|
Telco networks must be extremely tolerant of error conditions. Virtually every person and business depends on network connectivity.
Telco networks are critically important to modern life, and it would be
unacceptable if the networks were not reliable and robust against
attack.
Because of the fact that Telco network elements are communicating based on standards, it is not uncommon for these network elements to be exposed to traffic that doesn't quite measure up to the expectations written into the standards. Telco networks also need to be able to scale to handle a variety of kinds of “stress” that they might encounter during normal network operations.
When carriers are selecting new vendor equipment for use in their future network planning, it is critical that they establish a baseline, especially around interoperability of the protocols they intend to use within their service offerings. One of the best ways to accomplish this is via fuzzing, either based on open protocol specifications or on service-specific traffic flows that must be supported by the telco network. Mu's Protocol Fuzzing and Studio Zx modules make it easy to compare the results of many different target devices to see which is able to stand up to invalid or unexpected inputs around the traffic that must be supported by the network.
The purpose of resilience testing is to ensure the network continues to perform well - per the operator's definition of “well,” while under a variety of kinds of stress that can establish whether the IP Service is going to stand up to the worst conditions that a real network might experience. Resilience testing might be as innocuous as something an engineering or support team performs after Functional and/or Interoperability testing, and could amount to answering this question: "I know that it (the IP product or service) works, but if I run it at a heavy load or with varied user conditions, can it run overnight? Or over a weekend? Maybe for a week?"
This type of load is stateful, and equivalent to what happens in a real deployment where a network goes from 10 users to 100 users to 10,000 users over a period of time. The growth is likely steady, but is the IP Service able to stand up to the load? Clearly, the model of the IP Service is still the key to delivering this sort of resilience testing.
Resilience is also established through other tools, like Denial-of-Service (DoS) simulation. In a DoS testing model, the service is used to monitor the service with stateful, realistic service traffic while stateless DoS traffic is targeted at the key service components to ensure that each of them is not the weakest link in the service. For instance, how does a dDoS attack on the DNS server affect externally measurable service metrics?
Finally, resilience is established through fuzzing, wherein the service under test is still monitored (as above) using a valid service exchange to ensure that — despite a perfect storm of invalid or unexpected traffic — the service continues to function properly. In order for the provider to enjoy all the benefits of their network, the network must be practically bulletproof.
|
