Before officially renting a Singapore cloud server, many users struggle with the question: "How fast and stable is this machine?" The configuration list might say "1Gbps bandwidth" and "high-performance CPU," but the actual experience can be quite different. Instead of regretting it later when the server is incredibly slow, it's better to spend 30 minutes conducting a comprehensive performance test before ordering or immediately after receiving the machine.

This tutorial will guide you step-by-step on how to use professional tools to test the network latency, bandwidth throughput, and overall performance of a Singapore cloud server, ensuring you rent with confidence and peace of mind.

Why test a Singapore cloud server before commercial use?

Singapore is a Southeast Asian internet hub with numerous data centers and abundant bandwidth resources. However, the line quality, hardware configuration, and network optimization levels vary greatly among service providers.

Some Singapore servers use premium CN2 GIA lines, with latency of only 30-60ms when accessing from mainland China; others bypass the US or Japan, resulting in latency exceeding 150ms. The same configuration and the same price can lead to vastly different experiences.

More importantly, many service providers advertise "1Gbps bandwidth" which is actually shared bandwidth—if your neighbor uses high bandwidth, your speed will be compressed. Without testing, you'll have no idea what you're actually buying.

Therefore, before renting a Singapore cloud server, these tests are essential: network latency and packet loss, bandwidth throughput, disk I/O performance, and overall CPU and memory performance.

Preparation Before Testing

Before starting the test, confirm three things:

1. The server is operational and has obtained a public IP address—this is a prerequisite for testing.

2. Ensure the firewall/security group has the test ports open—for example, iperf3 uses TCP/UDP port 5201 by default, which needs to be allowed in the console beforehand.

3. Prepare a local computer (Windows/Mac/Linux are all acceptable)—all test commands should be executed in the local terminal.

Tip: It is recommended to test at different times (daytime and peak evening hours) to fully understand the server's actual network quality.

First Item: Network Latency and Packet Loss Test (ping + MTR)

Test Purpose: To determine the network latency and packet loss rate from your location to the Singapore server.

Basic Latency Test – ping

Execute on your local terminal:

Windows

ping -t Singapore server IP

Linux/Mac

ping -c 100 Singapore server IP

It is recommended to ping at least 100 times and take the average. Focus on three key numbers:

- Average Latency (avg): The lower the better. The RTT from mainland China to Singapore CN2 GIA lines is typically between 30-60ms.

- Packet Loss Rate (loss): Ideally, it should be 0%. If it exceeds 1%, it indicates network instability.

- Maximum Latency (max): If the maximum latency is significantly higher than the average, it indicates network jitter.

Route Tracing and Packet Loss Location – MTR

ping only tells you "whether the result is good or bad," while MTR tells you "where the problem lies."

Install MTR (Linux)

apt-get install mtr -y Ubuntu/Debian

yum install mtr -y CentOS

Run MTR

mtr -r -c 100 Singapore server IP

MTR will display each routing node a packet passes through, along with the packet loss rate and latency for each hop.

How to interpret:

- If a hop experiences high packet loss, but subsequent hops return to normal → it's likely that an intermediate node has ICMP rate limiting set up, not a real network problem.

- If a hop experiences high packet loss and all subsequent hops continue to experience high packet loss → the real network bottleneck is here.

For Singapore servers, pay close attention to whether the return route uses a CN2 GIA node (starting with 59.43). If the entire route uses a regular international line (starting with 202.97 or other overseas ASNs), stability during peak hours should be questioned.

Second Item: Bandwidth and Throughput Test (iperf3 + speedtest-cli)

Test Purpose: To confirm the actual bandwidth you purchased, whether it's the "nominal value" or a "reduced version".

Professional Bandwidth Testing – iperf3

iperf3 is widely recognized as the most accurate bandwidth testing tool, capable of measuring the server's true TCP throughput under specific network conditions.

Step 1: Start the iperf3 service on the server

Install iperf3

apt-get install iperf3 -y Ubuntu/Debian

yum install iperf3 -y CentOS

Start the server (listening on port 5201)

iperf3 -s

Step 2: Perform client testing on your local computer

Install iperf3 (using brew on Mac, download the exe file on Windows)

Execute the test (10 concurrent streams, 60 seconds test)

iperf3 -c Singapore server IP -p 5201 -P 10 -t 60

Parameter explanation:

- `-P 10`: Enables 10 concurrent threads, making fuller use of bandwidth

- `-t 60`: Test lasts 60 seconds

How to interpret the test results?

- If the measured bandwidth is close to the advertised value you purchased (e.g., purchased 100Mbps, measured above 90Mbps), the bandwidth meets the requirements.

- If the speed is significantly lower than the advertised value (e.g., purchased 100Mbps but only measured 20Mbps), there may be overselling or line congestion.

Note: iperf3 tests the end-to-end bandwidth from your local machine to the server, which is greatly affected by your local network environment. It is recommended to test in different network environments (e.g., office, home, 4G/5G).

Quick Public Network Speed ​​Test – speedtest-cli

If you don't want to bother with iperf3's dual-end configuration, you can use speedtest-cli for a quick approximation.

Install speedtest-cli

apt-get install speedtest-cli -y

Execute the speed test

speedtest-cli

speedtest-cli will automatically select the nearest Speedtest server to test, providing download speed, upload speed, and latency. While not as accurate as iperf3, it's sufficient for a quick reference.

Third Item: Server Comprehensive Performance Test (bench.sh / yabs.sh)

Test Purpose: To confirm whether the performance of hardware such as CPU, memory, and disk I/O meets the standards.

For users seeking ultimate performance—such as those running databases, high-frequency trading systems, or AI inference services—measuring only the network is far from sufficient; server computing power and disk I/O speed are equally crucial.

One-click benchmarking script—bench.sh

bench.sh is a classic one-click benchmarking script among VPS users, capable of quickly collecting system information, performing disk I/O tests, and testing network speed.

Execute on the server:

curl -Lso- bench.sh | bash

More comprehensive benchmarking — yabs.sh

For more detailed CPU benchmarks and disk IOPS data, use yabs.sh:

curl -sL yabs.sh | bash

Specific disk performance testing — fio

If you suspect disk performance issues, you can use fio for specific testing:

Install fio

apt-get install fio -y

4K random read/write test (simulating database scenarios)

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 \

--name=test --filename=test --bs=4k --iodepth=64 --size=1G \

--rw=randrw --rwmixread=75

Key metrics:

- IOPS: Input/output times per second, higher is better

- Latency: Read/write latency, less than 5ms is considered excellent.

Test Results Summary: What data qualifies as "acceptable"?

If multiple test results fall within the "caution required" range, it is recommended to try a different service provider or a different network connection.

Why Choose Jtti Singapore Cloud Servers?

If you are looking for a Singapore cloud server that can withstand all the above tests, Jtti is a choice worth considering.

Jtti Singapore cloud servers are equipped with Intel Xeon Gold processors and enterprise-grade NVMe SSD arrays, achieving enterprise-level standards in both computing performance and disk I/O—the data from the benchmark and fio tests will give you peace of mind.

More importantly, there's the network connectivity. Jtti's Singapore data center uses a premium CN2 GIA line, ensuring stable latency of 30-60ms for access from mainland China, avoiding detours and latency spikes during peak hours. Real-world testing shows that the CN2 GIA line has a near-zero packet loss rate and jitter typically below 5ms—meaning your website, application, or game server can provide users with a truly smooth experience.

In terms of configuration, Jtti offers a variety of options, from entry-level 1-core 1GB to enterprise-grade 8-core 16GB, supporting both Linux and Windows dual-boot systems to meet all needs from personal projects to enterprise-level deployments.

Choosing a Jtti Singapore cloud server means choosing a reliable machine that can withstand rigorous testing.

Visit the Jtti website now to choose the Singapore cloud server plan that best suits your needs!