ASP.NET Internals IIS and the Process Model and Handlers

Introduction

Tools like ASP.NET greatly simplify the development of a complex Web application. Although this is a great thing for general productivity, it can also keep you from understanding the fundamental communications between your Web server and your ASP.NET application code. Furthermore, there will be times when you need to butt-in and intercept HTTP requests, which requires a greater understanding of the process of communication between your Web server of choice, and ASP.NET.
This article explains how IIS and ASP.NET communicate, and describes some techniques for intercepting some of this communication. I’ll review how ASP.NET is configured to handle requests, and how applications and Web services are handled by default.
Then I’ll discuss how you might butt-in to those requests with HTTP handlers, handler factories and modules. You’ll see how they function individually and together through a series of examples.

<a href='http://ad.doubleclick.net/jump/atssn/tt_articles;bkg=FFFFFF;kw=;dcopt=;pos=1;sz=300x250;ptile=1;ord=6742805?''> <img src='http://ad.doubleclick.net/ad/atssn/tt_articles;bkg=FFFFFF;kw=;dcopt=;pos=1;sz=300x250;ptile=1;ord=6742805?'' width='300' height='250' border='0'></a>

From IIS to ASP.NET

IIS communicates with the .NET framework through unmanaged ISAPI extensions: aspnet_isapi.dll and aspnet_filter.dll. The aspnet_isapi.dll is an extension that serves as a request router, and the aspnet_filter.dll is a filter that primarily handles cookieless session states for ASP.NET. These unmanaged components, along with the state Windows service (aspnet_state.exe) and the ASP.NET worker process (aspnet_wp.exe) are the core of the ASP.NET processing model.
When the .NET framework is installed on a machine that has IIS installed, IIS is configured so that requests for specific extensions are handled by aspnet_isapi.dll. As a point of interest, the filter is also configured within IIS..

Requests for ASP.NET resources are forwarded by IIS to ASP.NET via this configured extension. This extension is the bridge between unmanaged and managed code. Before control is passed to your application, an ASP.NET application object must be instantiated (by the runtime) and configuration settings are considered to determine how this request should be handled. Machine.config and collective web.config files are processed collectively to support this process.

For this article, we are specifically interested in the <httpHandlers> configuration section. Settings in this section indicate which .NET type should handle the request. The default settings found in the machine.config file when .NET is installed, are as follows:
<httpHandlers>
   <add verb="*" path="trace.axd" type="System.Web.Handlers.TraceHandler"/>
   <add verb="*" path="*.aspx" type="System.Web.UI.PageHandlerFactory"/>
   <add verb="*" path="*.ashx" type="System.Web.UI.SimpleHandlerFactory"/>
   <add verb="*" path="*.asmx" type="System.Web.Services.Protocols.WebServiceHandlerFactory, 
  System.Web.Services, Version=1.0.3300.0, Culture=neutral,
  PublicKeyToken=b03f5f7f11d50a3a" validate="false"/> 
   <add verb="*" path="*.rem"
  type="System.Runtime.Remoting.Channels.Http.HttpRemotingHandlerFactory, 
  System.Runtime.Remoting, Version=1.0.3300.0, Culture=neutral,
  PublicKeyToken=b77a5c561934e089" validate="false"/>
   <add verb="*" path="*.soap"
  type="System.Runtime.Remoting.Channels.Http.HttpRemotingHandlerFactory, 
  System.Runtime.Remoting, Version=1.0.3300.0, Culture=neutral,
  PublicKeyToken=b77a5c561934e089" validate="false"/>
   <add verb="*" path="*.asax" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.ascx" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.config" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.cs" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.csproj" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.vb" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.vbproj" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.webinfo" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.asp" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.licx" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.resx" type="System.Web.HttpForbiddenHandler"/>
   <add verb="*" path="*.resources" type="System.Web.HttpForbiddenHandler"/>
   <add verb="GET,HEAD" path="*" type="System.Web.StaticFileHandler"/>
   <add verb="*" path="*" type="System.Web.HttpMethodNotAllowedHandler"/>
</httpHandlers>

The <httpHandlers> section indicates which HTTP handler factory, or handler, should be used to handle the request. In summary, the high-level workflow from IIS to your ASP.NET applications is as follows: •IIS receives request for a resource
•For resources mapped to ASP.NET ISAPI Extension (i.e., *.aspx, *.asmx) the request is passed to an unmanaged ASP.NET DLL which communicates with the HttpRuntime object
•The HttpRuntime object handles creation of the HttpApplication object (as needed), and the inspection of configuration settings, then passes control to the appropriate handler for the request
•The handler is created to process request, which ultimately sends a response
The following diagram illustrates this workflow:

<httpHandlers> settings can be modified at the global (machine.config) level, or overridden at the application (web.config) level. In other words, you can specify a different factory or handler to process particular resource requests. For example, to reject requests for *.rem objects, you can edit the machine.config, or the application web.config as follows:
<add verb="*" path="*.rem" type="System.Web.HttpForbiddenHandler"/
Associating the HttpForbiddenHandler with *.rem replaces the default behavior, which be to use HttpRemotingHandlerFactory. If you specify this override in the machine.config, this will impact all applications on the Web server.

HTTP Handlers and Handler Factories

As mentioned above, the ASP.NET runtime relies on HTTP handlers or handler factories to process requests. Configuration file settings associate an HTTP handler or handler factory class with specific resources. Let’s take a closer look at these settings.
The following entry associates the System.Web.UI.PageHandlerFactory class with *.aspx resources, for all HTTP verbs (i.e., GET, POST):
<add verb="*" path="*.aspx" type="System.Web.UI.PageHandlerFactory"/>
This next entry associates System.Web.HttpForbiddenHandler with *.config resources:
<add verb="*" path="*.config" type="System.Web.HttpForbiddenHandler"/>
These are preexisting classes provided with the base class library, but you can also build custom handlers and handler factories, by implementing the IHttpHandlerFactory and IHttpHandler interface (respectively). Before we talk about how to implement custom handlers and factories, let’s review their features. From a high level, HTTP handler factories are specified in order to dynamically return the correct HTTP handler object to manage the requested resource. If an HTTP handler is specified, it is instantiated directly by the runtime. Regardless, the end-result is to invoke a handler for the resource.
After the runtime passes control to a handler, it is the handler’s job to handle the request, instantiating the appropriate ASP.NET server-side objects, and send an HTTP response. For example, PageHandlerFactory returns a System.Web.UI.Page object for the requested *.aspx resource, and System.Web.HttpForbiddenHandler throws an HttpException indicating that the request is not supported.

Implementing IHttpHandler

Similar to ISAPI extensions, handlers provide low-level access to HTTP request and response objects. Implementing a custom handler allows you to process specific resources differently. You can intercept requests for those resources and override the response. For example, if you wanted to log the IP addresses of those requesting forbidden files, you could write a handler that logged information about those requests before throwing an HttpException. You may even want to send back a little love note to those making such request, as shown in the sample code, like this:

To create a custom HTTP handler, create a .NET component that implements IHttpHandler. This interface has the following members:
  • ProcessRequest() – this method is invoked by the runtime to handle the request
  • IsReusable – this property indicates if multiple requests can share the same handler
ProcessRequest() is passed the HttpContext for the request, which can be used to access HttpRequest, HttpResponse and HttpSessionState objects.
NOTE: A handler must implement IRequiresSessionState if it will access the session object.
Here is a simple example of an HTTP handler that writes output to the browser:
 public class ForbiddenLogHandler: IHttpHandler
   {
     public ForbiddenLogHandler()
     {
     }
 
     public virtual void ProcessRequest(HttpContext context)
     {
        
        context.Trace.Write("ForbiddenLogHandler.ProcessRequest()");
 
        HttpResponse rs = context.Response;
        HttpRequest rq = context.Request;
 
        rs.Write("<p><H1>We know who you are...</H1></p>");
 
        rs.Write("You were referred by " + rq.UrlReferrer + "<br>");
        rs.Write("Your IP is" + rq.UserHostAddress + "<br>");
        rs.Write("Your domain is" + rq.UserHostName + "<br>");
        rs.Write("<p>Why were you requesting a restricted resource?</p>");
 
     }
    
     public virtual bool IsReusable
     {
        get { return true; }
     }
   }
The following web.config section configures the ForbiddenLogHandler for any *.cs, *.resx, or *.config files within the application:
<httpHandlers>
<add verb="*" path="*.cs" type="WebHandlers.ForbiddenLogHandler,WebHandlers"/>
   <add verb="*" path="*.resx" type="WebHandlers.ForbiddenLogHandler,WebHandlers"/>
   <add verb="*" path="*.config" type="WebHandlers.ForbiddenLogHandler,WebHandlers"/>
</httpHandlers>

Implementing IHttpHandlerFactory

Where HTTP handlers may be useful in responding to requests for specific resources, a handler factory makes it possible to intercept a request, perform some pre-processing on the request, and then following a factory pattern, create the handler for the resource.
To create a custom handler factory, create a .NET component that implements IHttpHandlerFactory. This interface has the following members:
  • GetHandler() – wadda ya know, this method returns a valid IHttpHandler for the runtime to process the requested resource
  • ReleaseHandler() – provides an opportunity for the factory to reuse a handler
GetHandler() is called by the runtime, and must return null, or a valid IHttpHandler. Here is an example of an HTTP handler factory that counts the number of hits from a particular IP address, if the count is exceeded, returns a custom handler, if not, passes control to the default handler for the resource.
public class HitTrackingHandlerFactory: IHttpHandlerFactory
{
   public HitTrackingHandlerFactory()
   {
   }
 
   public virtual IHttpHandler GetHandler(HttpContext context, 
     String requestType, String url, String pathTranslated)
   {
 
     if (!HitLogHelper.CheckHitCount(context.Request.UserHostAddress, context.Request.UserHostName))
        return new HitsExceededHandler();
        
     Object handler = null;
   
     try 
     {
        String filename = url.Substring(url.LastIndexOf('/')+1);
        String file = filename.Substring(0, filename.IndexOf('.'));
        String ext = filename.Substring(filename.LastIndexOf('.')+1);
 
        if (ext == "aspx")
        {
           return System.Web.UI.PageParser.GetCompiledPageInstance(url, pathTranslated, context);
        }
        else if (ext == "asmx")
        {
           System.Web.Services.Protocols.WebServiceHandlerFactory fact = new
    System.Web.Services.Protocols.WebServiceHandlerFactory();
     
           handler = fact.GetHandler(context, context.Request.RequestType, url, pathTranslated);
        }
        else if (ext == "rem")
        {
           System.Runtime.Remoting.Channels.Http.HttpRemotingHandlerFactory fact = new
    System.Runtime.Remoting.Channels.Http.HttpRemotingHandlerFactory();
           handler = fact.GetHandler(context, context.Request.RequestType, url, pathTranslated);
        }
        else
        {
           throw new HttpException("Unable to process extension *." + ext);
        }
 
     }
     catch (Exception e)
     {
        throw new HttpException("HitTrackingHandlerFactory", e);
     }
     return (IHttpHandler)handler;
   }
 
 
   public virtual void ReleaseHandler(IHttpHandler handler)
   {
     return;
   }
}
 
public class HitsExceededHandler: IHttpHandler
{
   public HitsExceededHandler()
   {
   }
 
   public virtual void ProcessRequest(HttpContext context)
   {
     context.Response.Write("<h1>Number of hits exceeded!</h1><p>");
   }
    
   public virtual bool IsReusable
   {
     get { return true; }
   }
}

Implementing IHttpModule

A module provides an event-driven mechanism for interacting with application events, and HTTP request and response objects. Registered modules are instantiated before HTTP handlers and handler factories, so that they can register for application events. In fact, modules interact with Web applications in two ways: by listening to application events, and by firing custom events to applications that have registered to listen. The latter of these two tightly couples the module with your application. The former provides access to the request, response, and session objects at various stages in the processing of each HTTP request, in an independent component.
To create a custom module, create a .NET component that implements IHttpModule. This interface has the following members:
  • Init() – provides an opportunity for modules to register for application events
  • Dispose() – provides an opportunity to clean up resources allocated by
The following sample Init() method registers for the HttpApplication object’s PreRequestHandlerExecute, PostRequestHandlerExecute, and Error events:
protected void Init() 
{
  application.PreRequestHandlerExecute +=
  (new EventHandler(this.Application_PreRequestHandlerExecute));  
   application.PostRequestHandlerExecute +=
  (new EventHandler(this.Application_PostRequestHandlerExecute)); 
   application.Error += (new EventHandler(this.Application_Error));
}
You can also handle these events in individual applications in the global.asax, however using a module makes it possible to toggle the configuration externally, and deploy a common set of event handlers for multiple applications:
<httpModules>
   <add name="EventModule" type="WebHandlers.EventModule,WebHandlers"/>
   <add name="ErrorHandlerModule" type="WebHandlers.ErrorHandlerModule,WebHandlers"/>
</httpModules>
In such a global error handler, you could provide a catch all error handler that notifies the site administrator of uncaught exceptions.

Handlers vs. Modules

From an architectural perspective, handlers and modules serve a much different purpose. Handlers (and factories returning handlers) provide a way to intercept the request, and redirect how it is handled. The best a module can do in this respect, is send a custom response, and cancel the handler from completing its job. What they have in common is the ability to access to the underlying HttpApplication, HttpRequest, HttpResponse, and HttpSessionState objects (at certain points) so that you can manipulate and/or interact with this data.
The order of events for these components is as follows:
  • HttpApplication is created (if necessary)
  • HttpModules are created (in order of appearance in configuration files)
  • Modules will receive request notification and authentication events, prior to control passing to the appropriate handler.
  • HttpHandlerFactory/HttpHandler is created, to process the request
  • Resulting objects are created (I.e., the Page, Web Service or Remote object)
  • These objects, the application, and any modules continue to interact with one another via events
The following diagram illustrates the default behavior, in order of instantiation:

This next diagram illustrates where custom factories, handlers, modules and extensions would fit in the workflow:

The Code

Code samples supplied with this article demonstrate the order of events when custom handler factories, handlers and modules are all present. Code demonstrates the following:
  • The <trace> element of the web.config file is configured to display trace output in the Web page. When the page is loaded, you’ll be able to see trace statements, and you can also view the trace.axd file like so:

    http://localhost/ASPNETCS/WebResources/trace.axd
  • From the main Web page, you can click on invalid resources to see the results of the ForbiddenLogHandler, which handles *.resx, *.cs, and *.config requests.
  • A custom message is written to the response stream, but the data returned can also be interesting data to collect from those visiting your site, for data mining.
  • The main Web page also has a link to an invalid resource, which will generate an application error. The ErrorHandlerModule is configured to pick up errors, and e-mail the administrator a notification, while also writing custom HTML to the response stream.
  • After 2 passes though the HitTrackingHandlerFactory, you will no longer be able to browse the main page. An XML file is written by the factory, counting hits from particular IPs, and rejecting that IP after 2 hits. Note the file path is currently hard-coded, so you’ll need to create a directory for the file, and modify the code to match. Delete the file to reset the behavior. The factory also demonstrates forwarding the request to be handled by default behavior of ASP.NET configuration, if the request is approved.

Summary

This article should give you the tools you need to determine what architectural model is best for you application, with respect to intercepting HTTP requests, and modifying application behavior globally. For your reading pleasure, here are a few ideas:
  • Customize authentication, to provide a guest account for users that do not have permission
  • Pre-authenticate users for a “trial” usage of your application
  • Toggle logging behavior, to collect useful statistics about your clients, and perform data mining

0 comments:

Post a Comment