Secure applications do not just happen – they are the result of an organization deciding that they will produce secure applications. OWASP does not wish to mandate a particular approach or require an organization to pick up compliance with laws that do not affect them - every organization is different.
However, for a secure application, the following at a minimum are required:
- Organizational management which champions security
- A written information security policy properly derived from national standards
- A development methodology with adequate security checkpoints and activities
- Secure release and configuration management processes
Many of the controls within OWASP Development Guide 2.0 are influenced by requirements of national standards or control frameworks such as COBIT; typically controls selected out of OWASP will satisfy relevant ISO 27002(17799) and COBIT controls.
Organizational commitment to security
Organizations that have security buy-in from the highest levels will generally produce and procure applications that meet basic information security principles. This is the first of many steps along the range between ad hoc “possibly secure (but probably not)” to “best practices”.
Organizations that do not have management buy-in, or simply do not care about security, are extraordinarily unlikely to produce secure applications. Each secure organization documents its “taste” for risk in their information security policy, thus making it easy to determine which risks will be accepted, mitigated, or assigned.
Insecure organizations simply don’t know where this “taste” is set, and so when projects run by the insecure organization select controls, they will either end up implementing the wrong controls or not enough controls. Rare examples have been found where every control, including a kitchen sink tealeaf strainer has been implemented, usually at huge cost.
Most organizations produce information security policies derived from ISO 27002(17799), or if in the US, from COBIT, or occasionally both or other standards. There is no hard and fast rule for how to produce information security policies, but in general:
- If you’re publicly traded in most countries, you must have an information security policy
- If you’re publicly traded in the US, you must have an information security policy which is compliant with SOX requirements, which generally means COBIT controls
- If you’re privately held, but have more than a few employees or coders, you probably need one
- Popular FOSS projects, which are not typical organizations, should also have an information security policy
It is perfectly fine to mix and match controls from COBIT and ISO 27002(17799) and most any other respected information security standard; rarely do they disagree on the details. The method of production is sometimes tricky – if you “need” certified policy, you will need to engage qualified firms to help you.
OWASP’s Place at the Framework table
Organizations need to establish information security policy informed by relevant national legislation, industry regulation, merchant agreements, and subsidiary best practice guides, such as OWASP. It is impossible to draw a small diagram containing all relevant laws and regulations, so you should assume all of the relevant laws, standards, regulations, and guidelines are missing – you need to find out which affect your organization, customers (as applicable), and where the application is deployed.
IANAL: OWASP is not a qualified source of legal advice; you should seek your own legal advice.
COBIT is a popular risk management framework structured around four domains:
- Planning and organization
- Acquisition and implementation
- Delivery and support
Each of the four domains has 13 high level objectives, such as DS5 Ensure Systems Security. Each high level objective has a number of detailed objectives, such as 5.2 Identification, Authentication, and Access. Objectives can be fulfilled in a variety of methods that are likely to be different for each organization.
COBIT is typically used as a SOX control framework, or as a complement to ISO 27002(17799) controls.
OWASP does not dwell on the management and business risk aspects of COBIT. If you are implementing COBIT, OWASP is an excellent start for systems development risks and to ensure that custom and off-the-shelf applications comply with COBIT controls, but OWASP is not a COBIT compliance magic wand.
Where a COBIT objective is achieved with an OWASP control in this book, you will see “COBIT XXy z.z” to help direct you to the relevant portion of COBIT control documentation. Such controls should be a part of all COBIT compliant applications.
For more information about COBIT, please visit http://www.isaca.org/
ISO 27002(17799) is a risk-based Information Security Management framework directly derived from the AS/NZS 4444 and BS 7799 standards. It is an international standard and used heavily in most organizations, although not in the US. However, a few US organizations use ISO 27002(17799) as well, particularly if they have subsidiaries outside the US.
ISO 27002(17799) dates back to the mid-1990s, and some of the control objectives reflect this age – for example referring to administrative interfaces as “diagnostic ports”.
Organizations using ISO 27002(17799) can use OWASP for detailed guidance when selecting and implementing a wide range of ISO 17799 controls, particularly those in the Systems Development chapter, among others. Where a 27002(17799) objective is achieved with an OWASP control in this book, you will see “ISO 27002(17799) X.y.z” to help direct you to the relevant portion of ISO 27002(17799). Such controls should be a part of all ISO 27002(17799) compliant applications.
For more information about ISO 27002(17799), please visit http://www.computersecuritynow.com/ and the relevant standards bodies, such as Standards Australia (http://www.standards.com.au/), Standards New Zealand (http://www.standards.co.nz/), or British Standards International (http://www.bsi-global.com/).
Note that ISO 17799 was renamed to ISO 27002 in July 2007, to achieve numerical alignment with ISO's overall security standards set (ref: http://www.27000.org)
A primary motivator for many US organizations in adopting OWASP controls is to assist with ongoing Sarbanes-Oxley compliance. If an organization followed every control in this book, it would not automatically grant the organization SOX compliance. Therefore, The Guide is useful as a suitable control for application procurement and in-house development, as part of a wider compliance program.
However, SOX compliance is often used as a case for resource-starved IT managers to implement long neglected security controls, so it is important to understand what SOX actually requires. A summary of SOX, section 404 obtained from AICPA’s web site at http://www.aicpa.org/info/sarbanes_oxley_summary.htm states:
Section 404: Management Assessment of Internal Controls
Requires each annual report of an issuer to contain an "internal control report", which shall:
- state the responsibility of management for establishing and maintaining an adequate internal control structure and procedures for financial reporting; and
- contain an assessment, as of the end of the issuer's fiscal year, of the effectiveness of the internal control structure and procedures of the issuer for financial reporting.
This essentially states that management must establish and maintain internal financial control structures and procedures, and conduct an annual evaluation that verifies the controls are effective. As finance is no longer conducted using double entry ledger books, “SOX compliance” is often extended to mean IT.
The Development Guide can facilitate SOX compliance by providing effective controls for all applications, and not just for the purposes of financial reporting. It allows organizations to buy products which claim they use OWASP controls, or allows organizations to mandate to software suppliers that they must use OWASP controls to provide more secure software.
However, avoid using SOX as an excuse. SOX controls are intended to prevent another Enron, not to buy widgets that may or may not help. All controls, whether off the shelf widgets, training, code controls, or process changes, should be selected based on measurable results and ability to manage risk, and not just to “tick the boxes”.
High performing development shops employ a development methodology and some set of coding standards or conventions. As it turns out, the choice of development methodology is not as important as simply having one.
Ad hoc development is too unstructured to produce secure applications. Therefore, organizations who wish to produce secure code consistently need to utilize a methodology that supports that goal. Choose carefully – small teams should never consider heavy weight methodologies that identify many different roles, while large teams must choose methodologies that will scale to their needs.
Here are some key attributes to look for in selecting a development methodology:
- Strong acceptance of design, testing, and documentation processes
- Clear instances where security controls (such as threat risk analysis, peer reviews, code reviews, etc) can be slotted in
- Works well for the organization’s size and maturity
- Has the potential to reduce the current error rate and improve developer productivity
- Will scale as the organization or product line grows
Methodologies alone are not coding standards; each team will either need to determine what to use based upon common practice, or simply lay down the law based upon known best practices.
Inputs to consider:
- Architectural guidance (i.e., “The web tier cannot call the database directly”)
- Minimum documentation levels required
- Mandatory testing and coverage requirements
- Minimum levels of code commenting and the preferred comment style
- Proper use of exception handling
- Correct use of flow of control blocks (e.g., “All uses of conditional flow shall use explicit statement blocks”)
- Preferred variable, function, class, and table naming styles
- A preference for maintainable and readable code over clever or complex code
Indentation style and tabbing are a holy war, and from a security perspective, they simply do not matter that much. However, it should be noted that we no longer use 80x24 terminals, so vertical space usage is not as important as it once was. Indent and tabbing can be “fixed” using automated tools or simply a style within a code editor, so do not get overly fussy on this issue.
Source Code Control
High performance software engineering requires the use of regular improvements to code, along with associated testing regimes. All code and test changes must be able to be versioned and capable of being reverted.
This could be done by copying folders on a file server, but it is better performed by source code revision tools, such as Subversion, CVS, SourceSafe, or ClearCase.
Why include tests with a software revision? Most simply put, because tests for later builds will not necessarily match the tests required for earlier builds. So, it is vital that a test is applied to the build for which it was intended.
The use of policy frameworks does not automatically guarantee secure applications or standards compliance. However, it is very difficult to produce secure applications consistently without some structure in place to do so.
Select your policy framework carefully -- it should meet the needs of your organization today, while providing room for growth, too.
Finally, get started today! Incorporating a security-conscious development process is a crucial first step to delivering secure applications. Your policy framework and development process should leverage your local conventions, risk management goals, and applicable standards to ensure a secure and quality result.
[http://www.nsa.gov/snac/downloads_all.cfm?doc=%22Web%20Application%20Securit y%20Overview.pdf%2 NSA]