and ”VPS.Town A4 2C4G40G + 1TB HDD” are internal beta machines entrusted by the provider for testing, and are not guaranteed to match the retail-version experience. The other machines tested in this article, including all machines used for iPerf3 speed tests, were purchased at my own expense. Results may be time-sensitive.
Kernel parameters of the tested machines (/etc/sysctl.conf) are as follows

fs.file-max = 6815744
net.ipv4.tcp_no_metrics_save=1
net.ipv4.tcp_ecn=0
net.ipv4.tcp_frto=0
net.ipv4.tcp_mtu_probing=0
net.ipv4.tcp_rfc1337=0
net.ipv4.tcp_sack=1
net.ipv4.tcp_fack=1
net.ipv4.tcp_window_scaling=1
net.ipv4.tcp_adv_win_scale=1
net.ipv4.tcp_moderate_rcvbuf=1
net.core.rmem_max=33554432
net.core.wmem_max=33554432
net.ipv4.tcp_rmem=4096 87380 33554432
net.ipv4.tcp_wmem=4096 16384 33554432
net.ipv4.udp_rmem_min=8192
net.ipv4.udp_wmem_min=8192
net.ipv4.ip_forward=1
net.ipv4.conf.all.route_localnet=1
net.ipv4.conf.all.forwarding=1
net.ipv4.conf.default.forwarding=1
net.core.default_qdisc=fq
net.ipv4.tcp_congestion_control=bbr
net.ipv6.conf.all.forwarding=1
net.ipv6.conf.default.forwarding=1

10. Acknowledgements

Thanks to VPS.Town for providing two machines for testing.