As the title says, last year’s 2.5G RN DC02 plan is going to be let go, so I bought the 11.11 plans and tested them 
For RN Seattle, see this article:[Japan-USGSL] RackNerd Seattle Black Friday plan detailed review
1. Configuration
| Plan |
🇺🇸 RackNerd DC02 |
🇺🇸 RackNerd DC03 |
🇺🇸 RackNerd SJC |
| Location |
United States, Los Angeles |
United States, Los Angeles |
United States, San Jose |
| CPU |
2vCPU |
1vCPU |
1vCPU |
| RAM |
2GB |
1GB |
1GB |
| Disk |
35GB |
20GB |
20GB |
| Traffic |
10TB(in+out) |
6TB(in+out) |
6TB(in+out) |
| Speed |
1Gbps |
1Gbps |
1Gbps |
| Network |
1 x IPv4, 1 x /64 IPv6 (optional) |
1 x IPv4, 1 x /64 IPv6 (optional) |
1 x IPv4 |
| Price |
YearlyUSD$17.66 |
YearlyUSD$10.76 |
YearlyUSD$10.76 |
Purchase link: https://www.racknerd.com/BlackFriday (no affiliate; the 11.11 promo has ended, and Black Friday configurations differ)
[Test before buying] Official Looking Glass : DC02 | DC03 | SJC
2. Node Quality
bash <(curl -sL https://run.NodeQuality.com)
💻 Basic Information
🎬 IP Quality
Google access on DC03 and SJC has been ruined by neighboring users, so services like Gemini cannot be used. You can route that traffic through WARP, for example this tutorial by forum member @DuolaD this tutorial; or buy DNS unlock / use an exit node to solve it
| 🇺🇸 DC02 - 74 range |
🇺🇸 DC02 - 148 range |
🇺🇸 DC03 |
🇺🇸 SJC |
 |
 |
 |
 |
🌐 Network Quality
ordinary direct routes for all three Chinese networks
📍 Return Routes
3. Streaming Unlock
bash <(curl -L -s check.unlock.media)
DC02 74 range currently unlocks the most streaming services
| 🇺🇸 DC02 - 74 range |
🇺🇸 DC02 - 148 range |
🇺🇸 DC03 |
🇺🇸 SJC |
 |
 |
 |
 |
4. Major website & DNS latency tests
bash <(wget -qO- https://raw.githubusercontent.com/Cd1s/network-latency-tester/main/latency.sh)
Real connection latency to popular websites
Common DNS resolution speed
5. Latency Test (itdog.cn, 2025-11-07 23:00)
All three machines show jitter and packet loss on all three Chinese networks. DC03 Unicom has the worst jitter and packet loss; the other two are okay. DC02 Mobile has relatively high latency in some regions
6.Proxy Speed Test
During evening peak, all three machines perform just average because these are direct, unoptimized routes. Unicom has the best throughput, while Telecom and Mobile have a hard time ramping up. DC02 Unicom speeds are higher than the other two.
*Click a node name to view the speed-test screenshot
🇺🇸 DC02
🇺🇸 DC03
🇺🇸 SJC
7. International speed test - iPerf3 TCP single-thread (single 10-second test, 2025-11-07 18:00-19:00)
Internationally, DC02 single-thread speed is slightly lower than the other two
🇺🇸 DC02
| Region |
Node |
Average/max upload(Mbps) |
Average/max download(Mbps) |
Latency(ms) |
| 🇺🇸 Los Angeles |
DMIT T1 |
825 / 832 |
780 / 844 |
0.72 |
| 🇺🇸 Los Angeles |
GreenCloud |
828 / 835 |
811 / 835 |
0.84 |
| 🇺🇸 San Jose |
HostDZire |
827 / 836 |
805 / 878 |
12.40 |
| 🇺🇸 Seattle |
Misaka |
817 / 836 |
783 / 854 |
27.13 |
| 🇺🇸 Salt Lake City |
BAGE |
828 / 836 |
791 / 843 |
13.99 |
| 🇺🇸 New York |
Fourplex |
746 / 886 |
481 / 747 |
67.02 |
| 🇯🇵 Tokyo |
DMIT T1 |
760 / 837 |
577 / 782 |
98.92 |
| 🇯🇵 Tokyo |
BAGE JP (GSL) |
758 / 837 |
528 / 705 |
98.24 |
| 🇯🇵 Tokyo |
BAGE JPS (CTC) |
753 / 837 |
564 / 732 |
98.22 |
| 🇭🇰 Hong Kong |
DMIT T1 |
717 / 836 |
428 / 595 |
143.32 |
| 🇭🇰 Hong Kong |
BAGE HKS (CTC) |
539 / 827 |
318 / 598 |
152.60 |
| 🇭🇰 Hong Kong |
BAGE HKL |
623 / 838 |
298 / 499 |
150.68 |
| 🇸🇬 Singapore |
BAGE SG (GSL) |
695 / 836 |
360 / 586 |
164.47 |
| 🇸🇬 Singapore |
BAGE SGS (CTC) |
691 / 838 |
363 / 556 |
176.89 |
🇺🇸 DC03
| Region |
Node |
Average/max upload(Mbps) |
Average/max download(Mbps) |
Latency(ms) |
| 🇺🇸 Los Angeles |
DMIT T1 |
844 / 900 |
490 / 510 |
0.51 |
| 🇺🇸 Los Angeles |
GreenCloud |
808 / 857 |
889 / 907 |
1.10 |
| 🇺🇸 San Jose |
HostDZire |
804 / 902 |
849 / 891 |
12.53 |
| 🇺🇸 Seattle |
Misaka |
776 / 948 |
843 / 950 |
27.55 |
| 🇺🇸 Salt Lake City |
BAGE |
768 / 868 |
850 / 943 |
30.89 |
| 🇺🇸 New York |
Fourplex |
527 / 673 |
734 / 937 |
67.28 |
| 🇯🇵 Tokyo |
DMIT T1 |
579 / 787 |
623 / 809 |
99.26 |
| 🇯🇵 Tokyo |
BAGE JP (GSL) |
588 / 681 |
661 / 889 |
98.42 |
| 🇯🇵 Tokyo |
BAGE JPS (CTC) |
516 / 639 |
612 / 753 |
98.37 |
| 🇭🇰 Hong Kong |
DMIT T1 |
549 / 864 |
487 / 725 |
143.64 |
| 🇭🇰 Hong Kong |
BAGE HKS (CTC) |
610 / 781 |
461 / 642 |
164.53 |
| 🇭🇰 Hong Kong |
BAGE HKL |
554 / 875 |
541 / 833 |
151.05 |
| 🇸🇬 Singapore |
BAGE SG (GSL) |
423 / 601 |
422 / 748 |
164.28 |
| 🇸🇬 Singapore |
BAGE SGS (CTC) |
499 / 833 |
411 / 642 |
174.86 |
🇺🇸 SJC
| Region |
Node |
Average/max upload(Mbps) |
Average/max download(Mbps) |
Latency(ms) |
| 🇺🇸 Los Angeles |
DMIT T1 |
905 / 940 |
547 / 583 |
12.13 |
| 🇺🇸 Los Angeles |
GreenCloud |
905 / 926 |
909 / 946 |
12.28 |
| 🇺🇸 San Jose |
HostDZire |
897 / 916 |
894 / 919 |
1.67 |
| 🇺🇸 Seattle |
Misaka |
894 / 946 |
867 / 941 |
18.70 |
| 🇺🇸 Salt Lake City |
BAGE |
869 / 907 |
880 / 957 |
20.08 |
| 🇺🇸 New York |
Fourplex |
691 / 824 |
679 / 838 |
61.99 |
| 🇯🇵 Tokyo |
DMIT T1 |
682 / 836 |
640 / 848 |
110.36 |
| 🇯🇵 Tokyo |
BAGE JP (GSL) |
663 / 856 |
645 / 854 |
109.41 |
| 🇯🇵 Tokyo |
BAGE JPS (CTC) |
726 / 909 |
554 / 730 |
109.49 |
| 🇭🇰 Hong Kong |
DMIT T1 |
595 / 869 |
534 / 926 |
154.69 |
| 🇭🇰 Hong Kong |
BAGE HKS (CTC) |
678 / 850 |
530 / 717 |
162.14 |
| 🇭🇰 Hong Kong |
BAGE HKL |
576 / 725 |
393 / 678 |
159.96 |
| 🇸🇬 Singapore |
BAGE SG (GSL) |
585 / 830 |
392 / 630 |
175.71 |
| 🇸🇬 Singapore |
BAGE SGS (CTC) |
646 / 782 |
487 / 741 |
175.47 |
8. Summary
For pure proxy use, there is no need to blindly follow the hype and buy these, because none of the three machines is optimized and evening-peak performance may disappoint. Among the three Chinese networks, Unicom direct performs best; among the three machines, DC02 is the best overall for direct connection.
9. About The Tests
All machines used in this article (including all machines participating in iPerf3 tests) were purchased at my own expense; results may be time-sensitive
Kernel parameters of the test 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 Speedtest"(@Kyoma) for providing public-benefit speed testing services