Previous column

Next column

---

Security Use Cases Donald G. Firesmith, Software Engineering Institute, U.S.A.

COLUMN PDF Version

Abstract

Although use cases are a popular modeling approach for engineering functional requirements, they are often misused when it comes to engineering security requirements because requirements engineers unnecessarily specify security architectural mechanisms instead of security requirements. After discussing the relationships between misuse cases, security use cases, and security mechanisms, this column provides examples and guidelines for properly specifying essential (i.e., requirements-level) security use cases.

1 INTRODUCTION

Over the past decade, use cases have become one of the most popular modeling approaches for analyzing and specifying functional requirements. However, use case modeling has not been as successfully applied to engineering quality requirements, such as operational availability, performance, portability, reliability, reuse, security, and usability. When it comes to engineering security requirements, use cases are typically misused to unnecessarily specify security architectural mechanisms (e.g., the use of user identifiers and passwords) rather than actual security requirements (e.g., mandating some level of identification and authentication). Thus, typical example use cases for an automatic teller machine might include initial interactions for inserting an ATM card to identify the customer and entering a PIN number for authentication (i.e., verifying the identity of the customer). Whereas this is the current standard security mechanism for implementing identification and authentication requirements for ATM machines, it unnecessarily prevents the use of other, perhaps improved means of access control such as biometrics (e.g., face recognition, fingerprint analysis, or retinal scan).

Security requirements should be based on an analysis of the assets and services to be protected and the security threats from which these assets and services should be protected. Thus, as illustrated in Figure 1, there are clear relationships between assets and services, which are vulnerable to security threats, which necessitate security requirements, which require security mechanisms that counter these security threats and thereby protect the assets and services.

Fig. 1: Security Threats, Requirements, and Mechanisms

Historically, the emphasis of security engineering has been on the development and use of numerous security mechanisms to protect vulnerable assets and services by countering known security threats. The analysis and documentation of security threats and security requirements has received considerably less attention.

Misuse Cases for the Analysis of Security Threats

A relatively recent approach to addressing security threat analysis has been the development of misuse cases. As illustrated in Figure 2, misuse cases (a.k.a., abuse cases) are a specialized kind of use cases that are used to analyze and specify security threats [Sindre and Opdahl 2001] [Alexander2003]. Unlike normal use cases that document interactions between an application and its users, misuse cases concentrate on interactions between the application and its misusers (e.g., cracker or disgruntled employee) who seek to violate its security. Because the success criteria for a misuse case is a successful attack against an application, misuse cases are highly effective ways of analyzing security threats but are inappropriate for the analysis and specification of security requirements. Instead, security use cases should be used to specify requirements that the application shall successfully protect itself from its relevant security threats.

Fig. 2: Misuse Cases vs. Security Use Cases

The following table summarizes the primary differences between misuse cases and security use cases.

To further illustrate the differences between normal use cases, security use cases, and associated misuse cases, consider Figure 3. The traditional use cases for an automated teller machine might include Deposit Funds, Withdraw Funds, Transfer Funds, and Query Balance, all of which are specializations of a general Manage Accounts use case. To securely manage one’s accounts, one can specify security use cases to control access (identification, authentication, and authorization), ensure privacy (of data and communications), ensure integrity (of data and communications), and ensure nonrepudiation of transactions. The resulting four security use cases specify requirements that protect the ATM and its users from three security threats involving attacks by either crackers or thiefs.

Fig. 3: Example Security Use Cases and Misuse Cases

2 EXAMPLE SECURITY USE CASES

As documented in [Firesmith 2003], there are numerous kinds of security requirements. Although each kind of security requirement typically has its own security use case, given the limited space available in this column, I have selected access control (identification and authentication), integrity, and privacy to illustrate the proper use of security use cases. To maximize the reusability of the following use case path specifications, I have also kept them at the highest, most-generic level of abstraction (i.e., as paths through essential use cases). When reused on real projects, each path specification can easily be made more specific to the application being specified without devolving into an architecture or design level specification, often merely by replacing the general words “system” and “user” with the specific application name and the specific type of user.

Access Control Use Case

Access control is the extent to which a a business enterprise, application, component, or center controls access by its externals (e.g., human users and applications). Access control consists of identification, authentication (i.e., verification of identification), and authorization. The following three tables document example use case paths through a highly-reusable essential security use case that specifies access control requirements:

• Use Case Path - Attempted Spoofing using Valid User Identity
• Use Case Path - Attempted Identity and Authentication Theft
• Use Case Path - Attempted Spoofing using Social Engineering

Integrity Use Case

Integrity is the extent to which a business enterprise, application, component, or center ensures that its data and communications are not intentionally corrupted via unauthorized creation, modification, or deletion. The following three tables document example use case paths through a highly-reusable essential security use case that specifies integrity requirements:

• Use Case Path - System Data Protected
• Use Case Path - System Data Corrupted
• Use Case Path - System Message Integrity
• Use Case Path - Use Message Integrity
• Use Case Pase - Denial Of Service (DOS) Attack

Privacy Use Case

Privacy is the extent to which a business enterprise, application, component, or center keep its sensitive data and communications private from unauthorized individuals and programs. The following three tables document example use case paths through a highly-reusable essential security use case that specify privacy requirements:

• Use Case Path - Data Privacy
• Use Case Path - System Message Privacy
• Use Case Path - User Message Privacy

3 SECURITY USE CASE GUIDELINES

The following guidelines are recommended when developing security use cases during requirements engineering:

• Use essential (i.e., requirements only) use cases that do not specify unnecessary security architectural mechanisms such as user IDs, passwords, digital signatures, biometrics, encryption, etc. Leave such decisions to the architecture and security teams who are better qualified than the requirements team to make such decisions.
• Carefully differentiate requirements (e.g., by using the word “shall”) from ancellary information. System interactions, system actions, and the postconditions should be specified as requirements on the system, whereas preconditions, user interactions, and misuser interactions should not be.
• To avoid unnecessarily specifying design constraints, the clearly note if the sequencing of the interactions can occur in different orders.
• Explicitly document the individual paths through the security use cases in order to specify the actual security requirements.
• Base the security use cases on the different types of security requirements, which provide a natural organization to the use cases.
• Document the security threats that justify the individual paths through the security use case.
• Clearly distinquish between user and misuser interactions.
• Clearly distinguish between externally-visible system interactions and hidden system actions.
• Document both preconditions and postconditions, which capture the essence of the individual path.

4 CONCLUSION

Whereas misuse cases are an excellent means of analyzing security threats, they are inappropriate for analyzing and specifying security requirements because they are based on misusers successfully attacking the system. On the other hand, essential security use cases provide a highly-reusable way of organizing, analyzing, and specifying security requirements if they are kept at the appropriate level and if unnecessary architectural and design constraints are avoided.

REFERENCES

[Alexander2003] Ian Alexander: Misuse Case Help To Elicit Nonfunctional Requirements, IEE CCEJ, 2001, http://easyweb.easynet.co.uk/~iany/consultancy/papers.htm.

[Firesmith2003] Donald Firesmith: OPEN Process Framework (OPF) Website, http://www.donald-firesmith.com.

[Sindre and Opdahl 2001] Guttorm Sindre and Andreas Opdahl: Templates for Misuse Case Description, 2001, http://www.ifi.uib.no/conf/refsq2001/papers/p25.pdf.

ACKNOWLEDGEMENTS

The contents of this article are taken from my informational website on process engineering as well as the contents of my upcoming book on requirements engineering. Guttorm Sindre and Andreas Opdahl provided very valuable insights regarding the relationship between misuse cases and security use cases.

About the author

Donald Firesmith is a senior member of the technical staff at the Software Engineering Institute. He has worked exclusively with object technology since 1984 and has written 5 books on the subject. He is currently writing a book on requirements engineering. Most recently, he has developed a 1000+ page informational website on the OPEN Process Framework at http://www.donald-firesmith.com. He can be reached at donald_firesmith@hotmail.com.

Cite this column as follows: Donald Firesmith: "Security Use Cases", in Journal of Object Technology, vol. 2, no. 3, May-June 2003, pp. 53-64. http://www.jot.fm/issues/issue_2003_05/column6