Remote app setup

In some incremental upgrade scenarios, it's useful for the new ASP.NET Core app to be able to communicate with the original ASP.NET app.

Specifically, this capability is used, currently, for remote app authentication and remote session features.

Configuration

To enable the ASP.NET Core app to communicate with the ASP.NET app, it's necessary to make a couple small changes to each app.

ASP.NET app configuration

To set up the ASP.NET app to be able to receive requests from the ASP.NET Core app:

  1. Install the nuget package Microsoft.AspNetCore.SystemWebAdapters.FrameworkServices
  2. Call the AddRemoteAppServer extension method on the ISystemWebAdapterBuilder:
SystemWebAdapterConfiguration.AddSystemWebAdapters(this)
    .AddRemoteAppServer(options =>
    {
        // ApiKey is a string representing a GUID
        options.ApiKey = ConfigurationManager.AppSettings["RemoteAppApiKey"];
    });

In the options configuration method passed to the AddRemoteAppServer call, an API key must be specified. The API key is:

  • Used to secure the endpoint so that only trusted callers can make requests to it.
  • The same API key provided to the ASP.NET Core app when it is configured.
  • A string and must be parsable as a GUID. Hyphens in the key are optional.
  1. Optional : Add the SystemWebAdapterModule module to the web.config if it wasn't already added by NuGet. The SystemWebAdapterModule module is not added automatically when using SDK style projects for ASP.NET Core.
  <system.webServer>
    <modules>
+      <remove name="SystemWebAdapterModule" />
+      <add name="SystemWebAdapterModule" type="Microsoft.AspNetCore.SystemWebAdapters.SystemWebAdapterModule, Microsoft.AspNetCore.SystemWebAdapters.FrameworkServices" preCondition="managedHandler" />
    </modules>
</system.webServer>

ASP.NET Core app

To set up the ASP.NET Core app to be able to send requests to the ASP.NET app, you need to make a similar change, calling AddRemoteApp after registering System.Web adapter services with AddSystemWebAdapters.

builder.Services.AddSystemWebAdapters()
    .AddRemoteAppClient(options =>
    {
        options.RemoteAppUrl = new(builder.Configuration["ReverseProxy:Clusters:fallbackCluster:Destinations:fallbackApp:Address"]);
        options.ApiKey = builder.Configuration["RemoteAppApiKey"];
    });

In the preceding code:

  • The AddRemoteApp call is used to configure the remote app's URL and the shared secret API key.
  • The RemoteAppUrl property specifies the URL of the ASP.NET Framework app that the ASP.NET Core app communicates with. In this example, the URL is read from an existing configuration setting used by the YARP proxy that proxies requests to the ASP.NET Framework app as part of the incremental migration's strangler fig pattern.

With both the ASP.NET and ASP.NET Core app updated, extension methods can now be used to set up remote app authentication or remote session, as needed.

Securing the remote app connection

Because remote app features involve serving requests on new endpoints from the ASP.NET app, it's important that communication to and from the ASP.NET app be secure.

First, make sure that the API key string used to authenticate the ASP.NET Core app with the ASP.NET app is unique and kept secret. It is a best practice to not store the API key in source control. Instead, load it at runtime from a secure source such as Azure Key Vault or other secure runtime configuration. In order to encourage secure API keys, remote app connections require that the keys be non-empty GUIDs (128-bit hex numbers).

Second, because it's important for the ASP.NET Core app to be able to trust that it is requesting information from the correct ASP.NET app, the ASP.NET app should use HTTPS in any production scenarios so that the ASP.NET Core app can know responses are being served by a trusted source.