Testing the RFCHOST JP T1 series in two different data centers. This product is positioned as an exit node + DNS unlock + international transit, with a bit of direct-connect capability as well. I borrowed JP2 from forum member @Hconps for the test; there is not a huge difference, so just treat this as a JP2 review 
| Plan |
🇯🇵 RFCHOST JP2-T1ION-Balance |
🇯🇵 RFCHOST JP1-T1ION-Balance |
| Location |
Japan, Tokyo |
Japan, Tokyo |
| Hardware |
1vCPU, 0.5GB RAM, 10 GB SSD |
1vCPU, 0.5GB RAM, 10 GB SSD |
| Traffic |
1TB |
1TB |
| Route |
International route, no mainland China optimization |
International route, no mainland China optimization |
| Network |
1 x IPv4 |
1 x IPv4 |
| Cost |
Yearly USD$ 29.90 |
Yearly USD$ 29.90 |
💻 Basic Information
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🎬 IP Quality
DNS unlock was not enabled during testing
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🔗Route Information
Compared with JP1, JP2 has additional RETN and Telstra connectivity, but traffic from JP2 to the EU currently does not use RETN
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🌐 Network Quality
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
📍 Return Routes
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
📡TCP-Ping Test(itdog.cn, 22:00)
The packet loss is high; I suspect TCP-Ping was treated as DDoS traffic and scrubbed by mitigation
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
📡ICMP-Ping Test(itdog.cn, 23:00)
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
⏬ iPerf3 TCP single-thread speed test
🇨🇳 Guangdong Mobile Yunxi Huanjing @2000Mbps (2025-08-16 20:00)
China Mobile return goes via Lumen
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇨🇳 Shandong Unicom Yunxi Huanjing @2000Mbps (2025-08-16 20:00)
China Unicom return goes via SoftBank
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇨🇳 Ningbo Telecom Seagull Cloud @800Mbps (2025-08-16 20:00)
China Telecom return goes via IIJ
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇭🇰 Hong Kong RFCHOST JINX (2025-08-16 20:00)
Testing Hong Kong T1 pulling from Japan T1. At the moment, JP2 to HK JINX and CTC seems slower than JP1
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇭🇰 Hong Kong RFCHOST CTC (2025-08-16 20:00)
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇺🇸 Los Angeles RFCHOST T1 (2025-08-16 20:00)
Japan T1 pulling from US T1. During the test, US back to JP1 used GSL, with lower latency but lower throughput; back to JP2 used NTT, with slightly higher latency but better throughput
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇩🇪 Frankfurt BAGE @5Gbps (2025-08-16 20:00)
Germany currently returns to Japan via the US over GSL
| 🇯🇵 RFC JP2 T1 |
🇯🇵 RFC JP1 T1 |
 |
 |
🇯🇵 JP1 and JP2 mutual test
Both directions can reach 8Gbps+
| JP2 sending to JP1 |
JP1 sending to JP2 |
 |
 |
🗒️ Summary
After all, these are just different data centers, so there is basically not much difference. JP1 T1 will probably be migrated to JP2 eventually anyway, so treat this test as an archive
🔗 About The Tests
Thanks to forum member @Hconps for providing the JP2 machine for testing. Aside from that machine, all machines used in this article were purchased at my own expense; TCP tuning was limited to enabling BBR+FQ, with all other parameters left at defaults; results may be time-sensitive
🇯🇵 RFCHOST JP2-T1ION-Balance NodeQuality Report
🇯🇵 RFCHOST JP1-T1ION-Balance NodeQuality Report