Dynamic program analysis is the act of analyzing software that involves executing a program as opposed to static program analysis, which does not execute it.
Analysis can focus on different aspects of the software including but not limited to: behavior, test coverage, performance and security.
To be effective, the target program must be executed with sufficient test inputs[1] to address the ranges of possible inputs and outputs. Software testing measures, such as code coverage, and tools such as mutation testing, are used to identify where testing is inadequate.
See main article: Software testing.
Functional testing includes relatively common programming techniques such as unit testing, integration testing and system testing.[2]
Computing the code coverage of a test identifies code that is not tested; not covered by a test.
Although this analysis identifies code that is not tested it does not determine whether tested coded is adequately tested. Code can be executed even if the tests do not actually verify correct behavior.
See main article: Dynamic testing.
Dynamic testing involves executing a program on a set of test cases.
See main article: Fuzzing. Fuzzing is a testing technique that involves executing a program on a wide variety of inputs; often these inputs are randomly generated (at least in part). Gray-box fuzzers use code coverage to guide input generation.
See main article: Concolic testing.
Dynamic symbolic execution (also known as DSE or concolic execution) involves executing a test program on a concrete input, collecting the path constraints associated with the execution, and using a constraint solver (generally, an SMT solver) to generate new inputs that would cause the program to take a different control-flow path, thus increasing code coverage of the test suite.[3] DSE can considered a type of fuzzing ("white-box" fuzzing).
Dynamic data-flow analysis tracks the flow of information from sources to sinks. Forms of dynamic data-flow analysis include dynamic taint analysis and even dynamic symbolic execution.[4] [5]
Daikon is an implementation of dynamic invariant detection. Daikon runs a program, observes the values thatthe program computes, and then reports properties that were true over the observed executions, and thus likely true over all executions.
Dynamic analysis can be used to detect security problems.
See main article: Program slicing. For a given subset of a program’s behavior, program slicing consists of reducing the program to the minimum form that still produces the selected behavior. The reduced program is called a “slice” and is a faithful representation of the original program within the domain of the specified behavior subset.Generally, finding a slice is an unsolvable problem, but by specifying the target behavior subset by the values of a set of variables, it is possible to obtain approximate slices using a data-flow algorithm. These slices are usually used by developers during debugging to locate the source of errors.
Most performance analysis tools use dynamic program analysis techniques.
Most dynamic analysis involves instrumentation or transformation.
Since instrumentation can affect runtime performance, interpretation of test results must account for this to avoid misidentifying a performance problem.
DynInst is a runtime code-patching library that is useful in developing dynamic program analysis probes and applying them to compiled executables. Dyninst does not require source code or recompilation in general, however, non-stripped executables and executables with debugging symbols are easier to instrument.
Iroh.js is a runtime code analysis library for JavaScript. It keeps track of the code execution path, provides runtime listeners to listen for specific executed code patterns and allows the interception and manipulation of the program's execution behavior.