As a corporate developer by trade, I don't get much opportunity to create from-the-ground-up web sites; usually it's tweaks, fixes, and new functionality to existing sites. And with hobby sites, I often don't find the challenges I run into with enterprise systems; usually it's starting from Visual Studio's boilerplate project and adding whatever functionality I want to play around with, rarely deploying outside my own machine. So my experience creating a new enterprise-level site was a bit dated, and the technologies to do so have come a long way, and are much more ready to go out of the box. My intention with this post isn't so much to provide any groundbreaking insights, but to just tie together a lot of information in one place to make it easy to create a new site from scratch.

Architecture

One site I created earlier this year had an MVC 3 front end and a WCF 4-driven service layer. Using Visual Studio 2010, these project types are easy enough to add to a new solution. I created a third Class Library project to store common functionality the front end and services layers both needed to access, for example, the DataContract classes that the front end uses to call services in the service layer. By keeping DataContract classes in a separate project, I avoided the need for the front end to have an assembly/project reference directly to the services code, a bit cleaner and more flexible of an SOA implementation.

Consuming the service

Even by this point, VS has given you a lot. You have a working web site and a working service, neither of which do much but are great starting points. To wire up the front end and the services, I needed to create proxy classes and WCF client configuration information. I decided to use the SvcUtil.exe utility provided as part of the Windows SDK, which you should have installed if you installed VS. VS also provides an Add Service Reference command since the .NET 1.x ASMX days, which I've never really liked; it creates several .cs/.disco/etc. files, some of which contained hardcoded URL's, adding duplicate files (*1.cs, *2.cs, etc.) without doing a good job of cleaning up after itself. I've found SvcUtil much cleaner, as it outputs one C# file (containing several proxy classes) and a config file with settings, and it's easier to use to regenerate the proxy classes when the service changes, and to then maintain all your configuration in one place (your Web.config, instead of the Service Reference files). I provided it a reference to a copy of my common assembly so it doesn't try to recreate the data contract classes, had it use the type List<T> for collections, and modified the output files' names and .NET namespace, ending up with a command like:

svcutil.exe /l:cs /o:MyService.cs /config:MyService.config /r:MySite.Common.dll /ct:System.Collections.Generic.List`1 /n:*,MySite.Web.ServiceProxies http://localhost:59999/MyService.svc

I took the generated MyService.cs file and drop it in the web project, under a ServiceProxies folder, matching the namespace and keeping it separate from classes I coded manually. Integrating the config file took a little more work, but only needed to be done once as these settings didn't often change. A great thing Microsoft improved with WCF 4 is configuration; namely, you can use all the default settings and not have to specify them explicitly in your config file. Unfortunately, SvcUtil doesn't generate its config file this way. If you just copy & paste MyService.config's contents into your front end's Web.config, you'll copy a lot of settings you don't need, plus this will get unwieldy if you add more services in the future, each with its own custom binding. Really, as the only mandatory settings are the endpoint's ABC's (address, binding, and contract) you can get away with just this:

<system.serviceModel>
  <client>
    <endpoint address="http://localhost:59999/MyService.svc" binding="wsHttpBinding" contract="MySite.Web.ServiceProxies.IMyService" />
  </client>
</system.serviceModel>

By default, the services project uses basicHttpBinding. As you can see, I switched it to wsHttpBinding, a more modern standard. Using something like netTcpBinding would probably be faster and more efficient since the client & service are both written in .NET, but it requires additional server setup and open ports, whereas switching to wsHttpBinding is much simpler.

From an MVC controller action method, I instantiated the client, and invoked the method for my operation. As with any object that implements IDisposable, I wrapped it in C#'s using() statement, a tidy construct that ensures Dispose gets called no matter what, even if an exception occurs. Unfortunately there are problems with that, as WCF's ClientBase<TChannel> class doesn't implement Dispose according to Microsoft's own usage guidelines. I took an approach similar to Technology Toolbox's fix, except using partial classes instead of a wrapper class to extend the SvcUtil-generated proxy, making the fix more seamless from the controller's perspective, and theoretically, less code I have to change if and when Microsoft fixes this behavior.

User interface

The MVC 3 project template includes jQuery and some other common JavaScript libraries by default. I updated the ones I used to the latest versions using NuGet, available in VS via the Tools > Library Package Manager > Manage NuGet Packages for Solution... > Updates. I also used this dialog to remove packages I wasn't using. Given that it's smart enough to know the difference between the .js and .min.js files, I was hoping it would be smart enough to know which to include during build and publish operations, but this doesn't seem to be the case. I ended up using Cassette to perform the minification and bundling of my JavaScript and CSS files; ASP.NET 4.5 includes this functionality out of the box.

The web client to web server link via jQuery was easy enough. In my JavaScript function, unobtrusively wired up to a button's click event, I called $.ajax, corresponding to an action method that returns a JsonResult, accomplished by passing my model class to the Controller.Json() method, which jQuery helpfully translates from JSON to a JavaScript object.

$.ajax calls weren't perfectly straightforward. I tried using the simpler $.post method instead, but ran into trouble without specifying the contentType parameter, which $.post doesn't have. The url parameter is simple enough, though for flexibility in how the site is deployed, I used MVC's Url.Action method to get the URL, then sent this to JavaScript in a JavaScript string variable. If the request needed input data, I used the JSON.stringify function to convert a JavaScript object with the parameters into a JSON string, which MVC then parses into strongly-typed C# parameters. I also specified "json" for dataType, and "application/json; charset=utf-8" for contentType. For success and error, I provided my success and error handling functions, though success is a bit hairier. "Success" in this context indicates whether the HTTP request succeeds, not whether what you wanted the AJAX call to do on the web server was successful. For example, if you make an AJAX call to retrieve a piece of data, the success handler will be invoked for any 200 OK response, and the error handler will be invoked for failed requests, e.g. a 404 Not Found (if the server rejected the URL you provided in the url parameter) or 500 Internal Server Error (e.g. if your C# code threw an exception that wasn't caught). If an exception was caught and handled, or if the data requested wasn't found, this would likely go through the success handler, which would need to do further examination to verify it did in fact get back the data for which it asked. I discuss this more in the next section.

Logging and exception handling

At this point, I had a working application. If I ran into any errors or unexpected behavior, debugging was easy enough, but of course that's not an option on public web servers. Microsoft Enterprise Library 5.0 filled this gap nicely, with its Logging and Exception Handling functionality. First I installed Enterprise Library; NuGet as outlined above is probably the best way to do so. I needed a total of three assembly references--Microsoft.Practices.EnterpriseLibrary.ExceptionHandling, Microsoft.Practices.EnterpriseLibrary.ExceptionHandling.Logging, and Microsoft.Practices.EnterpriseLibrary.Logging. VS links with the handy Enterprise Library 5.0 Configuration Console, accessible by right-clicking your Web.config and choosing Edit Enterprise Library V5 Configuration. In this console, under Logging Settings, I set up a Rolling Flat File Trace Listener to write to log files but not let them get too large, using a Text Formatter with a simpler template than that provided by default. Logging to a different (or additional) destination is easy enough, but a flat file suited my needs. At this point, I verified it wrote as expected by calling the Microsoft.Practices.EnterpriseLibrary.Logging.Logger.Write method from my C# code.

With those settings verified, I went on to wire up Exception Handling with Logging. Back in the EntLib Configuration Console, under Exception Handling, I used a LoggingExceptionHandler, setting its Logging Category to the category I already had configured in the Logging Settings. Then, from code (e.g. a controller's OnException method, or any action method's catch block), I called the Microsoft.Practices.EnterpriseLibrary.ExceptionHandling.ExceptionPolicy.HandleException method, providing the exception and the exception policy name I had configured in the Exception Handling Settings.

Before I got this configured correctly, when I tried it out, nothing was logged. In working with .NET, I'm used to seeing an exception if something doesn't work or isn't set up correctly, but instead working with these EntLib modules reminds me more of JavaScript (before the "use strict" v5 days)--it just does nothing and leaves you to figure out why, I presume due in part to the listener pattern Microsoft followed with the Enterprise Library. First, I verified logging worked on its own. Then, verifying/correcting where each piece wires up to the next resolved my problem. Your C# code calls into the Exception Handling module, referencing the policy you pass the HandleException method; that policy's configuration contains a LoggingExceptionHandler that references a logCategory; that logCategory should be added in the loggingConfiguration's categorySources section; that category references a listener; that listener should be added in the loggingConfiguration's listeners section, which specifies the name of the log file.

One final note on error handling, as the proper way to handle WCF and MVC errors is a whole other very lengthy discussion. For AJAX calls to MVC action methods, depending on your configuration, an exception thrown here will result in ASP.NET'S Yellow Screen Of Death being sent back as a response, which is at best unnecessarily and uselessly verbose, and at worst a security risk as the internals of your application are exposed to potential hackers. I mitigated this by overriding my controller's OnException method, passing the exception off to the Exception Handling module as above. I created an ErrorModel class with as few properties as possible (e.g. an Error string), sending as little information to the client as possible, to both maximize bandwidth and mitigate risk. I then return an ErrorModel in JSON format for AJAX requests:

if (filterContext.HttpContext.Request.IsAjaxRequest())
{
    filterContext.Result = Json(new ErrorModel(...));
    filterContext.ExceptionHandled = true;
}

My $.ajax calls from the browser get a valid 200 OK response and go into the success handler. Before assuming everything is OK, I check if it's an ErrorModel or a model containing what I requested. If it's an ErrorModel, or null, I pass it to my error handler. If the client needs to handle different errors differently, ErrorModel can contain a flag, error code, string, etc. to differentiate, but again, sending as little information back as possible is ideal.

Summary

As any experienced ASP.NET developer knows, this is a far cry from where ASP.NET started when I began working with it 11 years ago. WCF services are far more powerful than ASMX ones, MVC is in many ways cleaner and certainly more unit test-friendly than Web Forms (if you don't consider the code/markup commingling you're doing again), the Enterprise Library makes error handling and logging almost entirely configuration-driven, AJAX makes a responsive UI more feasible, and jQuery makes JavaScript coding much less painful. It doesn't take much work to get a functional, maintainable, flexible application, though having it actually do something useful is a whole other matter.