Knowledgebase

If you are experiencing low network performance on your VPS their are a handful of factors that could be effecting the network speed.

CPU Overload

VPS instances rely heavily on the CPU to handle network traffic, especially when using a virtual network interface like VirtIO. A single-core instance may negatively impact network performance due to limited CPU resources.

Solutions:

1. Monitor CPU Usage: Regularly check your VPS's CPU usage. High CPU usage can adversely affect performance, as the virtual network interface depends on CPU resources.

2. Upgrade or Reboot: Consider upgrading your VPS plan to one with more CPU cores to enhance performance. Alternatively, reboot your VPS occasionally to clear any stuck processes.

It's important to note that a single-core instance might not always support higher network speeds effectively.

Incorrect or Slow Network Driver

We offer various network drivers for your VPS, with VirtIO being the most efficient. Using a slower driver like E1000 can negatively impact network performance.

Solutions:

• Ensure your VPS uses the VirtIO driver, optimized for performance.

• Follow this guide to enable the VirtIO driver.

While the E1000 driver supports Gigabit Ethernet speeds (up to 1 Gbit/s), it is less efficient than VirtIO in virtualized environments.

I/O Bottlenecks

Network performance can be impacted by high disk I/O activity on your VPS or the host node, particularly if using shared or network-attached storage. High I/O wait times or CPU steal time can degrade both your VPS and its network efficiency.

Solutions:

1. Monitor System Performance: Use the top command on your VPS to check system metrics. Focus on these key indicators:

   • %wa (I/O Wait): Indicates how much time the CPU waits for disk operations. High values signal I/O bottlenecks.

   • %st (Steal Time): Reflects the time the host node uses your VPS's CPU. Elevated levels indicate resource contention on the hypervisor.

2. Addressing High Values:

   • High %wa: Optimize your disk usage to alleviate I/O bottlenecks.

   • High %st: Contact our support team for further assistance if CPU steal time is significant.

Internet Congestion and MTR Diagnosis

Network congestion on the path between your VPS and your location can sometimes cause slow performance. To identify and troubleshoot this issue, it is important to analyse the network path in both directions.

Solution:

1. Run an MTR (My Traceroute):

   • From your location to your VPS.

   • From your VPS to your location.

   Ensure each test includes a minimum of 1000 pings to capture comprehensive data about the path, including any latency or packet loss.

2. Submit the Results: Send the MTR results to our support team. This information will assist in diagnosing and resolving any network congestion issues.