Recently I saw that Japan three-network 4837 seemed to have returned, so I bought one to test what is going on. 
Jtti, whose domain was registered on 2022-03-14, previously sold small-bandwidth optimized-route machines and was not very well known. Recently it launched HK&JP return-to-China optimized products advertising "dedicated" 200Mbps bandwidth. The JP route uses 4837 return for all three networks. This reminded me that the old RFCHOST JPCO Clawless also used three-network 4837 return, also had 100Mbps/200Mbps bandwidth, and also promoted something close to dedicated bandwidth in its user group. Combined with upstream JINX, it feels like a spiritual successor to Clawless. 
That said, the "dedicated" marketing is a bit too exaggerated: at 35 USD/year plus traffic billing, it is impossible to have that much dedicated China-bound bandwidth. As a machine itself, the value is acceptable; 200Mbps single-thread is enough for daily use, and Unicom/Mobile are also good. However, Jtti is not well known, and this is a non-refundable annual product, so I recommend watching a bit longer.
The provider has mixed reviews; long-term route-machine network performance may not be guaranteed, so be cautious with annual payment; there is no refund policy; the machine was self-purchased and this article has no affiliate links.
(1) Network quality
Speed: consistency between evening peak and daytime off-peak is excellent. In the Pearl River Delta and Yangtze River Delta, single-thread speeds on all three networks reach 160-180Mbps, and cross-network performance is decent.
Latency: Telecom has high latency because return uses cross-network 4837, even though outbound is optimized through CN2. NetQuality also measured serious Telecom return packet loss during the evening peak. Unicom has no cross-network issue, but latency is still 20-30 ms higher than Secretary Gen2/RFCHOST JPCO. Mobile latency is quite low. During the evening peak, latency on all three networks increases slightly.
Routes: return for all three networks is forced through 4837. Telecom outbound uses CN2->NTT, avoiding the Telecom CN2 NCP submarine-cable fault; Unicom outbound uses 4837->NTT, and Mobile outbound uses CMI.
International: despite the small 200Mbps bandwidth, it can almost fully saturate both local and US West destinations. Local latency is 1-2 ms, Seattle is 88-96 ms, and Los Angeles is 100 ms+.
(2) IP quality
IPQuality shows full streaming unlock, with partial local streaming unlock such as DMM. Confusingly, Google geolocates it to Singapore, so the location still needs to be pulled back. There are also forum reports of Google geolocating to China, without locked geolocation, so it depends on luck.
(3) Machine performance
Skylake architecture; CPU and I/O are fine for this price, and the 50G disk is quite generous.
Product Configuration
| Plan | 🇯🇵 Jtti JP.OP.1C1G |
|---|---|
| Location | Japan, Tokyo |
| CPU | 1 x Intel Xeon Platinum 8168 vCore |
| RAM | 1GB |
| Disk | 50GB |
| Traffic | 1TB(two-way billing) |
| Peak Bandwidth | 200Mbps |
| Return Route | three-network 4837 |
| Outbound Route | TelecomCN2/Unicom 4837/Mobile CMI |
| Network | 1 x IPv4 |
| Price | Annual USD35.00 |
Purchase address:https://www.jtti.cc/zh/activity/y2026-anniversary.html
Test Overview
| Live Probe | YABS | Node Quality | Mixed Benchmark | nws.sh | Streaming Unlock |
|---|---|---|---|---|---|
| [Link] | [Link] | [Link] | [Link] | [Link] | [Link] |
1. Network Quality
Single-Thread Return-to-China Speed Tests
| Evening Peak | Off-Peak |
|---|---|
NetQuality
bash ‹(curl -sL https://Check. Place) -N
| Evening Peak | Off-Peak |
|---|---|
Evening Peak Outbound TCP Ping
Test Site: itdog.cn
Test Time: 2026-03-10 22:00
| China Telecom | China Unicom | China Mobile |
|---|---|---|
Off-Peak Outbound TCP Ping
Test Site: itdog.cn
Test Time: 2026-03-10 10:00
| China Telecom | China Unicom | China Mobile |
|---|---|---|
Return Routes
bash ‹(curl -sL https://Check. Place) -N
Outbound Routes
Test Site: itdog.cn
Test Time: 2026-03-11 19:00
| ShanghaiTelecom | ShanghaiUnicom | Shanghai Mobile |
|---|---|---|
bgp.tools
Test Site: bgp.tools
Test Time: 2026-03-11 19:00
P.S. You can compare it with RFCHOST's machine: the same international JINX and the same single-line NTT.
Web Latency
bash <(wget -qO- https://raw.githubusercontent.com/danger-dream/network-latency-tester/main/latency.sh)
International Latency(iplark)
Test Site: iplark.com
Test Time: 2025-03-11 19:00
Own-node Tests
Test Method: iPerf3 TCP single-thread & ICMP Ping, one 10-second run
Test Time: 2026-03-11 20:00
| iPerf3 | ICMP Ping |
|---|---|
Cloudflare Local Speed Test
wget -qO- https://github.com/code-inflation/cfspeedtest/releases/latest/download/cfspeedtest-x86_64-unknown-linux-gnu.tar.gz | tar xz && ./cfspeedtest -n 25; rm ./cfspeedtest
Avg GET request latency 47.77 ms (RTT excluding server processing time)
Type Payload | min/max/avg in mbit/s
Download 100KB | min 9.10 max 24.26 avg 17.55
Download 1MB | min 78.12 max 112.63 avg 101.28
Download 10MB | min 162.61 max 178.56 avg 174.12
Upload 100KB | min 9.64 max 48.54 avg 37.13
Upload 1MB | min 66.95 max 276.44 avg 205.47
Upload 10MB | min 128.57 max 344.46 avg 216.49
Apple Local Speed Test
curl -sL nxtrace.org/speedtest | bash
▸ Idle Latency
[+] Endpoint: https://mensura.cdn-apple.com/api/v1/gm/small
[+] Samples: 20
➜ 18.36 ms median (min 16.84 / avg 18.70 / max 25.48) jitter 0.45 ms
▸ Download (single thread)
[+] Threads: 1
[+] Limit: 2G / 10s per thread
➜ 185 Mbps (225.0 MiB in 10.2s)
[+] Loaded latency: 38.37 ms (jitter 0.47 ms)
▸ Download (multi-thread)
[+] Threads: 4
[+] Limit: 2G / 10s per thread
➜ 183 Mbps (224.8 MiB in 10.3s, 4 threads)
[+] Loaded latency: 46.41 ms (jitter 0.53 ms)
▸ Upload (single thread)
[+] Threads: 1
[+] Limit: 2G / 10s per thread
➜ 203 Mbps (242.1 MiB in 10.0s)
[+] Loaded latency: 33.42 ms (jitter 6.23 ms)
▸ Upload (multi-thread)
[+] Threads: 4
[+] Limit: 2G / 10s per thread
➜ 204 Mbps (246.1 MiB in 10.1s, 4 threads)
[+] Loaded latency: 50.47 ms (jitter 0.92 ms)
Speedtest International Test
curl -sL nws.sh | bash
---------------------------------- nws.sh ---------------------------------
A simple script to bench network performance using speedtest-cli
---------------------------------------------------------------------------
Version : v2025.11.07
Global Speedtest : wget -qO- nws.sh | bash
Region Speedtest : wget -qO- nws.sh | bash -s -- -r <region>
Ping & Routing : wget -qO- nws.sh | bash -s -- -rt <region>
---------------------------------------------------------------------------
Speedtest.net (Region: GLOBAL)
---------------------------------------------------------------------------
Location Latency Loss DL Speed UP Speed Server
ISP: Jt Telecom International Pte.ltd.
Nearest 1.02 ms N/A 189.68 Mbps 204.18 Mbps GSL Networks - Tokyo
Bangalore, IN 103.91 ms 0.0% 192.17 Mbps 204.51 Mbps Bharti Airtel Ltd - Bangalore
Chennai, IN 126.97 ms 0.0% 193.03 Mbps 203.27 Mbps RailTel Corporation of India Ltd - Chennai
Mumbai, IN 137.64 ms 0.0% 207.74 Mbps 202.08 Mbps Melbicom - Mumbai
Seattle, US 92.49 ms N/A 191.92 Mbps 198.56 Mbps Comcast - Seattle, WA
Los Angeles, US 105.37 ms 0.0% 193.90 Mbps 205.39 Mbps ReliableSite Hosting - Los Angeles, CA
Dallas, US 142.62 ms 0.0% 192.56 Mbps 197.35 Mbps Hivelocity - Dallas, TX
Miami, US 213.83 ms 0.0% 178.21 Mbps 198.41 Mbps Frontier - Miami, FL
New York, US 144.09 ms 0.0% 192.47 Mbps 197.99 Mbps GSL Networks - New York, NY
Toronto, CA 174.48 ms 0.0% 195.12 Mbps 198.13 Mbps Rogers - Toronto, ON
Mexico City, MX 138.93 ms 0.0% 194.33 Mbps 207.36 Mbps INFINITUM - Ciudad de México
London, UK 221.23 ms 0.0% 174.88 Mbps 187.10 Mbps VeloxServ Communications - London
Amsterdam, NL FAILED
Paris, FR 220.22 ms 0.0% 192.38 Mbps 201.23 Mbps Scaleway - Paris
Frankfurt, DE 230.12 ms 0.0% 194.52 Mbps 206.52 Mbps Clouvider Ltd - Frankfurt am Main
Warsaw, PL 259.81 ms 0.0% 163.23 Mbps 197.89 Mbps Play - Warszawa
Bucharest, RO 253.57 ms 0.0% 188.98 Mbps 203.47 Mbps Digi
Moscow, RU FAILED
Jeddah, SA 179.78 ms 0.0% 207.55 Mbps 207.23 Mbps Saudi Telecom Company
Dubai, AE 147.75 ms N/A 203.63 Mbps 223.61 Mbps e& UAE - Dubai
Istanbul, TR 273.51 ms 0.0% 170.98 Mbps 194.79 Mbps Turkcell - Istanbul
Tehran, IR FAILED
Cairo, EG 278.84 ms 0.0% 193.52 Mbps 195.49 Mbps Telecom Egypt - Cairo
Tokyo, JP FAILED
Shanghai, CU-CN 59.96 ms N/A 191.67 Mbps 201.24 Mbps China Unicom 5G - Shanghai
Hong Kong, CN 46.75 ms 0.0% 209.72 Mbps 93.15 Mbps Misaka Network, Inc. - Hong Kong
Singapore, SG 74.03 ms 0.0% 193.38 Mbps 202.77 Mbps ViewQwest - Singapore
Jakarta, ID 79.31 ms 0.0% 191.92 Mbps 203.25 Mbps PT Solnet Indonesia - Jakarta
Sydney, AU 172.73 ms 0.0% 202.40 Mbps 193.59 Mbps Aussie Broadband - Sydney
---------------------------------------------------------------------------
Avg DL Speed : 192.00 Mbps
Avg UL Speed : 197.14 Mbps
Total DL Data : 6.78 GB
Total UL Data : 6.63 GB
Total Data : 13.42 GB
---------------------------------------------------------------------------
Duration : 12 min 30 sec
System Time : 10/03/2026 - 19:13:40 CST
Total Script Runs : 520831
---------------------------------------------------------------------------
Result : https://result.nws.sh/r/1773141221_81R641_GLOBAL.txt
---------------------------------------------------------------------------
2. IP Quality
IPQuality
bash ‹(curl -sL https://Check. Place) -I
Streaming Unlock
bash <(curl -L -s check.unlock.media)
Common IP Database Checks
export noninteractive=true && curl -L https://raw.githubusercontent.com/oneclickvirt/ecs/master/goecs.sh -o goecs.sh && chmod +x goecs.sh && ./goecs.sh install && goecs
--------------------------------------IP Quality Check--------------------------------------
The following are database IDs; results include the corresponding source database IDs
ipinfo Database [0] | scamalytics Database [1] | virustotal Database [2] | abuseipdb Database [3] | ip2location Database [4]
ip-api Database [5] | ipwhois Database [6] | ipregistry Database [7] | ipdata Database [8] | db-ip Database [9]
ipapiis Database [A] | ipapicom Database [B] | bigdatacloud Database [C] | dkly Database [D] | ipqualityscore Database [E]
ipintel Database [F] | ipfighter Database [G] | fraudlogix Database [H] | cloudflare Database [I] |
IPv4:
Security Score:
Trust Score(higher is better): 100 [8]
VPN Score(lower is better): 0 [8]
Proxy Score(lower is better): 0 [8]
Community Votes - Harmless: 0 [2]
Community Votes - Malicious: 0 [2]
Threat Score(lower is better): 1 [8]
Fraud Score(lower is better): 0 [E]
Abuse Score(lower is better): 0 [3]
ASN Abuse Score(lower is better): 0.005 (Low) [A]
Company Abuse Score(lower is better): 0.002 (Low) [A]
Threat Level: low [9 B]
Traffic Share: Human(higher is better)20% [I] Bot(lower is better)79% [I]
Blacklist Record Statistics (number of blacklist sites with records):
Harmless Records: 0 [2] Malicious Records: 0 [2] Suspicious Records: 0 [2] No-record Count: 94 [2]
Security Info:
Usage Type: isp [0 3 7 A] unknown [8 C] business [9]
Company Type: government [A] business [0] hosting [7]
Browser Type: Mainstream94% Other5% [I]
Device Type: Desktop96% Mobile3% Other0% [I]
OS Type: Mainstream96% Other3% [I]
Is Cloud Provider: Yes [7 D]
Is Data Center: No [0 5 8 A C G]
Is Mobile Device: No [5 A C G] Yes [E]
Is Proxy: No [0 4 5 7 8 9 A B C D E] Yes [G]
Is VPN: No [0 7 A C D E G]
Is Tor: No [0 3 7 8 A B C D E]
Is Tor Exit: No [7 D]
Is Web Crawler: No [9 A B E]
Is Anonymous: No [7 8 D]
Is Attacker: No [7 8 D]
Is Abuser: No [7 8 A C D E]
Is Threat: No [7 8 C D]
Is Relay: No [0 7 8 C D]
Is Bogon: No [7 8 A C D]
Is Bot: No [E]
DNS Blacklist: 309(Total_Check) 0(Clean) 8(Blacklisted) 12(Other)
iplark Check Score
Test Site: iplark.com
Test Time: 2026-03-11 19:00
3. Machine Performance
HardwareQuality
bash ‹(curl -sL https://Check. Place) -I
YABS
curl -sL https://yabs.sh | bash -s -- -5 -i
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
# Yet-Another-Bench-Script #
# v2025-04-20 #
# https://github.com/masonr/yet-another-bench-script #
# ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## ## #
Wed Mar 11 09:09:19 PM CST 2026
Basic System Information:
---------------------------------
Uptime : 1 days, 21 hours, 39 minutes
Processor : Intel(R) Xeon(R) Platinum 8168 CPU @ 2.70GHz
CPU cores : 1 @ 2693.670 MHz
AES-NI : ✔ Enabled
VM-x/AMD-V : ✔ Enabled
RAM : 949.5 MiB
Swap : 1024.0 MiB
Disk : 48.9 GiB
Distro : Debian GNU/Linux 13 (trixie)
Kernel : 6.12.63+deb13-cloud-amd64
VM Type : KVM
IPv4/IPv6 : ✔ Online / ❌ Offline
IPv4 Network Information:
---------------------------------
ISP : JT TELECOM INTERNATIONAL PTE.LTD.
ASN : AS137535 JT TELECOM INTERNATIONAL PTE.LTD.
Host : Hong Kong Communications International Co., Limited
Location : Tokyo, Tokyo (13)
Country : Japan
fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/sda2):
---------------------------------
Block Size | 4k (IOPS) | 64k (IOPS)
------ | --- ---- | ---- ----
Read | 40.01 MB/s (10.0k) | 553.34 MB/s (8.6k)
Write | 40.11 MB/s (10.0k) | 556.25 MB/s (8.6k)
Total | 80.13 MB/s (20.0k) | 1.10 GB/s (17.3k)
| |
Block Size | 512k (IOPS) | 1m (IOPS)
------ | --- ---- | ---- ----
Read | 1.01 GB/s (1.9k) | 1.00 GB/s (978)
Write | 1.06 GB/s (2.0k) | 1.06 GB/s (1.0k)
Total | 2.08 GB/s (4.0k) | 2.07 GB/s (2.0k)
Geekbench 5 Benchmark Test:
---------------------------------
Test | Value
|
Single Core | 820
Multi Core | 829
Full Test | https://browser.geekbench.com/v5/cpu/24171810
YABS completed in 3 min 26 sec
4. Appended at the End
About the Tests
The machines used in this test 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
Acknowledgements
Thanks to the administrators and volunteers of Haoyong Aiyong for providing public-interest speed test services; thanks to itdog.cn and iplark.com and the contributors of the test scripts
🔔 Thanks for reading this far. You can subscribe to my channel to receive reminders for new NS reviews.