A good way to start is to backup the current settings to be able to restore them in case the optimization turns out to make things worse. Inexperienced users might want to use the Optimal Settings button in the main interface to apply changes automatically. It has to be noted that the TCP Optimizer will also backup the settings before making any changes to the system.

Advanced users on the other hand can run a handful of tests that test the latency or the largest MTU value before they change dozens of settings manually. Changes range from setting the MTU value to the maximum connections per server and host resolution priority settings.

TCP Optimizer can be helpful if a user experiences problems with his Internet connection. It can be related to settings of the host system. If that is the case those settings can be optimized using the software.

One major problem is the reliability of those speed tests on the Internet. Some show lower speeds even if the Internet connection is able to handle faster speeds. This can be due to server capacities at the speed test host, the location of the servers in the world or general routing problems that affect the measured speed.

Speed.io is the first broadband speed test on the Internet that measured the speed of an Internet connection correctly. Every other speed test was usually missing a few thousand Kbits. The service measures the download and upload speed, the maximum number of connections and the response time of the Internet connection.

The download speed test result was 14806 Kbit and the result of the upload speed test was 990 Kbit which comes close to the maximum capacities of the 16 Mbit / 1 Mbit advertised speed of the Internet provider.

The developers of Speed.io claim to utilize 50.000 servers world wide for their speed tests which seems an awful lot but would be an incredible infrastructure if this was the case. Still, it does not really matter if they have 1 or 50K servers as long as the speed tests are accurate.