How to Benchmark VPS Network Performance with iPerf3, MTR, and Traceroute

Network performance is often the bottleneck that determines whether your VPS feels fast or sluggish to end users. Even a server with plenty of CPU and RAM will deliver poor application performance if the network latency is high, packet loss is frequent, or throughput is capped by the provider. If you are evaluating which VPS plan offers the best connectivity, compare VPS performance across providers to see real-world network benchmarks. This guide walks through practical network benchmarking using iPerf3, MTR, and traceroute — three tools that give you precise, actionable data about your VPS network.

Why Network Benchmarking Matters

VPS providers shape network performance in ways that are not always obvious from the spec sheet. A plan might advertise “1 Gbps port speed” but throttle throughput after a certain threshold, or route traffic through congested peering exchanges. Running your own benchmarks reveals the real performance your application will experience. The three metrics that matter most are:

• Throughput: The maximum data transfer rate between your VPS and clients, measured in Mbps or Gbps.
• Latency: The round-trip time for packets, measured in milliseconds. Lower is better for real-time applications.
• Packet loss: The percentage of packets that never reach their destination. Anything above 0.5% will noticeably impact performance.

Installing the Benchmark Tools

Install all three tools on your VPS:

Ubuntu/Debian:

sudo apt update
sudo apt install iperf3 mtr-tiny traceroute -y

Rocky Linux/RHEL 9:

sudo dnf install iperf3 mtr traceroute -y

Benchmarking Throughput with iPerf3

iPerf3 measures the maximum achievable throughput between two endpoints. You will need two machines: one running in server mode and one in client mode. For VPS benchmarking, run the server on your VPS and connect from your local machine (or use a second VPS in a different region).

On your VPS, start the iPerf3 server:

iperf3 -s

From your client machine, run a 30-second throughput test:

iperf3 -c YOUR_VPS_IP -t 30

Run the test in both directions by adding the -R flag for reverse mode (download from server perspective):

iperf3 -c YOUR_VPS_IP -t 30 -R

To test UDP throughput and jitter (important for VoIP and gaming):

iperf3 -c YOUR_VPS_IP -u -b 100M -t 30

iPerf3 outputs a summary showing transfer size, bandwidth, retransmissions, and jitter. Pay attention to retransmissions — a high retransmission rate indicates network congestion or packet loss somewhere along the path.

For the most accurate results, run iPerf3 tests at different times of day. VPS networks experience peak congestion in the evening hours (8 PM to 11 PM local time). Compare your morning and evening results to see if the provider shapes traffic during peak periods.

Analyzing the Route with MTR

MTR (My Traceroute) combines the functionality of traceroute and ping into a single tool. It continuously probes each hop between your VPS and a target, showing latency and packet loss per hop. This is invaluable for identifying exactly where network problems occur.

Run MTR from your VPS to a target destination:

mtr -r -c 100 google.com

The -r flag produces a report (non-interactive), and -c 100 sends 100 probes. The output shows each hop with loss percentage, latency, and packet counts. Key things to look for:

• Packet loss at early hops (hops 1–3): This suggests a problem with your VPS provider’s network or your own network configuration.
• Packet loss at middle hops (hops 4–10): This is typically a transit provider or peering issue. Escalate to your VPS provider with the MTR output.
• Latency spikes: A sudden jump from 10 ms to 150 ms at a specific hop indicates your traffic is being routed through a distant geographic location.
• Loss at the final hop only: This is often normal — the target server may prioritize other traffic over ICMP probes.

Run MTR in both directions — from your VPS to your local machine, and from your local machine to your VPS. One-way routing problems are common (asymmetric routing) and only show up when tested from both directions.

Using Traceroute for Quick Path Analysis

Traceroute gives you a one-shot view of the route path without the continuous monitoring of MTR. It is useful for quick checks and for documenting the route:

traceroute -n google.com

The -n flag skips DNS resolution for faster results. Traceroute shows each hop’s IP address and the three probe latencies. Typical output shows 8–15 hops for a cross-continent connection:

 1  10.0.0.1  0.346 ms  0.463 ms  0.330 ms
 2  192.168.1.1  0.685 ms  0.895 ms  0.670 ms
 3  203.0.113.1  2.134 ms  2.456 ms  2.001 ms
 4  198.51.100.1  15.234 ms  14.987 ms  15.102 ms

If any hop shows three asterisks (* * *), that router is not responding to ICMP probes — this is not necessarily a problem, as many routers intentionally drop ICMP traffic. However, if you see multiple consecutive asterisks followed by high latency or loss, there may be a routing issue.

Interpreting Your Benchmark Results

Here is how to evaluate the numbers you collect:

• Throughput below 100 Mbps on a “1 Gbps” port: The provider may be oversubscribing their network or throttling your connection. Run tests at multiple times to confirm.
• Latency above 100 ms for regional traffic: Check if your VPS is in a different geographic region than advertised, or if traffic is being routed through an unnecessary transit provider.
• Packet loss above 1%: This will cause TCP retransmissions and degrade application performance. Open a support ticket with your MTR output.
• High jitter (above 5 ms): For real-time applications like VoIP, video conferencing, or gaming, high jitter is worse than high latency. Consider switching providers or adding a jitter buffer.

Creating a Benchmark Report

Document your findings in a structured report. Here is a template you can follow:

• Date and time of test
• VPS provider, plan name, and location
• Target server or client IP and location
• iPerf3 upload throughput (Mbps)
• iPerf3 download throughput (Mbps)
• iPerf3 retransmission count
• MTR average latency and max loss % per hop
• Traceroute hop count

Run the full benchmark suite quarterly and after any provider maintenance or network changes. Keeping a historical record helps you spot performance degradation before it affects your users.

Conclusion

Benchmarking your VPS network with iPerf3, MTR, and traceroute gives you concrete data about real-world throughput, latency, and packet loss. These three tools cover the full picture: how fast data moves, where slowdowns occur, and whether packets are being lost along the way. When you are ready to choose or switch providers, see our provider benchmarks and comparison table to find a VPS that delivers the network performance your applications need.