Using GAE as a OData service for the data layer

One of the many challenges of software development is to allow your application to save its data into a reliable repository from anywhere in the world. With Google App Engine (GAE) and its cloud service, you can put the data layer of your application in the cloud and still use your preferred programming language to create your business and presentation layers. By exposing your GAE database using OData (http://www.odata.org/), you get a flexible way of storing and retrieving your data from any desktop client or mobile application including C# based applications. The best thing is that GAE offers free quotas unlike Microsoft Azure's offerings.

One OData library that I know works in GAE is odata-py ( http://code.google.com/p/odata-py/). After downloading the latest version (using SVN), you can start using the sample GAE application that comes with it. There is however, one small fix you need to make. Modify core.py to prevent an exception when getting the Key from the url when the key is missing; so in the parse_request_url method, instead of match.group('key')[1:-1] make it read:

      key_match = match.group('key')
      if (key_match and isinstance(key_match, str)):
            key = key_match[1:-1]   # strip quotes for URLs with the form http://app/odata.svc/Model('key')
      else:
            key = ''

One way to access the odata service from C#, is to add a service reference in Visual Studio to http://localhost:8080/odata.svc ( 8080 is the default port for GAE development) or to your realgaeapp.appspot.com/odata.svc. Basically the service definition lives in odata.svc in your GAE app. Visual Stuio will generate a proxy for your data service. The following is a sample on how to use the sample service from a C# console app:

static void Main(string[] args)
        {
            GAEODATAService.model.default_container proxy = new GAEODATAService.model.default_container(new Uri("http://localhost:8080/odata.svc"));
            //Uncomment below lines to add a Pet
            proxy.AddToPet(new Pet() { key = "0", name = "Lilo", weight_in_pounds = 10, type = "cat" }); //must be one of set(['bird', 'dog', 'cat'])
            proxy.SaveChanges();
          
            var all_pets = proxy.Pet;
            foreach (var pet in all_pets)
            {
                Console.Out.WriteLine("Pet {0}, weight: {1}, type: {2}", pet.name, pet.weight_in_pounds, pet.type);
            }
            Console.ReadLine();
        }

As you can see from the example, GAE's High Replication Datastore becomes your reliable, fast and free (if you stay below the free quota) data layer for your .NET application. With odata in GAE, you can create your rich desktop or mobile application and have it distributed as a click-once install (which you could possibly host as a static file in GAE), and not be afraid of loosing your user's data because the data is coming and going to the cloud. This way your business and presentation logic get offloaded from your overburdened servers to the user's computer or phone.

It is also important to note that not all the application are data centered CRUDs like the example above, but you can also program your GAE data layer to handle custom web methods.