Acegi Security provides comprehensive security services for J2EE-based enterprise software applications. There is a particular emphasis on supporting projects built using The Spring Framework, which is the leading J2EE solution for enterprise software development. If you're not using Spring for developing enterprise applications, we warmly encourage you to take a closer look at it. Some familiarity with Spring - and in particular dependency injection principles - will help you get up to speed with Acegi Security more easily.

People use Acegi Security for many reasons, but most are drawn to the project after finding the security features of J2EE's Servlet Specification or EJB Specification lack the depth required for typical enterprise application scenarios. Whilst mentioning these standards, it's important to recognise that they are not portable at a WAR or EAR level. Therefore, if you switch server environments, it is typically a lot of work to reconfigure your application's security in the new target environment. Using Acegi Security overcomes these problems, and also brings you dozens of other useful, entirely customisable security features.

As you probably know, security comprises two major operations. The first is known as "authentication", which is the process of establishing a principal is who they claim to be. A "principal" generally means a user, device or some other system which can perform an action in your application. "Authorization" refers to the process of deciding whether a principal is allowed to perform an action in your application. To arrive at the point where an authorization decision is needed, the identity of the principal has already been established by the authentication process. These concepts are common, and not at all specific to Acegi Security.

At an authentication level, Acegi Security supports a wide range of authentication models. Most of these authentication models are either provided by third parties, or are developed by relevant standards bodies such as the Internet Engineering Task Force. In addition, Acegi Security provides its own set of authentication features. Specifically, Acegi Security currently supports authentication with all of these technologies:

Many independent software vendors (ISVs) adopt Acegi Security because of this rich choice of authentication models. Doing so allows them to quickly integrate their solutions with whatever their end clients need, without undertaking a lot of engineering or requiring the client to change their environment. If none of the above authentication mechanisms suit your needs, Acegi Security is an open platform and it is quite simple to write your own authentication mechanism. Many corporate users of Acegi Security need to integrate with "legacy" systems that don't follow any particular security standards, and Acegi Security is happy to "play nicely" with such systems.

Sometimes the mere process of authentication isn't enough. Sometimes you need to also differentiate security based on the way a principal is interacting with your application. For example, you might want to ensure requests only arrive over HTTPS, in order to protect passwords from eavesdropping or end users from man-in-the-middle attacks. Or, you might want to ensure that an actual human being is making the requests and not some robot or other automated process. This is especially helpful to protect password recovery processes from brute force attacks, or simply to make it harder for people to duplicate your application's key content. To help you achieve these goals, Acegi Security fully supports automatic "channel security", together with JCaptcha integration for human user detection.

Irrespective of how authentication was undertaken, Acegi Security provides a deep set of authorization capabilities. There are three main areas of interest in respect of authorization, these being authorizing web requests, authorizing methods can be invoked, and authorizing access to individual domain object instances. To help you understand the differences, consider the authorization capabilities found in the Servlet Specification web pattern security, EJB Container Managed Security and file system security respectively. Acegi Security provides deep capabilities in all of these important areas, which we'll explore later in this reference guide.

Acegi Security began in late 2003, when a question was posed on the Spring Developers' mailing list asking whether there had been any consideration given to a Spring-based security implementation. At the time the Spring community was relatively small (especially by today's size!), and indeed Spring itself had only existed as a SourceForge project from early 2003. The response to the question was that it was a worthwhile area, although a lack of time currently prevented its exploration.

With that in mind, a simple security implementation was built and not released. A few weeks later another member of the Spring community inquired about security, and at the time this code was offered to them. Several other requests followed, and by January 2004 around twenty people were using the code. These pioneering users were joined by others who suggested a SourceForge project was in order, which was duly established in March 2004.

In those early days, the project didn't have any of its own authentication modules. Container Managed Security was relied upon for the authentication process, with Acegi Security instead focusing on authorization. This was suitable at first, but as more and more users requested additional container support, the fundamental limitation of container-specific authentication realm interfaces was experienced. There was also a related issue of adding new JARs to the container's classpath, which was a common source of end user confusion and misconfiguration.

Acegi Security-specific authentication services were subsequently introduced. Around a year later, the Acegi Security became an official Spring Framework subproject. The 1.0.0 final release was published in May 2006 - after more than two and a half years of active use in numerous production software projects and many hundreds of improvements and community contributions.

Today Acegi Security enjoys a strong and active open source community. There are thousands of messages about Acegi Security on the support forums. Fourteen developers work on the code itself, with an active community who also regularly share patches and support their peers.

It is useful to understand how Acegi Security release numbers work, as it will help you identify the effort (or lack thereof) involved in migrating to future releases of the project. Officially, we use the Apache Portable Runtime Project versioning guidelines, which can be viewed at http://apr.apache.org/versioning.html. We quote the introduction contained on that page for your convenience:

“Versions are denoted using a standard triplet of integers: MAJOR.MINOR.PATCH. The basic intent is that MAJOR versions are incompatible, large-scale upgrades of the API. MINOR versions retain source and binary compatibility with older minor versions, and changes in the PATCH level are perfectly compatible, forwards and backwards.”

Acegi Security is written to execute within a standard Java 1.3 Runtime Environment. It also supports Java 5.0, although the Java types which are specific to this release are packaged in a separate package with the suffix "tiger" in their JAR filename. As Acegi Security aims to operate in a self-contained manner, there is no need to place any special configuration files into your Java Runtime Environment. In particular, there is no need to configure a special Java Authentication and Authorization Service (JAAS) policy file or place Acegi Security into common classpath locations.

Similarly, if you are using an EJB Container or Servlet Container there is no need to put any special configuration files anywhere, nor include Acegi Security in a server classloader.

This above design offers maximum deployment time flexibility, as you can simply copy your target artifact (be it a JAR, WAR or EAR) from one system to another and it will immediately work.

Let's explore some of the most important shared components in Acegi Security. Components are considered "shared" if they are central to the framework and the framework cannot operate without them. These Java types represent the building blocks of the remaining system, so it's important to understand that they're there, even if you don't need to directly interact with them.

The most fundamental object is SecurityContextHolder. This is where we store details of the present security context of the application, which includes details of the principal currently using the application. By default the SecurityContextHolder uses a ThreadLocal to store these details, which means that the security context is always available to methods in the same thread of execution, even if the security context is not explicitly passed around as an argument to those methods. Using a ThreadLocal in this way is quite safe if care is taken to clear the thread after the present principal's request is processed. Of course, Acegi Security takes care of this for you automatically so there is no need to worry about it.

Some applications aren't entirely suitable for using a ThreadLocal, because of the specific way they work with threads. For example, a Swing client might want all threads in a Java Virtual Machine to use the same security context. For this situation you would use the SecurityContextHolder.MODE_GLOBAL. Other applications might want to have threads spawned by the secure thread also assume the same security identity. This is achieved by using SecurityContextHolder.MODE_INHERITABLETHREADLOCAL. You can change the mode from the default SecurityContextHolder.MODE_THREADLOCAL in two ways. The first is to set a system property. Alternatively, call a static method on SecurityContextHolder. Most applications won't need to change from the default, but if you do, take a look at the JavaDocs for SecurityContextHolder to learn more.

Inside the SecurityContextHolder we store details of the principal currently interacting with the application. Acegi Security uses an Authentication object to represent this information. Whilst you won't normally need to create an Authentication object yourself, it is fairly common for users to query the Authentication object. You can use the following code block - from anywhere in your application - to do this:

Object obj = SecurityContextHolder.getContext().getAuthentication().getPrincipal();

if (obj instanceof UserDetails) {
  String username = ((UserDetails)obj).getUsername();
} else {
  String username = obj.toString();
}

The above code introduces a number of interesting relationships and key objects. First, you will notice that there is an intermediate object between SecurityContextHolder and Authentication. The SecurityContextHolder.getContext() method is actually returning a SecurityContext. Acegi Security uses a few different SecurityContext implementations, such as if we need to store special information related to a request that is not principal-specific. A good example of this is our JCaptcha integration, which needs to know whether the current request came from a human user or not. Because such a decision has nothing at all to do with the principal the request may or may not be authenticated as, we store it in the SecurityContext.

Another item to note from the above code fragment is that you can obtain a principal from the Authentication object. The principal is just an Object. Most of the time this can be cast into a UserDetails object. UserDetails is a central interface in Acegi Security. It represents a principal, but in an extensible and application-specific way. Think of UserDetails as the adapter between your own user database and what Acegi Security needs inside the SecurityContextHolder. Being a representation of something from your own user database, quite often you will cast the UserDetails to the original object that your application provided, so you can call business-specific methods (like getEmail(), getEmployeeNumber() and so on).

By now you're probably wondering, so when do I provide a UserDetails object? How do I do that? I thought you said this thing was declarative and I didn't need to write any Java code - what gives? The short answer is that there is a special interface called UserDetailsService. The only method on this interface accepts a String-based username argument and returns a UserDetails. Most authentication providers that ship with Acegi Security delegate to a UserDetailsService as part of the authentication process. The UserDetailsService is used to build the Authentication object that is stored in the SecurityContextHolder. The good news is that we provide a number of UserDetailsService implementations, including one that uses an in-memory map and another that uses JDBC. Most users tend to write their own, though, with such implementations often simply sitting on top of an existing Data Access Object (DAO) that represents their employees, customers, or other users of the enterprise application. Remember the advantage that whatever your UserDetailsService returns can always be obtained from the SecurityContextHolder, as per the above code fragment.

Besides the principal, another important method provided by Authentication is getAuthorities(). This method provides an array of GrantedAuthority objects. A GrantedAuthority is, not surprisingly, an authority that is granted to the principal. Such authorities are usually "roles", such as ROLE_ADMINISTRATOR or ROLE_HR_SUPERVISOR. These roles are later on configured for web authorization, method authorization and domain object authorization. Other parts of Acegi Security are capable of interpreting these authorities, and expect them to be present. GrantedAuthority objects are usually loaded by the UserDetailsService.

Usually the GrantedAuthority objects are application-wide permissions. They are not specific to a given domain object. Thus, you wouldn't likely have a GrantedAuthority to represent a permission to Employee object number 54, because if there are thousands of such authorities you would quickly run out of memory (or, at the very least, cause the application to take a long time to authenticate a user). Of course, Acegi Security is expressly designed to handle this common requirement, but you'd instead use the project's domain object security capabilities for this purpose.

Last but not least, sometimes you will need to store the SecurityContext between HTTP requests. Other times the principal will re-authenticate on every request, although most of the time it will be stored. The HttpSessionContextIntegrationFilter is responsible for storing a SecurityContext between HTTP requests. As suggested by the name of the class, the HttpSession is used to store this information. You should never interact directly with the HttpSession for security purposes. There is simply no justification for doing so - always use the SecurityContextHolder instead.

Just to recap, the major building blocks of Acegi Security are:

Now that you've gained an understanding of these repeatedly-used components, let's take a closer look at the process of authentication.

As mentioned in the beginning of this reference guide, Acegi Security can participate in many different authentication environments. Whilst we recommend people use Acegi Security for authentication and not integrate with existing Container Managed Authentication, it is nevertheless supported - as is integrating with your own proprietary authentication system. Let's first explore authentication from the perspective of Acegi Security managing web security entirely on its own, which is illustrative of the most complex and most common situation.

Consider a typical web application's authentication process:

Acegi Security has distinct classes responsible for most of the steps described above. The main participants (in the order that they are used) are the ExceptionTranslationFilter, an AuthenticationEntryPoint, an authentication mechanism, and an AuthenticationProvider.

ExceptionTranslationFilter is an Acegi Security filter that has responsibility for detecting any Acegi Security exceptions that are thrown. Such exceptions will generally be thrown by an AbstractSecurityInterceptor, which is the main provider of authorization services. We will discuss AbstractSecurityInterceptor in the next section, but for now we just need to know that it produces Java exceptions and knows nothing about HTTP or how to go about authenticating a principal. Instead the ExceptionTranslationFilter offers this service, with specific responsibility for either returning error code 403 (if the principal has been authenticated and therefore simply lacks sufficient access - as per step seven above), or launching an AuthenticationEntryPoint (if the principal has not been authenticated and therefore we need to go commence step three).

The AuthenticationEntryPoint is responsible for step three in the above list. As you can imagine, each web application will have a default authentication strategy (well, this can be configured like nearly everything else in Acegi Security, but let's keep it simple for now). Each major authentication system will have its own AuthenticationEntryPoint implementation, which takes actions such as described in step three.

After your browser decides to submit your authentication credentials (either as an HTTP form post or HTTP header) there needs to be something on the server that "collects" these authentication details. By now we're at step six in the above list. In Acegi Security we have a special name for the function of collecting authentication details from a user agent (usually a web browser), and that name is "authentication mechanism". After the authentication details are collected from the user agent, an "Authentication request" object is built and then presented to an AuthenticationProvider.

The last played in the Acegi Security authentication process is an AuthenticationProvider. Quite simply, it is responsible for taking an Authentication request object and deciding whether or not it is valid. The provider will either throw an exception or return a fully populated Authentication object. Remember our good friends, UserDetails and UserDetailsService? If not, head back to the previous section and refresh your memory. Most AuthenticationProviders will ask a UserDetailsService to provide a UserDetails object. As mentioned earlier, most application will provide their own UserDetailsService, although some will be able to use the JDBC or in-memory implementation that ships with Acegi Security. The resultant UserDetails object - and particularly the GrantedAuthority[]s contained within the UserDetails object - will be used when building the fully populated Authentication object.

After the authentication mechanism receives back the fully-populated Authentication object, it will deem the request valid, put the Authentication into the SecurityContextHolder, and cause the original request to be retried (step seven above). If, on the other hand, the AuthenticationProvider rejected the request, the authentication mechanism will ask the user agent to retry (step two above).

Whilst this describes the typical authentication workflow, the good news is that Acegi Security doesn't mind how you put an Authentication inside the SecurityContextHolder. The only critical requirement is that the SecurityContextHolder contains an Authentication that represents a principal before the AbstractSecurityInterceptor needs to authorize a request.

You can (and many users do) write their own filters or MVC controllers to provide interoperability with authentication systems that are not based on Acegi Security. For example, you might be using Container Managed Authentication which makes the current user available from a ThreadLocal or JNDI location. Or you might work for a company that has a legacy proprietary authentication system, which is a corporate "standard" over which you have little control. In such situations it's quite easy to get Acegi Security to work, and still provide authorization capabilities. All you need to do is write a filter (or equivalent) that reads the third-party user information from a location, build an Acegi Security-specific Authentication object, and put it onto the SecurityContextHolder. It's quite easy to do this, and it is a fully-supported integration approach.

If you're familiar with AOP, you'd be aware there are different types of advice available: before, after, throws and around. An around advice is very useful, because an advisor can elect whether or not to proceed with a method invocation, whether or not to modify the response, and whether or not to throw an exception. Acegi Security provides an around advice for method invocations as well as web requests. We achieve an around advice for method invocations using AOP Alliance, and we achieve an around advice for web requests using a standard Filter.

For those not familiar with AOP, the key point to understand is that Acegi Security can help you protect method invocations as well as web requests. Most people are interested in securing method invocations on their services layer. This is because the services layer is where most business logic resides in current-generation J2EE applications (for clarification, the author disapproves of this design and instead advocates properly encapsulated domain objects together with the DTO, assembly, facade and transparent persistence patterns, but as anemic domain objects is the present mainstream approach, we'll talk about it here). If you just need to secure method invocations to the services layer, using the Spring's standard AOP platform (otherwise known as AOP Alliance) will be adequate. If you need to secure domain objects directly, you will likely find that AspectJ is worth considering.

You can elect to perform method authorization using AspectJ or AOP Alliance, or you can elect to perform web request authorization using filters. You can use zero, one, two or three of these approaches together. The mainstream usage is to perform some web request authorization, coupled with some AOP Alliance method invocation authorization on the services layer.

Acegi Security uses the term "secure object" to refer to any object that can have security applied to it. Each secure object supported by Acegi Security has its own class, which is a subclass of AbstractSecurityInterceptor. Importantly, by the time the AbstractSecurityInterceptor is run, the SecurityContextHolder will contain a valid Authentication if the principal has been authenticated.

The AbstractSecurityInterceptor provides a consistent workflow for handling secure object requests. This workflow includes looking up the "configuration attributes" associated with the present request. A "configuration attribute" can be thought of as a String that has special meaning to the classes used by AbstractSecurityInterceptor. They're normally configured against your AbstractSecurityInterceptor using XML. Anyway, the AbstractSecurityInterceptor will ask an AccessDecisionManager "here's the configuration attributes, here's the current Authentication object, and here's details of the current request - is this particular principal allowed to perform this particular operation?".

Assuming AccessDecisionManager decides to allow the request, the AbstractSecurityInterceptor will normally just proceed with the request. Having said that, on rare occasions users may want to replace the Authentication inside the SecurityContext with a different Authentication, which is handled by the AccessDecisionManager calling a RunAsManager. This might be useful in reasonably unusual situations, such as if a services layer method needs to call a remote system and present a different identity. Because Acegi Security automatically propagates security identity from one server to another (assuming you're using a properly-configured RMI or HttpInvoker remoting protocol client), this may be useful.

Following the secure object proceeding and then returning - which may mean a method invocation completing or a filter chain proceeding - the AbstractSecurityInterceptor gets one final chance to handle the invocation. At this stage the AbstractSecurityInterceptor is interested in possibly modifying the return object. We might want this to happen because an authorization decision couldn't be made "on the way in" to a secure object invocation. Being highly pluggable, AbstractSecurityInterceptor will pass control to an AfterInvocationManager to actually modify the object if needed. This class even can entirely replace the object, or throw an exception, or not change it in any way.

Because AbstractSecurityInterceptor is the central template class, it seems fitting that the first figure should be devoted to it.

Figure 1: The key "secure object" model

Only developers contemplating an entirely new way of intercepting and authorizing requests would need to use secure objects directly. For example, it would be possible to build a new secure object to secure calls to a messaging system. Anything that requires security and also provides a way of intercepting a call (like the AOP around advice semantics) is capable of being made into a secure object. Having said that, most Spring applications will simply use the three currently supported secure object types (AOP Alliance MethodInvocation, AspectJ JoinPoint and web request FilterInterceptor) with complete transparency.

Congratulations! You have enough of a high-level picture of Acegi Security to embark on your project. We've explored the shared components, how authentication works, and reviewed the common authorization concept of a "secure object". Everything that follows in this reference guide may or may not apply to your particular needs, and can be read in any order.

Acegi Security supports localization of exception messages that end users are likely to see. If your application is designed for English users, you don't need to do anything as by default all Acegi Security messages are in English. If you need to support other locales, everything you need to know is contained in this section.

All exception messages can be localized, including messages related to authentication failures and access being denied (authorization failures). Exceptions and logging that is focused on developers or system deployers (including incorrect attributes, interface contract violations, using incorrect constructors, startup time validation, debug-level logging) etc are not localized and instead are hard-coded in English within Acegi Security's code.

Shipping in the acegi-security-xx.jar you will find an org.acegisecurity package that in turn contains a messages.properties file. This should be referred to by your ApplicationContext, as Acegi Security classes implement Spring's MessageSourceAware interface and expect the message resolver to be dependency injected at application context startup time. Usually all you need to do is register a bean inside your application context to refer to the messages. An example is shown below:

<bean id="messageSource" class="org.springframework.context.support.ReloadableResourceBundleMessageSource">
  <property name="basename"><value>org/acegisecurity/messages</value></property>
</bean>        

The messages.properties is named in accordance with standard resource bundles and represents the default language supported by Acegi Securtiy messages. This default file is in English. If you do not register a message source, Acegi Security will still work correctly and fallback to hard-coded English versions of the messages.

If you wish to customize the messages.properties file, or support other languages, you should copy the file, rename it accordingly, and register it inside the above bean definition. There are not a large number of message keys inside this file, so localization should not be considered a major initiative. If you do perform localization of this file, please consider sharing your work with the community by logging a JIRA task and attaching your appropriately-named localized version of messages.properties.

Rounding out the discussion on localization is the Spring ThreadLocal known as org.springframework.context.i18n.LocaleContextHolder. You should set the LocaleContextHolder to represent the preferred Locale of each user. Acegi Security will attempt to locate a message from the message source using the Locale obtained from this ThreadLocal. Please refer to Spring documentation for further details on using LocaleContextHolder and the helper classes that can automatically set it for you (eg AcceptHeaderLocaleResolver, CookieLocaleResolver, FixedLocaleResolver, SessionLocaleResolver etc)

Acegi Security uses many filters, as referred to throughout the remainder of this reference guide. You have a choice in how these filters are added to your web application, in that you can use either FilterToBeanProxy or FilterChainProxy. We'll look at both below.

Most filters are configured using the FilterToBeanProxy. An example configuration from web.xml follows:

<filter>
  <filter-name>Acegi HTTP Request Security Filter</filter-name>
  <filter-class>org.acegisecurity.util.FilterToBeanProxy</filter-class>
  <init-param>
    <param-name>targetClass</param-name>
    <param-value>org.acegisecurity.ClassThatImplementsFilter</param-value>
  </init-param>
</filter>

Notice that the filter in web.xml is actually a FilterToBeanProxy, and not the filter that will actually implement the logic of the filter. What FilterToBeanProxy does is delegate the Filter's methods through to a bean which is obtained from the Spring application context. This enables the bean to benefit from the Spring application context lifecycle support and configuration flexibility. The bean must implement javax.servlet.Filter.

The FilterToBeanProxy only requires a single initialization parameter, targetClass or targetBean. The targetClass parameter locates the first object in the application context of the specified class, whilst targetBean locates the object by bean name. Like standard Spring web applications, the FilterToBeanProxy accesses the application context via WebApplicationContextUtils.getWebApplicationContext(ServletContext), so you should configure a ContextLoaderListener in web.xml.

There is a lifecycle issue to consider when hosting Filters in an IoC container instead of a servlet container. Specifically, which container should be responsible for calling the Filter's "startup" and "shutdown" methods? It is noted that the order of initialization and destruction of a Filter can vary by servlet container, and this can cause problems if one Filter depends on configuration settings established by an earlier initialized Filter. The Spring IoC container on the other hand has more comprehensive lifecycle/IoC interfaces (such as InitializingBean, DisposableBean, BeanNameAware, ApplicationContextAware and many others) as well as a well-understood interface contract, predictable method invocation ordering, autowiring support, and even options to avoid implementing Spring interfaces (eg the destroy-method attribute in Spring XML). For this reason we recommend the use of Spring lifecycle services instead of servlet container lifecycle services wherever possible. By default FilterToBeanProxy will not delegate init(FilterConfig) and destroy() methods through to the proxied bean. If you do require such invocations to be delegated, set the lifecycle initialization parameter to servlet-container-managed.

Rather than using FilterToBeanProxy, we strongly recommend to use FilterChainProxy instead. Whilst FilterToBeanProxy is a very useful class, the problem is that the lines of code required for <filter> and <filter-mapping> entries in web.xml explodes when using more than a few filters. To overcome this issue, Acegi Security provides a FilterChainProxy class. It is wired using a FilterToBeanProxy (just like in the example above), but the target class is org.acegisecurity.util.FilterChainProxy. The filter chain is then declared in the application context, using code such as this:

<bean id="filterChainProxy" class="org.acegisecurity.util.FilterChainProxy">
  <property name="filterInvocationDefinitionSource">
    <value>
      CONVERT_URL_TO_LOWERCASE_BEFORE_COMPARISON
      PATTERN_TYPE_APACHE_ANT
      /webServices/**=httpSessionContextIntegrationFilterWithASCFalse,basicProcessingFilter,exceptionTranslationFilter,filterSecurityInterceptor
      /**=httpSessionContextIntegrationFilterWithASCTrue,authenticationProcessingFilter,exceptionTranslationFilter,filterSecurityInterceptor
    </value>
  </property>
</bean>        

You may notice similarities with the way FilterSecurityInterceptor is declared. Both regular expressions and Ant Paths are supported, and the most specific URIs appear first. At runtime the FilterChainProxy will locate the first URI pattern that matches the current web request. Each of the corresponding configuration attributes represent the name of a bean defined in the application context. The filters will then be invoked in the order they are specified, with standard FilterChain behaviour being respected (a Filter can elect not to proceed with the chain if it wishes to end processing).

As you can see, FilterChainProxy requires the duplication of filter names for different request patterns (in the above example, exceptionTranslationFilter and filterSecurityInterceptor are duplicated). This design decision was made to enable FilterChainProxy to specify different Filter invocation orders for different URI patterns, and also to improve both the expressiveness (in terms of regular expressions, Ant Paths, and any custom FilterInvocationDefinitionSource implementations) and clarity of which Filters should be invoked.

You may have noticed we have declared two HttpSessionContextIntegrationFilters in the filter chain (ASC is short for allowSessionCreation, a property of HttpSessionContextIntegrationFilter). As web services will never present a jsessionid on future requests, creating HttpSessions for such user agents would be wasteful. If you had a high-volume application which required maximum scalability, we recommend you use the approach shown above. For smaller applications, using a single HttpSessionContextIntegrationFilter (with its default allowSessionCreation as true) would likely be sufficient.

In relation to lifecycle issues, the FilterChainProxy will always delegate init(FilterConfig) and destroy() methods through to the underlaying Filters if such methods are called against FilterChainProxy itself. In this case, FilterChainProxy guarantees to only initialize and destroy each Filter once, irrespective of how many times it is declared by the FilterInvocationDefinitionSource. You control the overall choice as to whether these methods are called or not via the lifecycle initialization parameter of the FilterToBeanProxy that proxies FilterChainProxy. As discussed above, by default any servlet container lifecycle invocations are not delegated through to FilterChainProxy.

You can also omit a URI pattern from the filter chain by using the token #NONE# on the right-hand side of the <URI Pattern> = <Filter Chain> expression. For example, using the example above, if you wanted to exclude the /webservices location completely, you would modify the corresponding line in the bean declaration to be

/webServices/**=#NONE#
        

Note that anything matching this path will then have no authentication or authorization services applied and will be freely accessible.

The order that filters are defined in web.xml is very important. Irrespective of which filters you are actually using, the order of the <filter-mapping>s should be as follows:

All of the above filters use FilterToBeanProxy or FilterChainProxy. It is recommended that a single FilterToBeanProxy proxy through to a single FilterChainProxy for each application, with that FilterChainProxy defining all of Acegi Security Filters.

If you're using SiteMesh, ensure Acegi Security filters execute before the SiteMesh filters are called. This enables the SecurityContextHolder to be populated in time for use by SiteMesh decorators

In addition to coordinating the authentication and authorization requirements of your application, Acegi Security is also able to ensure unauthenticated web requests have certain properties. These properties may include being of a particular transport type, having a particular HttpSession attribute set and so on. The most common requirement is for your web requests to be received using a particular transport protocol, such as HTTPS.

An important issue in considering transport security is that of session hijacking. Your web container manages a HttpSession by reference to a jsessionid that is sent to user agents either via a cookie or URL rewriting. If the jsessionid is ever sent over HTTP, there is a possibility that session identifier can be intercepted and used to impersonate the user after they complete the authentication process. This is because most web containers maintain the same session identifier for a given user, even after they switch from HTTP to HTTPS pages.

If session hijacking is considered too significant a risk for your particular application, the only option is to use HTTPS for every request. This means the jsessionid is never sent across an insecure channel. You will need to ensure your web.xml-defined <welcome-file> points to an HTTPS location, and the application never directs the user to an HTTP location. Acegi Security provides a solution to assist with the latter.

To utilise Acegi Security's channel security services, add the following lines to web.xml:

<filter>
  <filter-name>Acegi Channel Processing Filter</filter-name>
  <filter-class>org.acegisecurity.util.FilterToBeanProxy</filter-class>
  <init-param>
    <param-name>targetClass</param-name>
    <param-value>org.acegisecurity.securechannel.ChannelProcessingFilter</param-value>
  </init-param>
</filter>

<filter-mapping>
  <filter-name>Acegi Channel Processing Filter</filter-name>
  <url-pattern>/*</url-pattern>
</filter-mapping>

        

As usual when running FilterToBeanProxy, you will also need to configure the filter in your application context:

<bean id="channelProcessingFilter" class="org.acegisecurity.securechannel.ChannelProcessingFilter">
  <property name="channelDecisionManager"><ref bean="channelDecisionManager"/></property>
  <property name="filterInvocationDefinitionSource">
    <value>
      CONVERT_URL_TO_LOWERCASE_BEFORE_COMPARISON
      \A/secure/.*\Z=REQUIRES_SECURE_CHANNEL
      \A/acegilogin.jsp.*\Z=REQUIRES_SECURE_CHANNEL
      \A/j_acegi_security_check.*\Z=REQUIRES_SECURE_CHANNEL
      \A.*\Z=REQUIRES_INSECURE_CHANNEL
    </value>
  </property>
</bean>

<bean id="channelDecisionManager" class="org.acegisecurity.securechannel.ChannelDecisionManagerImpl">
  <property name="channelProcessors">
    <list>
      <ref bean="secureChannelProcessor"/>
      <ref bean="insecureChannelProcessor"/>
    </list>
  </property>
</bean>

<bean id="secureChannelProcessor" class="org.acegisecurity.securechannel.SecureChannelProcessor"/>
<bean id="insecureChannelProcessor" class="org.acegisecurity.securechannel.InsecureChannelProcessor"/>

        

Like FilterSecurityInterceptor, Apache Ant style paths are also supported by the ChannelProcessingFilter.

The ChannelProcessingFilter operates by filtering all web requests and determining the configuration attributes that apply. It then delegates to the ChannelDecisionManager. The default implementation, ChannelDecisionManagerImpl, should suffice in most cases. It simply delegates through the list of configured ChannelProcessor instances. A ChannelProcessor will review the request, and if it is unhappy with the request (eg it was received across the incorrect transport protocol), it will perform a redirect, throw an exception or take whatever other action is appropriate.

Included with Acegi Security are two concrete ChannelProcessor implementations: SecureChannelProcessor ensures requests with a configuration attribute of REQUIRES_SECURE_CHANNEL are received over HTTPS, whilst InsecureChannelProcessor ensures requests with a configuration attribute of REQUIRES_INSECURE_CHANNEL are received over HTTP. Both implementations delegate to a ChannelEntryPoint if the required transport protocol is not used. The two ChannelEntryPoint implementations included with Acegi Security simply redirect the request to HTTP and HTTPS as appropriate. Appropriate defaults are assigned to the ChannelProcessor implementations for the configuration attribute keywords they respond to and the ChannelEntryPoint they delegate to, although you have the ability to override these using the application context.

Note that the redirections are absolute (eg http://www.company.com:8080/app/page), not relative (eg /app/page). During testing it was discovered that Internet Explorer 6 Service Pack 1 has a bug whereby it does not respond correctly to a redirection instruction which also changes the port to use. Accordingly, absolute URLs are used in conjunction with bug detection logic in the PortResolverImpl that is wired up by default to many Acegi Security beans. Please refer to the JavaDocs for PortResolverImpl for further details.

You should note that using a secure channel is recommended if usernames and passwords are to be kept secure during the login process. If you do decide to use ChannelProcessingFilter with form-based login, please ensure that your login page is set to REQUIRES_SECURE_CHANNEL, and that the AuthenticationProcessingFilterEntryPoint.forceHttps property is true.

Once configured, using the channel security filter is very easy. Simply request pages without regard to the protocol (ie HTTP or HTTPS) or port (eg 80, 8080, 443, 8443 etc). Obviously you'll still need a way of making the initial request (probably via the web.xml <welcome-file> or a well-known home page URL), but once this is done the filter will perform redirects as defined by your application context.

You can also add your own ChannelProcessor implementations to the ChannelDecisionManagerImpl. For example, you might set a HttpSession attribute when a human user is detected via a "enter the contents of this graphic" procedure. Your ChannelProcessor would respond to say REQUIRES_HUMAN_USER configuration attributes and redirect to an appropriate entry point to start the human user validation process if the HttpSession attribute is not currently set.

To decide whether a security check belongs in a ChannelProcessor or an AccessDecisionVoter, remember that the former is designed to handle unauthenticated requests, whilst the latter is designed to handle authenticated requests. The latter therefore has access to the granted authorities of the authenticated principal. In addition, problems detected by a ChannelProcessor will generally cause an HTTP/HTTPS redirection so its requirements can be met, whilst problems detected by an AccessDecisionVoter will ultimately result in an AccessDeniedException (depending on the governing AccessDecisionManager).

Acegi Security comes bundled with several JSP tag libraries that eases JSP writing. The tag libraries are known as authz and provide a range of different services.

All taglib classes are included in the core acegi-security-xx.jar file, with the authz.tld located in the JAR's META-INF directory. This means for JSP 1.2+ web containers you can simply include the JAR in the WAR's WEB-INF/lib directory and it will be available. If you're using a JSP 1.1 container, you'll need to declare the JSP taglib in your web.xml file, and include authz.tld in the WEB-INF/lib directory. The following fragment is added to web.xml:

<taglib>
  <taglib-uri>http://acegisecurity.org/authz</taglib-uri>
  <taglib-location>/WEB-INF/authz.tld</taglib-location>
</taglib>       

Now that you've configured the tag libraries, refer to the individual reference guide sections for details on how to use them.

If you're using Acegi Security-provided authentication approaches, you'll usually need to configure a web filter, together with an AuthenticationProvider and AuthenticationEntryPoint. In this section we are going to explore an example application that needs to support both form-based authentication (ie so a nice HTML page is presented to a user for them to login) plus BASIC authentication (ie so a web service or similar can access protected resources).

In the web.xml, this application will need a single Acegi Security filter in order to use the FilterChainProxy. Nearly every Acegi Security application will have such an entry, and it looks like this:

<filter>
  <filter-name>Acegi Filter Chain Proxy</filter-name>
  <filter-class>org.acegisecurity.util.FilterToBeanProxy</filter-class>
  <init-param>
    <param-name>targetClass</param-name>
    <param-value>org.acegisecurity.util.FilterChainProxy</param-value>
  </init-param>
</filter>

<filter-mapping>
  <filter-name>Acegi Filter Chain Proxy</filter-name>
  <url-pattern>/*</url-pattern>
</filter-mapping>

The above declarations will cause every web request to be passed through to Acegi Security's FilterChainProxy. As explained in the filters section of this reference guide, the FilterChainProxy is a generally-useful class that enables web requests to be passed to different filters based on the URL patterns. Those delegated filters are managed inside the application context, so they can benefit from dependency injection. Let's have a look at what the FilterChainProxy bean definition would look like inside your application context:

<bean id="filterChainProxy" class="org.acegisecurity.util.FilterChainProxy">
  <property name="filterInvocationDefinitionSource">
    <value>
      CONVERT_URL_TO_LOWERCASE_BEFORE_COMPARISON
      PATTERN_TYPE_APACHE_ANT
      /**=httpSessionContextIntegrationFilter,logoutFilter,authenticationProcessingFilter,basicProcessingFilter,securityContextHolderAwareRequestFilter,rememberMeProcessingFilter,anonymousProcessingFilter,exceptionTranslationFilter,filterInvocationInterceptor,switchUserProcessingFilter
    </value>
  </property>
</bean>

Internally Acegi Security will use a PropertyEditor to convert the string presented in the above XML fragment into a FilterInvocationDefinitionSource object. What's important to note at this stage is that a series of filters will be run - in the order specified by the declaration - and each of those filters are actually the <bean id> of another bean inside the application context. So, in our case some extra beans will also appear in the application context, and they'll be named httpSessionContextIntegrationFilter, logoutFilter and so on. The order that the filters should appear is discussed in the filters section of the reference guide - although they are correct in the above example.

In our example we have the AuthenticationProcessingFilter and BasicProcessingFilter being used. These are the "authentication mechanisms" that respond to form-based authentication and BASIC HTTP header-based authentication respectively (we discussed the role of authentication mechanisms earlier in this reference guide). If you weren't using form or BASIC authentication, neither of these beans would be defined. You'd instead define filters applicable to your desired authentication environment, such as DigestProcessingFilter or CasProcessingFilter. Refer to the individual chapters of this part of the reference guide to learn how to configure each of these authentication mechanisms.

Recall that HttpSessionContextIntegrationFilter keeps the contents of the SecurityContext between invocations inside an HTTP session. This means the authentication mechanisms are only used once, being when the principal initially tries to authenticate. The rest of the time the authentication mechanisms sit there and silently pass the request through to the next filter in the chain. That is a practical requirement due to the fact that few authentication approaches present credentials on each and every call (BASIC authentication being a notable exception), but what happens if a principal's account gets cancelled or disabled or otherwise changed (eg an increase or decrease in GrantedAuthority[]s) after the initial authentication step? Let's look at how that is handled now.

The major authorization provider for secure objects has previously been introduced as AbstractSecurityInterceptor. This class needs to have access to an AuthenticationManager. It also has configurable settings to indicate whether an Authentication object should be re-authenticated on each secure object invocation. By default it just accepts any Authentication inside the SecurityContextHolder is authenticated if Authentication.isAuthenticated() returns true. This is great for performance, but not ideal if you want to ensure up-to-the-moment authentication validity. For such cases you'll probably want to set the AbstractSecurityInterceptor.alwaysReauthenticate property to true.

You might be asking yourself, "what's this AuthenticationManager?". We haven't explored it before, but we have discussed the concept of an AuthenticationProvider. Quite simply, an AuthenticationManager is responsible for passing requests through a chain of AuthenticationProviders. It's a little like the filter chain we discussed earlier, although there are some differences. There is only one AuthenticationManager implementation shipped with Acegi Security, so let's look at how it's configured for the example we're using in this chapter:

<bean id="authenticationManager" class="org.acegisecurity.providers.ProviderManager">
  <property name="providers">
    <list>
      <ref local="daoAuthenticationProvider"/>
      <ref local="anonymousAuthenticationProvider"/>
      <ref local="rememberMeAuthenticationProvider"/>
    </list>
  </property>
</bean>

It's probably worth mentioning at this point that your authentication mechanisms (which are usually filters) are also injected with a reference to the AuthenticationManager. So both AbstractSecurityInterceptor as well as the authentication mechanisms will use the above ProviderManager to poll a list of AuthenticationProviders.

In our example we have three providers. They are tried in the order shown (which is implied by the use of a List instead of a Set), with each provider able to attempt authentication, or skip authentication by simply returning null. If all implementations return null, the ProviderManager will throw a suitable exception. If you're interested in learning more about chaining providers, please refer to the ProviderManager JavaDocs.

The providers to use will sometimes be interchangeable with the authentication mechanisms, whilst at other times they will depend on a specific authentication mechanism. For example, the DaoAuthenticationProvider just needs a string-based username and password. Various authentication mechanisms result in the collection of a string-based username and password, including (but not limited to) BASIC and form authentication. Equally, some authentication mechanisms create an authentication request object which can only be interpreted by a single type of AuthenticationProvider. An example of this one-to-one mapping would be JA-SIG CAS, which uses the notion of a service ticket which can therefore only be authenticated by CasAuthenticationProvider. A further example of a one-to-one mapping would be the LDAP authentication mechanism, which can only be processed an the LdapAuthenticationProvider. The specifics of such relationships are detailed in the JavaDocs for each class, plus the authentication approach-specific chapters of this reference guide. You need not be terribly concerned about this implementation detail, because if you forget to register a suitable provider, you'll simply receive a ProviderNotFoundException when an attempt to authenticate is made.

After configuring the correct authentication mechanisms in the FilterChainProxy, and ensuring that a corresponding AuthenticationProvider is registered in the ProviderManager, your last step is to configure an AuthenticationEntryPoint. Recall that earlier we discussed the role of ExceptionTranslationFilter, which is used when HTTP-based requests should receive back an HTTP header or HTTP redirect in order to start authentication. Continuing on with our earlier example:

<bean id="exceptionTranslationFilter" class="org.acegisecurity.ui.ExceptionTranslationFilter">
  <property name="authenticationEntryPoint"><ref local="authenticationProcessingFilterEntryPoint"/></property>
  <property name="accessDeniedHandler">
    <bean class="org.acegisecurity.ui.AccessDeniedHandlerImpl">
      <property name="errorPage" value="/accessDenied.jsp"/>
    </bean>
  </property>
</bean>

<bean id="authenticationProcessingFilterEntryPoint" class="org.acegisecurity.ui.webapp.AuthenticationProcessingFilterEntryPoint">
  <property name="loginFormUrl"><value>/acegilogin.jsp</value></property>
  <property name="forceHttps"><value>false</value></property>
</bean>

Notice that the ExceptionTranslationFilter requires two collaborators. The first, AccessDeniedHandlerImpl, uses a RequestDispatcher forward to display the specified access denied error page. We use a forward so that the SecurityContextHolder still contains details of the principal, which may be useful for display to the user (in old releases of Acegi Security we relied upon the servlet container to handle a 403 error message, which lacked this useful contextual information). AccessDeniedHandlerImpl will also set the HTTP header to 403, which is the official error code to indicate access denied. In the case of the AuthentionEntryPoint, here we're setting what action we would like taken when an unauthenticated principal attempts to perform a protected operation. Because in our example we're going to be using form-based authentication, we specify AuthenticationProcessinFilterEntryPoint and the URL of the login page. Your application will usually only have one entry point, and most authentication approaches define their own specific AuthenticationEntryPoint. Details of which entry point to use for each authentication approach is discussed in the authentication approach-specific chapters of this reference guide.

As mentioned in the first part of the reference guide, most authentication providers take advantage of the UserDetails and UserDetailsService interfaces. The contract for this latter interface consists of a single method:

public UserDetails loadUserByUsername(String username) throws UsernameNotFoundException, DataAccessException;

The returned UserDetails is an interface that provides getters that guarantee non-null provision of basic authentication information such as the username, password, granted authorities and whether the user is enabled or disabled. Most authentication providers will use a UserDetailsService, even if the username and password are not actually used as part of the authentication decision. Generally such provider will be using the returned UserDetails object just for its GrantedAuthority[] information, because some other system (like LDAP or X509 or CAS etc) has undertaken the responsibility of actually validating the credentials.

A single concrete implementation of UserDetails is provided with Acegi Security, being the User class. Acegi Security users will need to decide when writing their UserDetailsService what concrete UserDetails class to return. In most cases User will be used directly or subclassed, although special circumstances (such as object relational mappers) may require users to write their own UserDetails implementation from scratch. This is not such an unusual situation, and users should not hesitate to simply return their normal domain object that represents a user of the system. This is especially common given that UserDetails is often used to store additional principal-related properties (such as their telephone number and email address), so that they can be easily used by web views.

Given UserDetailsService is so simple to implement, it should be easy for users to retrieve authentication information using a persistence strategy of their choice. Having said that, Acegi Security does include a couple of useful base implementations, which we'll look at below.

<bean id="inMemoryDaoImpl" class="org.acegisecurity.userdetails.memory.InMemoryDaoImpl">
  <property name="userMap">
    <value>
      marissa=koala,ROLE_TELLER,ROLE_SUPERVISOR
      dianne=emu,ROLE_TELLER
      scott=wombat,ROLE_TELLER
      peter=opal,disabled,ROLE_TELLER
    </value>
  </property>
</bean>        

<bean id="dataSource" class="org.springframework.jdbc.datasource.DriverManagerDataSource">
  <property name="driverClassName"><value>org.hsqldb.jdbcDriver</value></property>
  <property name="url"><value>jdbc:hsqldb:hsql://localhost:9001</value></property>
  <property name="username"><value>sa</value></property>
  <property name="password"><value></value></property>
</bean>

<bean id="jdbcDaoImpl" class="org.acegisecurity.userdetails.jdbc.JdbcDaoImpl">
  <property name="dataSource"><ref bean="dataSource"/></property>
</bean>        

Acegi Security is able to prevent a principal from concurrently authenticating to the same application more than a specified number of times. Many ISVs take advantage of this to enforce licensing, whilst network administrators like this feature because it helps prevent people from sharing login names. You can, for example, stop user "Batman" from logging onto the web application from two different sessions.

To use concurrent session support, you'll need to add the following to web.xml:

<listener>
  <listener-class>org.acegisecurity.ui.session.HttpSessionEventPublisher</listener-class>
</listener>        

In addition, you will need to add the org.acegisecurity.concurrent.ConcurrentSessionFilter to your FilterChainProxy. The ConcurrentSessionFilter requires two properties, sessionRegistry, which generally points to an instance of SessionRegistryImpl, and expiredUrl, which points to the page to display when a session has expired.

The web.xml HttpSessionEventPublisher causes an ApplicationEvent to be published to the Spring ApplicationContext every time a HttpSession commences or terminates. This is critical, as it allows the SessionRegistryImpl to be notified when a session ends.

You will also need to wire up the ConcurrentSessionControllerImpl and refer to it from your ProviderManager bean:

<bean id="authenticationManager" class="org.acegisecurity.providers.ProviderManager">
  <property name="providers">
    <!-- your providers go here -->
  </property>
  <property name="sessionController"><ref bean="concurrentSessionController"/></property>
</bean>

<bean id="concurrentSessionController" class="org.acegisecurity.concurrent.ConcurrentSessionControllerImpl">
  <property name="maximumSessions"><value>1</value></property>
  <property name="sessionRegistry"><ref local="sessionRegistry"/></property>
</bean>

<bean id="sessionRegistry" class="org.acegisecurity.concurrent.SessionRegistryImpl"/>

AuthenticationTag is used to simply output a property of the current principal's Authentication.getPrincipal() object to the web page.

The following JSP fragment illustrates how to use the AuthenticationTag:

<authz:authentication operation="username"/>

This tag would cause the principal's name to be output. Here we are assuming the Authentication.getPrincipal() is a UserDetails object, which is generally the case when using the typical DaoAuthenticationProvider.

Acegi Security includes a production-quality AuthenticationProvider implementation called DaoAuthenticationProvider. This authentication provider is compatible with all of the authentication mechanisms that generate a UsernamePasswordAuthenticationToken, and is probably the most commonly used provider in the framework. Like most of the other authentication providers, the DaoAuthenticationProvider leverages a UserDetailsService in order to lookup the username, password and GrantedAuthority[]s. Unlike most of the other authentication providers that leverage UserDetailsService, this authentication provider actually requires the password to be presented, and the provider will actually evaluate the validity or otherwise of the password presented in an authentication request object.

Aside from adding DaoAuthenticationProvider to your ProviderManager list (as discussed at the start of this part of the reference guide), and ensuring a suitable authentication mechanism is configured to present a UsernamePasswordAuthenticationToken, the configuration of the provider itself is rather simple:

<bean id="daoAuthenticationProvider" class="org.acegisecurity.providers.dao.DaoAuthenticationProvider">
  <property name="userDetailsService"><ref bean="inMemoryDaoImpl"/></property>
  <property name="saltSource"><ref bean="saltSource"/></property>
  <property name="passwordEncoder"><ref bean="passwordEncoder"/></property>
</bean>        

The PasswordEncoder and SaltSource are optional. A PasswordEncoder provides encoding and decoding of passwords presented in the UserDetails object that is returned from the configured UserDetailsService. A SaltSource enables the passwords to be populated with a "salt", which enhances the security of the passwords in the authentication repository. PasswordEncoder implementations are provided with Acegi Security covering MD5, SHA and cleartext encodings. Two SaltSource implementations are also provided: SystemWideSaltSource which encodes all passwords with the same salt, and ReflectionSaltSource, which inspects a given property of the returned UserDetails object to obtain the salt. Please refer to the JavaDocs for further details on these optional features.

In addition to the properties above, the DaoAuthenticationProvider supports optional caching of UserDetails objects. The UserCache interface enables the DaoAuthenticationProvider to place a UserDetails object into the cache, and retrieve it from the cache upon subsequent authentication attempts for the same username. By default the DaoAuthenticationProvider uses the NullUserCache, which performs no caching. A usable caching implementation is also provided, EhCacheBasedUserCache, which is configured as follows:

<bean id="daoAuthenticationProvider" class="org.acegisecurity.providers.dao.DaoAuthenticationProvider">
  <property name="userDetailsService"><ref bean="userDetailsService"/></property>
  <property name="userCache"><ref bean="userCache"/></property>
</bean>

<bean id="cacheManager" class="org.springframework.cache.ehcache.EhCacheManagerFactoryBean">
  <property name="configLocation">
    <value>classpath:/ehcache-failsafe.xml</value>
  </property>
</bean>

<bean id="userCacheBackend" class="org.springframework.cache.ehcache.EhCacheFactoryBean">
  <property name="cacheManager">
    <ref local="cacheManager"/>
  </property>
  <property name="cacheName">
    <value>userCache</value>
  </property>
</bean>

<bean id="userCache" class="org.acegisecurity.providers.dao.cache.EhCacheBasedUserCache">
  <property name="cache"><ref local="userCacheBackend"/></property>
</bean>        

All Acegi Security EH-CACHE implementations (including EhCacheBasedUserCache) require an EH-CACHE Cache object. The Cache object can be obtained from wherever you like, although we recommend you use Spring's factory classes as shown in the above configuration. If using Spring's factory classes, please refer to the Spring documentation for further details on how to optimise the cache storage location, memory usage, eviction policies, timeouts etc.

A design decision was made not to support account locking in the DaoAuthenticationProvider, as doing so would have increased the complexity of the UserDetailsService interface. For instance, a method would be required to increase the count of unsuccessful authentication attempts. Such functionality could be easily provided by leveraging the application event publishing features discussed below.

DaoAuthenticationProvider returns an Authentication object which in turn has its principal property set. The principal will be either a String