Stateful Testing

Although Frame Blasting will provide you with numbers that give a good insight into the performance of the DUT/SUT, it can’t simulate every situation. Some protocols, like TCP can’t be reproduced with Frame Blasting alone. These protocols control their own speed according to the current circumstances.

Stateful testing is very useful when reproducing problems or possible configurations. It is the ideal tool for simulating real live networks and evaluating network performance. It will help you understand the impact of configurations on user-experience, network speed, etc.

We currently support DHCP, IGMP, ICMP, ARP and TCP for IPv4. For IPv6, we support MLD, DHCPv6, router, and neighbor discovery. TCP is the only protocol that will be used for benchmarking. The other protocols will support the normal operation of the test.

Note: The only protocol currently supported for stateful testing is TCP.

The result of the stateful test using TCP is the average speed. Besides that, ByteBlower will inform you why the TCP connection has been closed. Normally, this will be because the requested data is exchanged. If not, ByteBlower will inform you about the exception.

TCP also keeps track of a congestion window. The minimum and maximum size of the congestion window is also returned at the end of the test. The congestion window is a good indication of how difficult it was to exchange the data. A small congestion window is an indication that the network was saturated.

Conclusion: Protocol testing is ideal for reproducing network behaviour. Results are not always exactly the same, but it is a good simulation of a real network.