The PCI SSC (Council) is responsible for managing the security standards, while compliance with the PCI set of standards is enforced by the founding members of the Council: American Express, Discover Financial Services, JCB International, MasterCard Worldwide and Visa Inc. Non-compliant companies who maintain a relationship with one or more of the card brands, either directly or through an acquirer risk losing their ability to process credit card payments and being audited and/or fined.

All in-scope companies must validate their compliance annually. This validation can be conducted by Qualified Security Assessors, i.e. companies that have completed a three-step certification process by the PCI SSC which recognises them as being qualified to assess compliance to the PCI DSS standard. However, smaller companies have the option to use a Self-Assessment Questionnaire. Whether this questionnaire needs to be validated by a QSA depends on the requirements of the card brands in that merchant’s region.

The current version of the standard specifies 12 requirements for compliance, organised into 6 logically related groups, which are called “control objectives.”

1. Build and Maintain a Secure Network
   • Requirement 1: Install and maintain a firewall configuration to protect cardholder data
   • Requirement 2: Do not use vendor-supplied defaults for system passwords and other security parameters
2. Protect Cardholder Data
   • Requirement 3: Protect stored cardholder data
   • Requirement 4: Encrypt transmission of cardholder data across open, public networks
3. Maintain a Vulnerability Management Program
   • Requirement 5: Use and regularly update anti-virus software
   • Requirement 6: Develop and maintain secure systems and applications
4. Implement Strong Access Control Measures
   • Requirement 7: Restrict access to cardholder data by business need-to-know
   • Requirement 8: Assign a unique ID to each person with computer access
   • Requirement 9: Restrict physical access to cardholder data
5. Regularly Monitor and Test Networks
   • Requirement 10: Track and monitor all access to network resources and cardholder data
   • Requirement 11: Regularly test security systems and processes
6. Maintain an Information Security Policy
   • Requirement 12: Maintain a policy that addresses information security

Compliance with these requirements can be summarized into 3 main stages:

• Collecting and storing: Secure collection and tamper-proof storage of all log data so that it is available for analysis.
• Reporting: Being able to prove compliance on the spot if audited and present evidence that controls are in place for protecting data.
• Monitoring and alerting: Have systems in place such as auto-alerting, to help administrators constantly monitor access and usage of data. Administrators are warned of problems immediately and can rapidly address them. These systems should also extend to the log data itself –- there must be proof that log data is being collected and stored.

What does this actually mean for web application developers?

It is considerably more expensive and more time-consuming to recover from a security incident than to take preventative measures ahead of time. If you follow the guidelines below, you will go along way to securing you application in line with the PCI DSS regulations. Many of the measures apply to general application security, but since PCI DSS is all about security, they are worth mentioning.

Server-level Security:

• Separate web- and database-servers on to different physical machines.
• Secure the web- and database-servers with traditional techniques. Only authorised accounts should have the capabilities to run tasks on the machine. That means not giving admin-rights to the user account.
• Keep servers up-to-date with the latest patches and software releases.
• Minimise the number of services running on the server. This means limiting the services to only those required for the web- or database-servers to function.
• Secure information in transit between servers. This may mean physically securing the network to prevent evesdropping via encryption or obfuscating the data amongst innocuous ‘noise’.
• Secure the database server behind a firewall.

Application-level Security:

• Separate ColdFusion, the webserver and database server user accounts. They should never be under the same system account.
• Create a database user specifically for your ColdFusion datasource and restrict it to only the activities required for the application. The user should not have database-owner rights, access to databases not relating to the application or access to the system tables.
• Revoke privileges in the ColdFusion datasource definition to prevent the SQL commands CREATE, DROP, GRANT, REVOKE and ALTER.
• General settings in the ColdFusion Administrator:
  • Check the Disable access to internal ColdFusion Java components option.
  • Check the Enable Global Script Protection option.
  • Add a Missing Template Handler.
  • Add a Site-wide Error Handler.
  • Reduce the Maximum size of post data from 100MB.
  • Enable Timeout Requests, and set to 60 seconds or less.
  • Disable Robust Exception Handling on production servers.

Web Application-level Security:

• Use secure HTTP to transfer data and/or when logged into ‘administration’ secutions of your web application.
• Timeout sessions after 15 minutes and on browser close.
• Provide multi-level login processes. For example, lock the application after 3 failed attempts for a period of 10 minutes.
• Do not identify whether the username or password are incorrect, simply notify the user that their login failed and that they must try again.
• Encrypt passwords stored in the database with a standard such as SHA-256 or ‘stronger’.
• Use CAPTCHAs (textual and aural) to prevent automated robots hacking into your application.
• Run regular penetration tests on your application to identify potential problems.
• Encrypt credit card information held in the database or other storage mechanism. Only store credit card data in line with the PCI DSS regulations.

Code-level Security:

• Application.cfc – Set the scriptProtect Application variable to true to enable application-wide cross-site script protection.
• CFQueryParam – This tag, importantly, verifies the data type of a query parameter and, for RDBMSs that support bind variables, enables ColdFusion to use bind variables in the SQL statement. Bind variable usage enhances performance when executing a cfquery statement multiple times. There are limitations to the use of the cfqueryparam tag. In ColdFusion 7 for example, you cannot use them in queries using the cachedWithin attribute. Similarly, they cannot be used in ORDER BY clauses, although the use of conditional logic should resolve the need for order by variables.
• Functions – As a rule of thumb, validate all the data being passed into a query prior to it being used. ColdFusion MX 7 saw the introduction of the isValid() function. This function tests whether a value meets a validation or data type rule and can be used to replace a large number of type-specific functions such as isArray(), isBinary(), isBoolean(), isDate(), isNumeric() and isSimpleValue() etc.
• Stored Procedures – I often favour the use of stored procedures over standard queries. Not only do they add an additional level of performance, they provide an additional level of security; ColdFusion does not do any raw processing of queries in the web code, it simply passes variables down the wire to the database server.

Conclusion

The goal of the PCI Data Security Standard is to protect cardholder data that is processed, stored or transmitted by merchants. The security controls and processes required by PCI DSS are vital for protecting cardholder account data, including the PAN – the primary account number printed on the front of a payment card. Merchants and any other service providers involved with payment card processing must never store sensitive authentication data after authorisation. This includes sensitive data that is printed on a card, or stored on a card’s magnetic stripe or chip – and personal identification numbers entered by the cardholder.

By following the points made above, you will go a long way to meeting the PCI DSS guidelines, whilst also securing your infrastructure and applications in a more general sense.

Caveat: The views and comments written in this article are provided as a guideline. I hold no responsibility for the security of your applications and data based upon the information provided.

Take this scenario. A company I once worked for had what can be described as a business directory built upon a licensed, yet bastardised, version of a popular ColdFusion-based CMS. It didn’t work that well! The decision was made to redevelop the application in Java. It took two years to reach the same level of functionality! What happened next? Ruby-on-Rails is what! The rest is history and beyond the topic of this post.

So, in effect, the application almost went full-circle in its development paradigm — both ColdFusion and Ruby-on-Rails can be considered Rapid Application Development environments, Java, certainly not. Why did the decision makers not stick with ColdFusion and put time aside to actually build it properly in the first place? To put it simply, they lost faith in ColdFusion; it was largely mis-understood.

The weakness of every programming language does not lie with the language itself per se — albeit it can have an important influencing factor — but rather with the ability, or indeed inability, of the developer to leverage the language in the most efficient and optimal way.

ColdFusion, like every other programming language has had and I’m sure still does have its fair share of poor developers; those people simply working with it as a means-to-an-end, rather than those passionate about the language, those people programming without understanding the fundamentals of programming or the implications of their poorly written code. This is apparent from .NET to Java, ColdFusion to Ruby, JavaScript to ActionScript.

Let’s not dilly-dally, bicker or insult one another about which is best, which one is dying and which one is not worth the computer it is compiled on. What is important is to understand the merits of each language and decide which one best suits the application, not only in technical terms, but also in terms of time-to-market, cost of development, availability of a skilled workforce etc.

ColdFusion, whether rightly or wrongly in some people’s opinion, can sit proudly amongst its peers and provide a truly compelling alternative.

Here’s how (in no particular order):

1. Low Total Cost of Ownership – frequently, ColdFusion is described as expensive, it simply isn’t especially if you consider the natively supported functions. But to put it bluntly, if your company cannot afford the cost of ColdFusion standard, or indeed ColdFusion hosting, you have bigger things to worry about regarding the profitability of the company; you won’t be able to afford much of anything! The problem becomes not the product. ColdFusion applications are quicker to develop and developers are vastly cheaper to employ than their peers in Java or Ruby, just look at ITJobsWatch for examples.
2. Rapid Application Development – ColdFusion vastly simplifies tasks. What would take other languages numerous lines of code to produce is efficiently encapsulated either in a tag or function or as a setting in the administrator. This is a simplistic yet indicative example: where else can you connected to a database simply with one line of code or indeed simply by name? ColdFusion changed the idea of specifying development time in terms of months and years to weeks and months or small features a matter of hours and days. Simplicity is not the mother or all evil. To be pragmatic, simplification reduces costs.
3. Rich Internet Applications – ColdFusion may or may not have pioneered the RIA paradigm, but it has played a significant supporting role to Flash and now Flex. ColdFusion natively supports Flash remoting, providing the all important data access tier.
4. Platform Maturity – ColdFusion 8 is built upon the latest version of Java (1.6). Along with internal improvements to the ColdFusion application, this has afforded ColdFusion unprecedented speed improvements and stability.
5. Language Maturity – with each major release of ColdFusion comes many language enhancements added to the core. This means that previous addons, for example image manipulation, which came at a premium are now standard. Adobe and other companies that produce CFML engines are now participating in a CFML advisory committee, which aims to set standards for the core language. This is not only a sign of maturity but a letter of intent by the industry that will mean your application will work on any engine, assuming no proprietary functionality is used.
6. The Ultimate Middleware – ColdFusion sits comfortably between any backend and front end system. Be it interfacing with a host of databases, Java, .NET, COM, Corba or connecting to classic HTML or rich Flash, Flex and AJAX frontends with little or no configuration.
7. Feature Rich – what other web technology natively supports PDF generation, charting, enterprise-level search, AJAX, image manipulation, Atom and RSS creation, Zip and JAR file manipulation, a server monitor, Flex integration, encryption libraries, all important database connectors, webservice creation, XML manipulation, inbuilt reporting application (similar to Crystal Reports), email, FTP to name but a few? I hazard a guess at none, unless you’re happy to pay a premium.
8. Platform Independent – since ColdFusion 6, when Macromedia redeveloped the entire application in Java, ColdFusion has been platform independent. You can install it on practically any machine.
9. OpenSource Alternatives – BlueDragon and Railo are both significant alternatives to Adobe ColdFusion and both have opensource alternatives, the latter of the two having recently joined the JBoss community. Adobe are also considering providing a free edition to academic institutions.
10. The Future – many commentators have mentioned Hibernate as a significant addition to the next release of ColdFusion, version 9. But having seen the prerelease notes, that is not all that will be added. Alas I’m under NDA, but rest assured, there is going to be a significant intake of breath when developers get hold of the next release. ColdFusion 8 was firmly geared towards middle management with fuzzy additions, ColdFusion 9 is set to re-address the balance with compelling language and functionality enhancements.

ColdFusion evangelism needs to step up a gear! Adobe certainly doesn’t afford much marketing budget to the product, prefering The Community do the hard work. It is not always easy convincing the decision makers that ColdFusion is a good product of choice, without Adobe’s unnerving support, but we have to work hard, break down those barriers, encroach on events outside the comfortable sphere of the ColdFusion world and demonstrate ColdFusion’s match-winning ability.

ColdFusion isn’t dying, it’s simply niche. Every niche has its place.

UPDATE: If you would like to view the ensuing debate regarding ColdFusion prompted by Aral Balkan, feel free to do so. This post should serve as a positive reminder of ColdFusion’s virtues, alongside the need for a balanced and polite debate.

But what is a User Agent? A User Agent is the client application used with a particular network protocol; the phrase is most commonly used in reference to those which access the Web. Web user agents range from web browsers and e-mail clients to search engine crawlers (spiders), as well as mobile phones, screen readers and braille browsers used by people with disabilities. When Internet users visit a web site, a text string is generally sent to identify the user agent to the server. This forms part of the HTTP request, prefixed with user-agent: and typically includes information such as the application name, version, host operating system, and language. Bots, such as web crawlers, often also include a URL and/or e-mail address so that the webmaster can contact the operator of the bot.

By simply typing about:version into Chrome’s address bar you will be presented with the following information:

Google Chrome
0.2.149.29 (1798)
Official Build
Google Inc.
Copyright © 2006-2008 Google Inc. All Rights Reserved.
Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US) AppleWebKit/525.13 (KHTML, like Gecko) Chrome/0.2.149.29 Safari/525.13

As you can see Chrome’s version information provides limited detail about the browser. The last line is the important one. It is the HTTP User-Agent header:

Mozilla/5.0 (Windows; U; Windows NT 6.0; en-US) AppleWebKit/525.13 (KHTML, like Gecko) Chrome/0.2.149.29 Safari/525.13.

If you know the RFC 2616 specification on the HyperText Transfer Protocol — which incidentally, I gladly don’t — you would know that the User Agent, or more formally, product token, should be short and to the point:

Product tokens SHOULD be short and to the point. They MUST NOT be used for advertising or other non-essential information. Although any token character MAY appear in a product-version, this token SHOULD only be used for a version identifier (i.e., successive versions of the same product SHOULD only differ in the product-version portion of the product value).

Clearly this isn’t the case! One of Google’s reason’s behind creating the Chrome browser was to start afresh. It would have therefore been truely amazing if they had made the string simply Chrome/0.2.149.27.

Unfortunately, browser sniffing makes an ever-growing UA string the path of least resistance for browser vendors.

So, what does Chrome’s User Agent string actually mean:

• Mozilla/ – This means that browser has the kind of capabilities that Netscape 1.1 had compared to Mosaic and Lynx.
• 5.0 – This means that the browser engine is from the post-Browser War Web Standards era as opposed to being from the Browser War era.
• (Windows; – This means that general windowing system flavor the browser runs on is Windows (as opposed to, for example, Apple and X11).
• U; – This means that the browser has at least the level of cryptographic capability / encryption strength that U.S. versions of browsers had in the late 1990s.
• Windows NT 6.0; – This indicates the operating system the browser is running on. In this instance, the browser is running on Vista.
• en-US) – This indicates the user interface language of the browser (U.S. English in this case). This may be used to choose between different content languages even though HTTP has a different header for that purpose.
• AppleWebKit/ – This indicates that the engine of the browser is WebKit as opposed to being Gecko. Developers should not do user agent sniffing as a rule, but if they still do, this is what they should be sniffing.
• 525.13 – This is the WebKit version from which Chrome branched its copy. Site admins could use this to detect old versions with known bugs.
• (KHTML, like Gecko) – This introduces the substring Gecko into the UA string while pointing out to human readers that Webkit was forked from KHTML. Without this substring, Chrome might be put in the same category as IE and Netscape 4.
• Chrome/ – This string identifies the browser as actually Google Chrome.
• 0.2.149.27 – This is the Chrome version. This could be used to detect old versions with known bugs.
• Safari/ – This means that the browser is like Safari as opposed to being like Firefox.
• 525.13 – This just repeats the WebKit version in order to have some version but not the irrelevant Safari.app version.

Why should you be worried about security?

The Web is changing many of the assumptions that people have historically made about computer security and publishing. As the Internet makes it possible for web servers to publish information to millions of users, it also makes it possible for computer hackers, crackers, criminals, vandals, and other “bad guys” to break into the very computers on which the web servers are running. Once subverted, web servers can be used by attackers as a launching point for conducting further attacks against users and organisations.

It is considerably more expensive and more time-consuming to recover from a security incident than to take preventative measures ahead of time.

This blog post started on the premise of protecting your website from a SQL Injection Attack. However, it is also appropriate to discuss, at a relatively high level, how to secure your server architecture and applications.

Server-Level Security

• Separate web- and database-servers on to different physical machines.
• Secure the web- and database-servers with traditional techniques. Only authorised accounts should have the capabilities to run tasks on the machine. That means not giving admin-rights to the user account.
• Keep servers up-to-date with the latest patches and software releases.
• Minimise the number of services running on the server. This means limiting the services to only those required for the web- or database-servers to function.
• Secure information in transit between servers. This may mean physically securing the network to prevent evesdropping via encryption or obfuscating the data amongst innocuous ‘noise’.
• Secure the database server behind a firewall.

Application-Level Security

• Separate ColdFusion, the webserver and database server user accounts. They should never be under the same system account.
• Create a database user specifically for your ColdFusion datasource and restrict it to only the activities required for the application. The user should not have database-owner rights, access to databases not relating to the application or access to the system tables.
• Revoke privileges in the ColdFusion datasource definition to prevent the SQL commands CREATE, DROP, GRANT, REVOKE and ALTER.
• General settings in the ColdFusion Administrator:
  • Check the Disable access to internal ColdFusion Java components option.
  • Check the Enable Global Script Protection option.
  • Add a Missing Template Handler.
  • Add a Site-wide Error Handler.
  • Reduce the Maximum size of post data from 100MB.
  • Enable Timeout Requests, and set to 60 seconds or less.
  • Disable Robust Exception Handling on production servers.

Code-Level Security

• Application.cfc – Set the scriptProtect Application variable to true to enable application-wide cross-site script protection.
• CFQueryParam – This tag, importantly, verifies the data type of a query parameter and, for RDBMSs that support bind variables, enables ColdFusion to use bind variables in the SQL statement. Bind variable usage enhances performance when executing a cfquery statement multiple times.

  <cfquery name="qry" datasource="#APPLICATION.dsn#">
  SELECT column1, column2, column3
  FROM tableName
  WHERE column4 = <cfqueryparam value="#variable1#" cfsqltype="cf_sql_bit" />
  AND column5 LIKE <cfqueryparam value="%#variable2#%" cfsqltype="cf_sql_varchar" maxlength="200" />
  AND column6 IN (<cfqueryparam value="#variable3#" cfsqltype="cf_sql_integer" list="true" />)
  </cfquery>

  There are limitations to the use of the cfqueryparam tag. In ColdFusion 7 for example, you cannot use them in queries using the cachedWithin attribute. Similarly, they cannot be used in ORDER BY clauses, although the use of conditional logic should resolve the need for order by variables.

• Functions – As a rule of thumb, validate all the data being passed into a query prior to it being used. ColdFusion MX 7 saw the introduction of the isValid() function. This function tests whether a value meets a validation or data type rule and can be used to replace a large number of type-specific functions such as isArray(), isBinary(), isBoolean(), isDate(), isNumeric() and isSimpleValue() etc.
• Stored Procedures – I often favour the use of stored procedures over standard queries. Not only do they add an additional level of performance, they provide an additional level of security; ColdFusion does not do any raw processing of queries in the web code, it simply passes variables down the wire to the database server.

Additional Resources

• Web Security, Privacy and Commerce
• O’Reilly’s The Web Application Hacker’s Handbook
• Adobe’s whitepaper – ColdFusion 8 Developer Security Guidlines (PDF, 281k)
• Adobe’s whitepaper – ColdFusion 7 Developer Security Guidlines (PDF, 217k)
• Adobe DevNet – Learning Stored Procedure Basics in ColdFusion 8
• 0×000000 # The Hacker Webzine’s article on Attacking ColdFusion
• Three part series from Mark Kruger (ColdFusion Muse) – Part 1, Part 2, Part 3
• Brad Wood’s article on CFQueryParam is not just for security.