Whitebox TestingWhite box testing is very different in nature from black box testing. In black box testing, focus of all the activities is only on the functionality of system and not on what is happening inside the system.
Purpose of white box testing is to make sure that
* Functionality is proper
* Information on the code coverage
White box is primarily development teams job, but now test engineers have also started helping development team in this effort by contributing in writing unit test cases, generating data for unit test cases etc.
White box testing can be performed at various level starting from unit to system testing. Only distinction between black box and white box is the system knowledge, in white box testing you execute or select your test cases based on the code/architectural knowledge of system under test.
What is White box Testing?
White box testing (a.k.a. clear box testing, glass box testing or structural testing) uses an internal perspective of the system to design test cases based on internal structure. It requires programming skills to identify all paths through the software. The tester chooses test case inputs to exercise paths through the code and determines the appropriate outputs. In electrical hardware testing, every node in a circuit may be probed and measured; an example is in-circuit testing (ICT).
Since the tests are based on the actual implementation, if the implementation changes, the tests probably will need to change, too. For example ICT needs updates if component values change, and needs modified/new fixture if the circuit changes. This adds financial resistance to the change process, thus buggy products may stay buggy. Automated optical inspection (AOI) offers similar component level correctness checking without the cost of ICT fixtures, however changes still require test updates.
While white box testing is applicable at the unit, integration and system levels of the software testing process, it is typically applied to the unit. While it normally tests paths within a unit, it can also test paths between units during integration, and between subsystems during a system level test. Though this method of test design can uncover an overwhelming number of test cases, it might not detect unimplemented parts of the specification or missing requirements, but one can be sure that all paths through the test object are executed.
There are different type of coverage that can be targeted in white box testing
* Statement coverage
* Function coverage
* Decision coverage
* Decision and Statement coverage
Conditional Testing
The first improvement to white box techniques is to ensure that the Boolean controlling expressions are adequately tested, a process known as condition testing.
The process of condition testing ensures that a controlling expression has been adequately exercised whilst the software is under test by constructing a constraint set for every expression and then ensuring that every member on the constraint set is included in the values which are presented to the expression. This may require additional test runs to be included in the test plan.
To introduce the concept of constraint sets the simplest possible Boolean condition, a single Boolean variable or a negated Boolean variable, will be considered. These conditions may take forms such as:
if DateValid then while not DateValid then
The constraint set for both of these expressions is {t ,f } which indicates that to adequately test these expressions they should be tested twice with DateValid having the values True and False.
Perhaps the next simplest Boolean condition consists of a simple relational expression of the form value operator value, where the operator can be one of: is equal to ( = ), is not equal to ( /= ), is greater than ( > ), is less than ( <), is greater than or equal to ( >= ) and is less than or equal to ( <= ). It can be noted that the negation of the simple Boolean variable above had no effect upon the constraint set and that the six relational operators can be divided into three pairs of operators and their negations. Is equal to is a negation of is not equal to, is greater than is a negation of is less than or equal to and is less than is a negation of is greater than or equal to. Thus the condition set for a relational expression can be expressed as {=, >, <}, which indicates that to adequately test a relational expression it must be tested three times with values which ensure that the two values are equal, that the first value is less than the second value and that the first value is greater than the second value.
More complex control expressions involve the use of the Boolean operators, and, or and xor, which combine the values of two Boolean values. To construct a constraint set for a simple Boolean expression of the form BooleanValue operator BooleanValue all possible combinations of True and False have to be considered. This gives the constraint set for the expression as {{t ,t } {t ,f } {f ,t } {f ,f }}. If both BooleanValues are simple or negated Boolean variables then no further development of this set is required. However if one or both of the BooleanValues are relational expressions then the constraint set for the relational expression will have to be combined with this constraint set. The combination takes the form of noting that the equality condition is equivalent to true and the both inequality conditions are equivalent to false. Thus every true condition in the condition set is replaced with ' = ' and every false replaced twice, once with ' > ' and once with ' < '.
Thus if only one the left hand BooleanValue is a relational expression the condition set would be {{= ,t } {= ,f } {> ,t } {< ,t } {> ,f } {< ,f }}. And if both BooleanValues are relation expressions this would become {{= ,= } {= ,> } {= ,< } {> ,= } {< ,= } {> ,> } {> ,< } {< ,> } {< ,< }}.
An increase in the complexity of the Boolean expression by the addition of more operators will introduce implicit or explicit bracketing of the order of evaluation which will be reflected in the condition set and will increase the number of terms in the set. For example a Boolean expression of the following form:
BooleanValue1 operator1 BooleanValue2 operator3nbsp;BooleanValue3
Has the implicit bracketing:
(BooleanValue1 operator1 BooleanValue2) operato3 BooleanValue3
The constraint set for the complete expression would be {{e1,t}{e1,f}}, where e1 is the condition set of the bracketed sub-expression and when it is used to expand this constraint set gives {{t,t,t} {t,f,t} {f,t,t} {f,f,t} {t,t,f} {t,f,f} {f,t,f} {f,f,f}}. If any of the BooleanValues are themselves relational expressions this will increase the number of terms in the condition set. In this example the worst case would be if all three values were relational expressions and would produce a total of 27 terms in the condition set. This would imply that 27 tests are required to adequately test the expression. As the number of Boolean operators increases the number of terms in a condition set increases exponentially and comprehensive testing of the expression becomes more complicated and less likely. It is this consideration which led to the advice, initially given in Section 1, to keep Boolean control expressions as simple as possible, and one way to do this is to use Boolean variables rather than expressions within such control conditions.
Data Life Cycle Testing
Keeping control expressions simple can be argued to simply distribute the complexity from the control expressions into the other parts of the subprogram and an effective testing strategy should recognise this and account for it. One approach to this consideration is known as data lifecyle testing, and it is based upon the consideration that a variable is at some stage created, and subsequently may have its value changed or used in a controlling expression several times before being destroyed. If only locally declared Boolean variables used in control conditions are considered then an examination of the source code will indicate the place in the source code where the variable is created, places where it is given a value, where the value is used as part of a control expression and the place where it is destroyed.
This approach to testing requires all possible feasible lifecycles of the variable to be covered whilst the module is under test. In the case of a Boolean variable this should include the possibility of a Boolean variable being given the values True and False at each place where it is given a value. A typical outline sketch of a possible lifecycle of a controlling Boolean variable within a subprogram might be as follows:
~~~ SomeSubProgram( ~~~ ) is
ControlVar : BOOLEAN := FALSE;
begin -- SomeSubProgram
~~~
while not ControlVar loop
~~~
ControlVar := SomeExpression;
end loop;
~~~
end SomeSubProg;
In this sketch ~~~ indicates the parts of the subprogram which are not relevant to the lifecycle. In this example there are two places where the variable ControlVar is given a value, the location where it is created and the assignment within the loop. Additionally there is one place where it is used as a control expression. There are two possible lifecycles to consider, one which can be characterised as { f t } indicating that the variable is created with the value False and given the value True upon the first iteration of the loop. The other lifecycle can be characterised as { f, f, ... t }, which differs from the first by indicating that the variable is given the value False on the first iteration, following which there is the possibility of more iterations where it is also given the value False, being given the value True on the last iteration.
Other possible lifecycles such as { f } or { f t f .. t } can be shown from a consideration of the source code not to be possible. The first is not possible as the default value will ensure that the loop iterates causing the variable to experience at least two values during its lifecycle. The second is not possible as soon as the variable is given the value True within the loop, the loop and subsequently the subprogram will terminate, causing the variable to be destroyed.
This quick look at the variable lifecycle approach only indicates the basis of this approach to white box testing. It should also indicate that this is one of the most laborious and thus expensive and difficult techniques to apply. As it is expensive and does not add a great deal to the testing considerations which have already been discussed, it is not widely used.
Loop Testing
The final white box consideration which will be introduced is the testing of loops, which have been shown to be the most common cause of faults in subprograms. If a loop, definite or indefinite, is intended to iterate n times then the test plan should include the following seven considerations and possible faults.
* That the loop might iterate zero times.
* That the loop might iterate once
* That the loop might iterate twice
* That the loop might iterate several times
* That the loop might iterate n - 1 times
* That the loop might iterate n times
* That the loop might iterate n + 1 times
* That the loop might iterate infinite times
All feasible possibilities should be exercised whilst the software is under test. The last possibility, an infinite loop, is a very noticeable, and common fault. All loops should be constructed in such a way that it is guaranteed that they will at some stage come to an end. However this does not necessarily guarantee that they come to an end after the correct number of iterations, a loop which iterates one time too many or one time too few is probably the most common loop fault. Of these possibilities an additional iteration may hopefully cause a CONSTRAINT_ERROR exception to be raised announcing its presence. Otherwise the n - 1 and n + 1 loop faults can be very difficult to detect and correct.
A loop executing zero times may be part of the design, in which case it should be explicitly tested that it does so when required and does not do so when it is not required. Otherwise, if the loop should never execute zero times, and it does, this can also be a very subtle fault to locate. The additional considerations, once, twice and many are included to increase confidence that the loop is operating correctly.
The next consideration is the testing of nested loops. One approach to this is to combine the test considerations of the innermost loop with those of the outermost loop. As there are 7 considerations for a simple loop, this will give 49 considerations for two levels of nesting and 343 considerations for a triple nested loop, this is clearly not a feasible proposition.
What is possible is to start by testing the innermost loop, with all other loops set to iterate the minimum number of times. Once the innermost loop has been tested it should be configured so that it will iterate the minimum number of times and the next outermost loop tested. Testing of nested loops can continue in this manner, effectively testing each nested loop in sequence rather than in combination, which will result in the number of required tests increasing arithmetically ( 7, 14, 21 ..) rather than geometrically ( 7, 49, 343 .. ).
Test Strategy
The first step in planning white box testing is to develop a test strategy based on risk analysis. The purpose of a test strategy is to clarify the major activities involved, key decisions made, and challenges faced in the testing effort. This includes identifying testing scope, testing techniques, coverage metrics, test environment, and test staff skill requirements. The test strategy must account for the fact that time and budget constraints prohibit testing every component of a software system and should balance test effectiveness with test efficiency based on risks to the system. The level of effectiveness necessary depends on the use of software and its consequence of failure. The higher the cost of failure for software, the more sophisticated and rigorous a testing approach must be to ensure effectiveness. Risk analysis provides the right context and information to derive a test strategy.
Test strategy is essentially a management activity. A test manager (or similar role) is responsible for developing and managing a test strategy. The members of the development and testing team help the test manager develop the test strategy.
Test Plan
The test plan should manifest the test strategy. The main purpose of having a test plan is to organize the subsequent testing process. It includes test areas covered, test technique implementation, test case and data selection, test results validation, test cycles, and entry and exit criteria based on coverage metrics. In general, the test plan should incorporate both a high-level outline of which areas are to be tested and what methodologies are to be used and a general description of test cases, including prerequisites, setup, execution, and a description of what to look for in the test results. The high-level outline is useful for administration, planning, and reporting, while the more detailed descriptions are meant to make the test process go smoothly.
While not all testers like using test plans, they provide a number of benefits:
* Test plans provide a written record of what is to be done.
* Test plans allow project stakeholders to sign off on the intended testing effort. This helps ensure that the stakeholders agree with and will continue to support the test effort.
* Test plans provide a way to measure progress. This allows testers to determine whether they are on schedule, and also provides a concise way to report progress to the stakeholders.
* Due to time and budget constraints, it is often impossible to test all components of a software system. A test plan allows the analyst to succinctly record what the testing priorities are.
* Test plans provide excellent documentation for testing subsequent releases—they can be used to develop regression test suites and/or provide guidance to develop new tests.
A test manager (or similar role) is responsible for developing and managing a test plan. The development managers are also part of test plan development, since the schedules in the test plan are closely tied to that of the development schedules.
Test Case Development
The last activity within the test planning stage is test case development. Test case description includes preconditions, generic or specific test inputs, expected results, and steps to perform to execute the test. There are many definitions and formats for test case description. In general, the intent of the test case is to capture what the particular test is designed to accomplish. Risk analysis, test strategy, and the test plan should guide test case development.
The members of the testing team are responsible for developing and documenting test cases.
Test Automation
Test automation provides automated support for the process of managing and executing tests, especially for repeating past tests. All the tests developed for the system should be collected into a test suite. Whenever the system changes, the suite of tests that correspond to the changes or those that represent a set of regression tests can be run again to see if the software behaves as expected. Test drivers or suite drivers support executing test suites. Test drivers basically help in setup, execution, assertion, and teardown for each of the tests. In addition to driving test execution, test automation requires some automated mechanisms to generate test inputs and validate test results. The nature of test data generation and test results validation largely depends on the software under test and on the testing intentions of particular tests.
In addition to test automation development, stubs or scaffolding development is also required. Test scaffolding provides some of the infrastructure needed in order to test efficiently. White box testing mostly requires some software development to support executing particular tests. This software establishes an environment around the test, including states and values for data structures, runtime error injection, and acts as stubs for some external components. Much of what scaffolding is for depends on the software under test. However, as a best practice, it is preferable to separate the test data inputs from the code that delivers them, typically by putting inputs in one or more separate data files. This simplifies test maintenance and allows for reuse of test code. The members of the testing team are responsible for test automation and supporting software development. Typically, a member of the test team is dedicated to the development effort.
Test Environment
Testing requires the existence of a test environment. Establishing and managing a proper test environment is critical to the efficiency and effectiveness of testing. For simple application programs, the test environment generally consists of a single computer, but for enterprise-level software systems, the test environment is much more complex, and the software is usually closely coupled to the environment.
For security testing, it is often necessary for the tester to have more control over the environment than in many other testing activities. This is because the tester must be able to examine and manipulate software/environment interactions at a greater level of detail, in search of weaknesses that could be exploited by an attacker. The tester must also be able to control these interactions. The test environment should be isolated, especially if, for example, a test technique produces potentially destructive results during test that might invalidate the results of any concurrent test or other activity. Testing malware (malicious software) can also be dangerous without strict isolation.
The test manager is responsible for coordinating test environment preparation. Depending on the type of environment required, the members of the development, testing, and build management teams are involved in test environment preparation.
Test Execution
Test execution involves running test cases developed for the system and reporting test results. The first step in test execution is generally to validate the infrastructure needed for running tests in the first place. This infrastructure primarily encompasses the test environment and test automation, including stubs that might be needed to run individual components, synthetic data used for testing or populating databases that the software needs to run, and other applications that interact with the software. The issues being sought are those that will prevent the software under test from being executed or else cause it to fail for reasons not related to faults in the software itself.
The members of the test team are responsible for test execution and reporting.
Testing Techniques
Data-Flow Analysis
Data-flow analysis can be used to increase program understanding and to develop test cases based on data flow within the program. The data-flow testing technique is based on investigating the ways values are associated with variables and the ways that these associations affect the execution of the program. Data-flow analysis focuses on occurrences of variables, following paths from the definition (or initialization) of a variable to its uses. The variable values may be used for computing values for defining other variables or used as predicate variables to decide whether a predicate is true for traversing a specific execution path. A data-flow analysis for an entire program involving all variables and traversing all usage paths requires immense computational resources; however, this technique can be applied for select variables. The simplest approach is to validate the usage of select sets of variables by executing a path that starts with definition and ends at uses of the definition. The path and the usage of the data can help in identifying suspicious code blocks and in developing test cases to validate the runtime behavior of the software. For example, for a chosen data definition-to-use path, with well-crafted test data, testing can uncover time-of-check-to-time-of-use (TOCTTOU) flaws. The ”Security Testing” section in explains the data mutation technique, which deals with perturbing environment data. The same technique can be applied to internal data as well, with the help of data-flow analysis.
Code-Based Fault Injection
The fault injection technique perturbs program states by injecting software source code to force changes into the state of the program as it executes. Instrumentation is the process of non-intrusively inserting code into the software that is being analyzed and then compiling and executing the modified (or instrumented) software. Assertions are added to the code to raise a flag when a violation condition is encountered. This form of testing measures how software behaves when it is forced into anomalous circumstances. Basically this technique forces non-normative behavior of the software, and the resulting understanding can help determine whether a program has vulnerabilities that can lead to security violations. This technique can be used to force error conditions to exercise the error handling code, change execution paths, input unexpected (or abnormal) data, change return values, etc. In runtime fault injection is explained and advocated over code-based fault injection methods. One of the drawbacks of code based methods listed in the book is the lack of access to source code. However, in this content area, the assumptions are that source code is available and that the testers have the knowledge and expertise to understand the code for security implications.
Abuse Cases
Abuse cases help security testers view the software under test in the same light as attackers do. Abuse cases capture the non-normative behavior of the system. While in [McGraw 04c] abuse cases are described more as a design analysis technique than as a white box testing technique, the same technique can be used to develop innovative and effective test cases mirroring the way attackers would view the system. With access to the source code, a tester is in a better position to quickly see where the weak spots are compared to an outside attacker. The abuse case can also be applied to interactions between components within the system to capture abnormal behavior, should a component misbehave. The technique can also be used to validate design decisions and assumptions. The simplest, most practical method for creating abuse cases is usually through a process of informed brainstorming, involving security, reliability, and subject matter expertise. Known attack patterns form a rich source for developing abuse cases.
Trust Boundaries Mapping
Defining zones of varying trust in an application helps identify vulnerable areas of communication and possible attack paths for security violations. Certain components of a system have trust relationships (sometimes implicit, sometime explicit) with other parts of the system. Some of these trust relationships offer ”trust elevation” possibilities—that is, these components can escalate trust privileges of a user when data or control flow cross internal boundaries from a region of less trust to a region of more trust [Hoglund 04]. For systems that have n-tier architecture or that rely on several third-party components, the potential for missing trust validation checks is high, so drawing trust boundaries becomes critical for such systems. Drawing clear boundaries of trust on component interactions and identifying data validation points (or chokepoints, as described in [Howard 02]) helps in validating those chokepoints and testing some of the design assumptions behind trust relationships. Combining trust zone mapping with data-flow analysis helps identify data that move from one trust zone to another and whether data checkpoints are sufficient to prevent trust elevation possibilities. This insight can be used to create effective test cases.
Code Coverage Analysis
Code coverage is an important type of test effectiveness measurement. Code coverage is a way of determining which code statements or paths have been exercised during testing. With respect to testing, coverage analysis helps in identifying areas of code not exercised by a set of test cases. Alternatively, coverage analysis can also help in identifying redundant test cases that do not increase coverage. During ad hoc testing (testing performed without adhering to any specific test approach or process), coverage analysis can greatly reduce the time to determine the code paths exercised and thus improve understanding of code behavior. There are various measures for coverage, such as path coverage, path testing, statement coverage, multiple condition coverage, and function coverage. When planning to use coverage analysis, establish the coverage measure and the minimum percentage of coverage required. Many tools are available for code coverage analysis. It is important to note that coverage analysis should be used to measure test coverage and should not be used to create tests. After performing coverage analysis, if certain code paths or statements were found to be not covered by the tests, the questions to ask are whether the code path should be covered and why the tests missed those paths. A risk-based approach should be employed to decide whether additional tests are required. Covering all the code paths or statements does not guarantee that the software does not have faults; however, the missed code paths or statements should definitely be inspected. One obvious risk is that unexercised code will include Trojan horse functionality, whereby seemingly innocuous code can carry out an attack. Less obvious (but more pervasive) is the risk that unexercised code has serious bugs that can be leveraged into a successful attack.
Classes of Tests
Creating security tests other than ones that directly map to security specifications is challenging, especially tests that intend to exercise the non-normative behavior of the system. When creating such tests, it is helpful to view the software under test from multiple angles, including the data the system is handling, the environment the system will be operating in, the users of the software (including software components), the options available to configure the system, and the error handling behavior of the system. There is an obvious interaction and overlap between the different views; however, treating each one with specific focus provides a unique perspective that is very helpful in developing effective tests.
Data
All input data should be untrusted until proven otherwise, and all data must be validated as it crosses the boundary between trusted and untrusted environments [Howard 02]. Data sensitivity/criticality plays a big role in data-based testing; however, this does not imply that other data can be ignored—non-sensitive data could allow a hacker to control a system. When creating tests, it is important to test and observe the validity of data at different points in the software. Tests based on data and data flow should explore incorrectly formed data and stressing the size of the data. The section ”Attacking with Data Mutation” describes different properties of data and how to mutate data based on given properties. To understand different attack patterns relevant to program input, refer to chapter six, ”Crafting (Malicious) Input,” . Tests should validate data from all channels, including web inputs, databases, networks, file systems, and environment variables. Risk analysis should guide the selection of tests and the data set to be exercised.
Environment
Software can only be considered secure if it behaves securely under all operating environments. The environment includes other systems, users, hardware, resources, networks, etc. A common cause of software field failure is miscommunication between the software and its environment. Understanding the environment in which the software operates, and the interactions between the software and its environment, helps in uncovering vulnerable areas of the system. Understanding dependency on external resources (memory, network bandwidth, databases, etc.) helps in exploring the behavior of the software under different stress conditions. Another common source of input to programs is environment variables. If the environment variables can be manipulated, then they can have security implications. In general, analyzing entities outside the direct control of the system provides good insights in developing tests to ensure the robustness of the software under test, given the dependencies.
Component Interfaces
Applications usually communicate with other software systems. Within an application, components interface with each other to provide services and exchange data. Common causes of failure at interfaces are misunderstanding of data usage, data lengths, data validation, assumptions, trust relationships, etc. Understanding the interfaces exposed by components is essential in exposing security bugs hidden in the interactions between components. The need for such understanding and testing becomes paramount when third-party software is used or when the source code is not available for a particular component. Another important benefit of understanding component interfaces is validation of principles of compartmentalization. The basic idea behind compartmentalization is to minimize the amount of damage that can be done to a system by breaking up the system into as few units as possible while still isolating code that has security privileges. Test cases can be developed to validate compartmentalization and to explore failure behavior of components in the event of security violations and how the failure affects other components.
Configuration
In many cases, software comes with various parameters set by default, possibly with no regard for security. Often, functional testing is performed only with the default settings, thus leaving sections of code related to non-default settings untested. Two main concerns with configuration parameters with respect to security are storing sensitive data in configuration files and configuration parameters changing the flow of execution paths. For example, user privileges, user roles, or user passwords are stored in the configuration files, which could be manipulated to elevate privilege, change roles, or access the system as a valid user. Configuration settings that change the path of execution could exercise vulnerable code sections that were not developed with security in mind. The change of flow also applies to cases where the settings are changed from one security level to another, where the code sections are developed with security in mind. For example, changing an endpoint from requiring authentication to not requiring authentication means the endpoint can be accessed by everyone. When a system has multiple configurable options, testing all combinations of configuration can be time consuming; however, with access to source code, a risk-based approach can help in selecting combinations that have higher probability in exposing security violations. In addition, coverage analysis should aid in determining gaps in test coverage of code paths.
Error handling
The most neglected code paths during the testing process are error handling routines. Error handling in this paper includes exception handling, error recovery, and fault tolerance routines. Functionality tests are normally geared towards validating requirements, which generally do not describe negative (or error) scenarios. Even when negative functional tests are created, they don’t test for non-normative behavior or extreme error conditions, which can have security implications. For example, functional stress testing is not performed with an objective to break the system to expose security vulnerability. Validating the error handling behavior of the system is critical during security testing, especially subjecting the system to unusual and unexpected error conditions. Unusual errors are those that have a low probability of occurrence during normal usage. Unexpected errors are those that are not explicitly specified in the design specification, and the developers did not think of handling the error. For example, a system call may throw an ”unable to load library” error, which may not be explicitly listed in the design documentation as an error to be handled. All aspects of error handling should be verified and validated, including error propagation, error observability, and error recovery. Error propagation is how the errors are propagated through the call chain. Error observability is how the error is identified and what parameters are passed as error messages. Error recovery is getting back to a state conforming to specifications. For example, return codes for errors may not be checked, leading to uninitialized variables and garbage data in buffers; if the memory is manipulated before causing a failure, the uninitialized memory may contain attacker-supplied data. Another common mistake to look for is when sensitive information is included as part of the error messages.
Tools
Source Code Analysis
Source code analysis is the process of checking source code for coding problems based on a fixed set of patterns or rules that might indicate possible security vulnerabilities. Static analysis tools scan the source code and automatically detect errors that typically pass through compilers and become latent problems. The strength of static analysis depends on the fixed set of patterns or rules used by the tool; static analysis does not find all security issues. The output of the static analysis tool still requires human evaluation. For white box testing, the results of the source code analysis provide a useful insight into the security posture of the application and aid in test case development. White box testing should verify that the potential vulnerabilities uncovered by the static tool do not lead to security violations. Some static analysis tools provide data-flow and control-flow analysis support, which are useful during test case development.
A detailed discussion about the analysis or the tools is outside the scope of this content area. This section addresses how source code analysis tools aid in white box testing. For a more detailed discussion on the analysis and the tools, refer to the Code Analysis Tools content area.
Program Understanding Tools
In general, white box testers should have access to the same tools, documentation, and environment as the developers and functional testers on the project do. In addition, tools that aid in program understanding, such as software visualization, code navigation, debugging, and disassembly tools, greatly enhance productivity during testing.
Coverage Analysis
Code coverage tools measure how thoroughly tests exercise programs. There are many different coverage measures, including statement coverage, branch coverage, and multiple-condition coverage. The coverage tool would modify the code to record the statements executed and which expressions evaluate which way (the true case or the false case of the expression). Modification is done either to the source code or to the executable the compiler generates. There are several commercial and freeware coverage tools available. Coverage tool selection should be based on the type of coverage measure selected, the language of the source code, the size of the program, and the ease of integration into the build process.
Profiling
Profiling allows testers to learn where the software under test is spending most of its time and the sequence of function calls as well. This information can show which pieces of the software are slower than expected and which functions are called more often than expected. From a security testing perspective, the knowledge of performance bottlenecks help uncover vulnerable areas that are not apparent during static analysis. The call graph produced by the profiling tool is helpful in program understanding. Certain profiling tools also detect memory leaks and memory access errors (potential sources of security violations). In general, the functional testing team or the development team should have access to the profiling tool, and the security testers should use the same tool to understand the dynamic behavior of the software under test.
Results to Expect
Any security testing method aims to ensure that the software under test meets the security goals of the system and is robust and resistant to malicious attacks. Security testing involves taking two diverse approaches: one, testing security mechanisms to ensure that their functionality is properly implemented; and two, performing risk-based security testing motivated by understanding and simulating the attacker’s approach. White box security testing follows both these approaches and uncovers programming and implementation errors. The types of errors uncovered during white box testing are several and are very context sensitive to the software under test. Some examples of errors uncovered include
* data inputs compromising security
* sensitive data being exposed to unauthorized users
* improper control flows compromising security
* incorrect implementations of security functionality
* unintended software behavior that has security implications
* design flaws not apparent from the design specification
* boundary limitations not apparent at the interface level
White box testing greatly enhances overall test effectiveness and test coverage. It can greatly improve productivity in uncovering bugs that are hard to find with black box testing or other testing methods alone.
Business Case
The goal of security testing is to ensure the robustness of the software under test, even in the presence of a malicious attack. The designers and the specification might outline a secure design, the developers might be diligent and write secure code, but it’s the testing process that determines whether the software is secure in the real world. Testing is an essential form of assurance. Testing is laborious, time consuming, and expensive, so the choice of testing (black box, or white box, or a combination) should be based on the risks to the system. Risk analysis provides the right context and information to make tradeoffs between time and effort to achieve test effectiveness.
White box testing is typically very effective in validating design decisions and assumptions and finding programming errors and implementation errors in software. For example, an application may use cryptography to secure data from specific threats, but an implementation error such as a poor key management technique can still leave the application vulnerable to security violations. White box testing can uncover such implementation errors.
Benefits
There are many benefits to white box testing, including the following:
* Analyzing source code and developing tests based on the implementation details enables testers to find programming errors quickly. For example, a white box tester looking at the implementation can quickly uncover a way, say, through an error handling mechanism, to expose secret data processed by a component. Finding such vulnerabilities through black box testing require comparatively more effort than found through white box testing. This increases the productivity of testing effort.
* Executing some (hard to set up) black box tests as white box tests reduces complexity in test setup and execution. For example, to drive a specific input into a component, buried inside the software, may require elaborate setup for black box testing but may be done more directly through white box testing by isolating the component and testing it on its own. This reduces the overall cost (in terms of time and effort) required to perform such tests.
* Validating design decisions and assumptions quickly through white box testing increases effectiveness. The design specification may outline a secure design, but the implementation may not exactly capture the design intent. For example, a design might outline the use of protected channels for data transfer between two components, but the implementation may be using an unprotected method for temporary storage before the data transfer. This increases the productivity of testing effort.
* Finding ”unintended” features can be quicker during white box testing. Security testing is not just about finding vulnerabilities in the intended functionality of the software but also about examining unintended functionality introduced during implementation. Having access to the source code improves understanding and uncovering the additional unintended behavior of the software. For example, a component may have additional functions exposed to support interactions with some other component, but the same functions may be used to expose protected data from a third component. Depending on the nature of the ”unintended” functionality, it may require a lot more effort to uncover such problems through black box testing alone.
Costs
The main cost drivers for white box testing are the following:
* specialized skill requirements: White box testing is knowledge intensive. White box testers should not only know how to analyze code for security issues but also understand different tools and techniques to test the software. Security testing is not just validating designed functionality but also proving that the defensive mechanisms work correctly. This requires invaluable experience and expertise. Testers who can perform such tasks are expensive and hard to get.
* support software development and tools: White box testing requires development of support software and tools to perform testing. Both the support software and the tools are largely based on the context of the software under test and the type of test technique employed. The type of tools used includes program understanding tools, coverage tools, fault injection tools, and source code analyzers.
* analysis and testing time: White box testing is time consuming, especially when applied to the whole system. Analyzing design and source code in detail for security testing is time consuming, but is an essential part of white box testing. Tools (source code analyzers, debuggers, etc.) and program understanding techniques (flow graphs, data-flow graphs, etc.) help in speeding up analysis. White box testing directly identifies implementation bugs, but whether the bugs can be exploited requires further analysis work. The consequences of failure help determine the amount of testing time and effort dedicated to certain areas.
Alternatives
There are alternatives to white box testing:
* White box testing can complement black box testing to increase overall test effectiveness. Based on risk assessment, certain areas of the software may require more scrutiny than others. White box testing could be performed for specific high-risk areas, and black box testing could be performed for the whole system. By complementing the two testing methods, more tests can be developed, focusing on both implementation issues and usage issues.
* Gray box testing can be used to combine both white box and black box testing methods in a powerful way. In a typical case, white box analysis is used to find vulnerable areas, and black box testing is then used to develop working attacks against these areas. The white box analysis increases productivity in finding vulnerable areas, while the black box testing method of driving data inputs decreases the cost of test setup and test execution.
All testing methods can reveal possible software risks and potential exploits. White box testing directly identifies more bugs in the software. White box testing is time consuming and expensive and requires specialized skills. As with any testing method, white box testing has benefits, associated costs, and alternatives. An effective testing approach balances efficiency and effectiveness in order to identify the greatest number of critical defects for the least cost.
Skills and Training
White box testing requires knowledge of software security design and coding practices, an understanding of an attacker’s mindset, knowledge of known attack patterns, vulnerabilities and threats, and the use of different testing tools and techniques. White box testing brings together the skills of a security developer, an attacker, and a tester.
Books like Writing Secure Code [Howard 02], 19 Deadly Sins [Howard 05], and Building Secure Software [Viega 02] help educate software professionals on how to write secure software, and books like How to Break Software Security [Whittaker 03] and Exploiting Software [Hoglund 04] help educate professionals on how to think like an attacker. There are several good books on software testing, including the two classics Software Testing Techniques [Beizer 90] and Testing Computer Software [Kaner 99]. There are several valuable information sites that detail known vulnerabilities, attack patterns, security tools, etc. White box testing is knowledge intensive and relies on expertise and experience.
Case Study
A large merchant organization involved in online business was in the process of developing an online e-commerce web site. In an effort to allow customers to electronically and efficiently transfer funds from customer checking accounts to merchant accounts, the merchant organization had outsourced its payment processing to a third-party Internet-enabled financial transaction payment firm. The third-party payments software provided customized interfaces to facilitate payment processing between the customers and the merchant organization.
A high-level security risk analysis was conducted on the system. Risk assessment identified transactions processing between the payments interface and the application as one of the risks. The impact of fraudulent transactions is serious for both the customers and for the merchant organization. The customers could suffer significant financial loss and hardships resulting from unauthorized transactions that could deplete account balances. The credibility and the reputation of the merchant organization could be severely damaged as a result of fraudulent transactions should they become publicized.
A thorough white box testing was conducted on the modules using the payments interface. First, all the component interfaces were identified and illustrated as interface diagrams; second, trust relationship boundaries were drawn on the component interactions; and third, data flows between the components were drawn. Abuse cases were developed based on this information. One of the abuse cases pointed to exercising the payments processing functionality as an anonymous user. The trust relationship mapping and data flow showed a path where inputs from the users were not validated or authenticated. A test case was developed to submit an account transfer from an external account to the merchant account anonymously and the account transfer was completed successfully, which is a critical software failure: unauthorized transactions via unauthenticated channel were allowed.
Risk assessment also identified a weak authentication component in the payment customer service component of the system. As in the case above, trust relationship boundaries and data-flow analysis were conducted on this component. After analysis and testing, it was shown that an attacker could gain direct access to the merchant administrative accounts. With this access, an attacker could redirect transactions from a merchant account to another, non-merchant account.
Using the two exploits described above, white box testers showed that an attacker could funnel payments from an unwitting customer to the merchant account and then redirect the transaction from the merchant account to a non-merchant account. The two bugs put together form a serious breach of security with significant business impact.
This example shows that risk analysis yields business-relevant results and points to specific vulnerable areas of the software. Subsequent to risk analysis, performing white box testing in concentrated areas aids in quickly uncovering design assumptions and implementation errors. This example also shows that finding an exploit does not mean that the job is done. Collecting information about various bugs in the software and analyzing them together shows how multiple exploits can be stringed together to form a full attack.
Conclusion
White box testing for security is useful and effective. It should follow a risk-based approach to balance the testing effort with consequences of software failure. Architectural and design-level risk analysis provide the right context to plan and perform white box testing. White box testing can be used with black box testing to improve overall test effectiveness. It uncovers programming and implementation errors.
