Standard
ABSTRACT
The present document provides a contribution to the evolution of network performance testing towards a professional degree of transparency. This begins with a consistent framework of definitions and technical terms. The elements of the testing process are then described within this context.
Apart from the obvious direct parameters of throughput testing, such as time windows or transferred data volumes, there are numerous other elements which can have an impact on data values obtained. In this sense, methodology and definition of metrics cannot be decoupled from each other. The process starts with selecting the boundaries to the system under test, i.e. insertion or demarcation points. Next comes the way the system under test is accessed. For instance, if the test is run over a radio access network using a mobile device such as a smartphone, the type and degree of influence needs to be assessed. The type of stimulus is likewise important, such as the protocol type, the structure of data traffic (e.g. TCP or UDP based), and the number of parallel connections. Depending on these selections, other choices also become parameters for testing. An example would be to use some kind of real application to create a particular type of traffic, versus using synthetically generated traffic.