[Japan/US GSL] RackNerd Seattle Black Friday plan, evening peak speed tests and international connectivity, detailed test

Last time I tested RackNerd DC02/DC03/SJC，
; this time I also tested Seattle, so now I have tested all of RN's West Coast VPS nodes.

1. Configuration

RackNerd Black Friday event link (no AFF): https://www.racknerd.com/BlackFriday
Looking Glass: https://lg-sea.racknerd.com

2. Node Quality

bash <(curl -sL https://run.NodeQuality.com)

💻Basic information

The CPU is just average.

Basic System Information:
---------------------------------
Uptime     : 0 days, 2 hours, 31 minutes
Processor  : Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
CPU cores  : 2 @ 2799.998 MHz
AES-NI     : ✔ Enabled
VM-x/AMD-V : ❌ Disabled
RAM        : 2.4 GiB
Swap       : 0.0 KiB
Disk       : 46.0 GiB
Distro     : Debian GNU/Linux 12 (bookworm)
Kernel     : 6.12.57+deb13-cloud-amd64
VM Type    : KVM
IPv4/IPv6  : ✔ Online / ❌ Offline

IPv4 Network Information:
---------------------------------
ISP        : HostPapa
ASN        : AS36352 HostPapa
Host       : RackNerd LLC
Location   : Seattle, Washington (WA)
Country    : United States

fio Disk Speed Tests (Mixed R/W 50/50) (Partition -):
---------------------------------
Block Size | 4k            (IOPS) | 64k           (IOPS)
  ------   | ---            ----  | ----           ---- 
Read       | 84.43 MB/s   (21.1k) | 472.90 MB/s   (7.3k)
Write      | 84.65 MB/s   (21.1k) | 475.39 MB/s   (7.4k)
Total      | 169.09 MB/s  (42.2k) | 948.29 MB/s  (14.8k)
           |                      |                     
Block Size | 512k          (IOPS) | 1m            (IOPS)
  ------   | ---            ----  | ----           ---- 
Read       | 702.74 MB/s   (1.3k) | 638.63 MB/s    (623)
Write      | 740.08 MB/s   (1.4k) | 681.16 MB/s    (665)
Total      | 1.44 GB/s     (2.8k) | 1.31 GB/s     (1.2k)

Geekbench 5 Benchmark Test:
---------------------------------
Test            | Value                         
                |                               
Single Core     | 478                           
Multi Core      | 857                           
Full Test       | https://browser.geekbench.com/v5/cpu/23913688

 SysBench CPU Test (Fast Mode, 1-Pass @ 5sec)
---------------------------------
 1 Thread Test (Single Core) Score:          770 Scores
 2 Thread Test (Multi-Core) Score:          1554 Scores
 SysBench Memory Test (Fast Mode, 1-Pass @ 5sec)
---------------------------------
 Single-thread Read Test:          15969.52 MB/s
 Single-thread Write Test:          12779.93 MB/s

🎬IP quality

🌐Network quality

Direct-connection latency is quite high.

📍Return route

NodeQuality link

3. Streaming unlock

D+ is not unlocked.

bash <(curl -L -s check.unlock.media)

International streaming	Local streaming

4. Website & DNS speed tests

Website latency feels a bit high...

bash <(wget -qO- https://raw.githubusercontent.com/Cd1s/network-latency-tester/main/latency.sh)

Website latency	DNS resolution

5. Proxy speed tests (2025-11-19 23:00-24:00)

Most Telecom and Mobile nodes cannot really run; some Unicom nodes can get going.

[Telecom speed-test records] | [Unicom speed-test records] | [Mobile speed-test records]

6. International connectivity - iPerf3 TCP IPv4 Single-threadspeed test (10 seconds per run, 2025-11-20 01:00-01:30)

[Test records]

What is GSL?

Global Secure Layer (GSL) is a tier-2 network operator [...] providing global carrier-grade internet services, using industry-leading DDoS protection technology for end-to-end security [...] focused on low latency, high resilience, and scalability... (source）

From my own use, GSL performs well in Asia-Pacific and the US West Coast, with latency kept fairly low (GSL Tokyo<->Seattle around 82ms, GSL Tokyo<->Los Angeles around 99ms).
Many Hong Kong machines connect to GSL. First, it helps with DDoS protection; second, through GSL<->Lumen<->CMI, they can obtain cheap China Mobile direct routing toward Mainland China.

Japan-US routes

DMIT and BAGE both use GSL in both directions and have the lowest latency. YxVM uses GSL outbound and NTT return. RFC outbound detours via Hong Kong and returns via NTT, giving the highest latency.

🇯🇵DMIT<->🇺🇸RackNerd	🇯🇵YxVM<->🇺🇸RackNerd	🇯🇵RFCHOST<->🇺🇸RackNerd
Outbound	Outbound	Outbound
Return	Return	Return

Hong Kong-US routes

DMIT and RN Seattle use GSL both ways via Tokyo and have the lowest latency. YxVM outbound uses RETN->Lumen and returns through GSL via Tokyo. RFC JINX outbound uses RETN->Lumen and returns via Lumen, with the highest latency.

🇭🇰DMIT<->🇺🇸RackNerd	🇭🇰YxVM<->🇺🇸RackNerd	🇭🇰RFCHOST<->🇺🇸RackNerd
Outbound	Outbound	Outbound
Return	Return	Return

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: ColoCrossing 

 Nearest          0.83 ms     0.0%    609.06 Mbps    847.28 Mbps    Astound Broadband - Seattle, WA 

 Bangalore, IN    242.54 ms   0.0%    564.90 Mbps    369.13 Mbps    Bharti Airtel Ltd - Bangalore 
 Chennai, IN      242.74 ms   0.0%    590.49 Mbps    402.84 Mbps    RailTel Corporation of India Ltd - Chennai 
 Mumbai, IN       267.61 ms   0.0%    673.18 Mbps    637.83 Mbps    Melbicom - Mumbai 

 Seattle, US      1.77 ms     N/A     862.81 Mbps    847.29 Mbps    Comcast - Seattle, WA 
 Los Angeles, US  29.33 ms    0.0%    741.68 Mbps    800.02 Mbps    ReliableSite Hosting - Los Angeles, CA 
 Dallas, US       47.27 ms    0.0%    712.74 Mbps    814.60 Mbps    Hivelocity - Dallas, TX 
 Miami, US        83.10 ms    0.0%    698.00 Mbps    829.52 Mbps    Frontier - Miami, FL 
 New York, US     55.94 ms    0.0%    926.15 Mbps    798.49 Mbps    GSL Networks - New York, NY 
 Toronto, CA      60.50 ms    0.0%    465.89 Mbps    779.45 Mbps    Rogers - Toronto, ON 
 Mexico City, MX  76.72 ms    N/A     716.90 Mbps    783.78 Mbps    INFINITUM - Ciudad de México 

 London, UK       140.29 ms   0.0%    902.84 Mbps    752.09 Mbps    VeloxServ Communications - London 
 Amsterdam, NL    143.61 ms   0.0%    777.74 Mbps    713.57 Mbps    31173 Services AB - Amsterdam 
 Paris, FR        137.46 ms   0.0%    861.03 Mbps    770.14 Mbps    Scaleway - Paris 
 Frankfurt, DE    151.89 ms   0.0%    634.35 Mbps    572.36 Mbps    Clouvider Ltd - Frankfurt am Main 
 Warsaw, PL       168.45 ms   0.0%    846.91 Mbps    714.41 Mbps    Play - Warszawa 
 Bucharest, RO    182.22 ms   0.0%    566.76 Mbps    743.76 Mbps    Digi        
 Moscow, RU       190.82 ms   0.0%    711.04 Mbps    697.20 Mbps    Misaka Network, Inc. - Moscow 

 Jeddah, SA       194.57 ms   0.0%    843.92 Mbps    718.25 Mbps    Saudi Telecom Company 
 Dubai, AE        244.09 ms   N/A     656.21 Mbps    784.60 Mbps    e& UAE - Dubai 
 Istanbul, TR     181.86 ms   0.0%    827.55 Mbps    688.08 Mbps    Turkcell - Istanbul 
 Tehran, IR       228.79 ms   0.0%    554.69 Mbps    367.02 Mbps    Irancell - Tehran 
 Cairo, EG        196.08 ms   0.3%    542.26 Mbps    399.05 Mbps    Telecom Egypt - Cairo 

 Tokyo, JP        133.52 ms   69.1%   545.08 Mbps    532.09 Mbps    fdcservers.net - Tokyo 
 Shanghai, CU-CN  216.41 ms   N/A     276.64 Mbps    664.13 Mbps    China Unicom 5G - Shanghai 
 Hong Kong, CN    133.72 ms   N/A     466.92 Mbps    404.46 Mbps    Misaka Network, Inc. - Hong Kong 
 Singapore, SG    211.37 ms   0.0%    783.74 Mbps    432.56 Mbps    ViewQwest - Singapore 
 Jakarta, ID      229.11 ms   0.0%    521.81 Mbps    393.51 Mbps    PT Solnet Indonesia - Jakarta 

 Sydney, AU       167.53 ms   0.0%    908.30 Mbps    722.55 Mbps    Aussie Broadband - Sydney 
---------------------------------------------------------------------------
 Avg DL Speed       : 682.40 Mbps
 Avg UL Speed       : 654.49 Mbps

 Total DL Data      : 27.96 GB
 Total UL Data      : 24.57 GB
 Total Data         : 52.53 GB
---------------------------------------------------------------------------
 Duration           : 14 min 23 sec
 System Time        : 20/11/2025 - 00:51:27 CST
 Total Script Runs  : 509144
---------------------------------------------------------------------------
 Result             : https://result.nws.sh/r/1763570166_3KZZX7_GLOBAL.txt
---------------------------------------------------------------------------

8. Summary

Direct connection to Seattle detours via Los Angeles and has no optimization, so direct performance is worse than DC02. I do not recommend direct use.

It is a fairly affordable Japan-US GSL peer (the lowest plan is around $10) and can be used with Shanghai-Japan private lines or Japan-optimized machines to reduce latency to the United States.

Update: however, someone in the comments also pointed out that RN performs poorly to major providers. I compared it with Misaka. The comparison is revealing: both are in Seattle, but RN's latency to mainstream websites is still much higher than Misaka's. So for this RN model, if used as a proxy, the latency saved from Tokyo to Seattle is given back in website latency.

9. About the tests

The 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 “Impart Speed Test“(@Kyoma) for providing public-interest speed-test service.