Windows Identity Foundation, first steps…

I’ve been slowly working through the excellent book Programming Windows Identity Foundation by Vittorio Bertocci. I was getting a little restless though and wanted to see some code so I found this walkthrough and decided to play along. Things didn’t go as smoothly as I had hoped but I did learn more than I bargained for…

One of the first requirements to get the sample running is to ensure that you have SSL enabled on your default web site. This is not a common task for most developers so I’ll elaborate a little:

Setting up HTTPS

IIS Manager supports the creation of a self-signed certificate which is sufficient for development purposes. The server configuration provides a‘Server Certificates’ option as below, in the Actions menu there is a ‘Create Self-Signed Certificate…’ item.


There’s not much to the certificate creation process, enter a friendly name for the certificate. In my case I lacked imagination and went with ‘TestCertficateForWIF’. The certificate is created in the machine certificate store so running certmgr.msc doesn’t help as it opens the user store. Instead I ran the mmc.exe directly and added the certificate manager snap-in explicitly, when asked to choose a store I went with the local machine store.




Looking in the Personal | Certificates node reveals the newly created certificate.


Setting up an HTTPS binding for the Default Web Site is now possible. Select the site in IIS manager and then choose to Edit Bindings… from the context menu.


The dialog allows you to add a new HTTPS binding, you just select the certificate you want to use as part of the encryption process.


I next ran through the various steps in the walkthrough but when I tried the run the completed sample I got a KeySet error.

Additional Notes:

  • The certificate name for the DemoSTS web.config only requires CN=, not two.


  • As we are using Windows authentication the console client does not need to pass credentials explicitly. When I set the credentials manually to a local test account I would see my domain account as the name in the returned claim.


Troubleshooting the Sample

A quick search suggested that the AppPool account my services were running as did not have access to the private key of the certificate. OK, back into the machine certificate store and ‘Manage Private Keys…’ for the certificate.


The web applications for the services were mapped to the ApplicationPoolIdentity (I’m running IIS7.5) so I tried adding the read right to the ‘IIS AppPool\DefaultAppPool’ account. This didn’t seem to help so I resorted to creating a specific service account and assigning it the read permission for the certificate.


I created a new application pool to run as this new ‘service.sts’ user and set the web applications use this application pool. This was good and resolved by KeySet error but I was now getting a fault back from my secure WCF service. After a little head scratching I fired up Fiddler to watch the traffic:


OK – I could see the secure WCF service calling the DemoSTS, the DemoSTS doing it’s work and then calling back to the secure service, then a 500 failure. Looking at the response message for the 500:


For some reason I was getting an ‘Invalid Security Token’ error. I knew the error was in the secure WCF service but not much more. While looking through the web.config for the service, I found commented out trace configuration:


So I enabled the tracing and re-ran the client. The WIFTrace.e2e file popped into the service directory and I used the Microsoft Service Trace Viewer to look at the log:


Looking at the error detail:


‘The issuer of the security token was not recognized by the IssuerNameRegistry…’, that looked familiar so back to the web.config.


<service name=”SecureWCFService.Service”>


<add value=”; />


<issuerNameRegistry type=”Microsoft.IdentityModel.Tokens.ConfigurationBasedIssuerNameRegistry, Microsoft.IdentityModel, Version=, Culture=neutral, PublicKeyToken=31bf3856ad364e35″>


<add thumbprint=”?????????????????????????????” name=”; />





I’ve removed the actual thumbprint, but here was where the service was configured to accept tokens from a STS using a particular certificate identified by it’s thumbprint. I needed the thumbprint of the certificate I had created, easily done via PowerShell:

> $certificate = Get-ChildItem -Path Cert:\LocalMachine\My | where { $_.Subject -match ‘CN\’ }


The thumbprint provided by PowerShell did not match my web.config so I updated the config.

Happy days, the sample now ran:


Workflow Services & MSMQ Revisited

I’ recently dusted off a WCF sample I’d written and blogged about a year or two ago. During the process of getting it to work again, I discovered the blog posting is incorrect so I’m reposting with corrections and additional explanation.

Tom Hollander published a great set of posts on this topic which I needed…

We needed a quick proof of concept to show that a workflow service could be activated via a message sent over MSMQ. First part was workflow design and coding, this was the easy part. All I wanted to do was accept a custom type, in this case a TimeEntry, from a SubmitTime service operation that belonged to an ITimeEntryContract. On receiving the time entry I would simple log the fact that it arrived into the event log. This is pretty much the “Hello, World!” of the services demos. The second part was getting the configuration correct…

One of the promises of WCF is that it gives us a unified communication model regardless of the protocol: net.tcp, http, msmq, net.pipe and it does. The best description I’ve heard for WCF is that it is a channel factory, and you configure the channels declaratively in the .config file (you can of course use code too if you prefer). The key benefit is that the service contract and implementation, for the main part, can be channel agnostic. Of course there are the exceptions to prove the rule such as a void return being required by an MSMQ channel but for the most part it holds. As it turns out, it was true that I needed to make no changes to code to move from a default http endpoint to a MSMQ endpoint. I need need to do a lot of configuration and setup though which is not that well documented. This post hopes to correct that is some small way.

First up the easy part, writing the code.

In Visual Studio 2010 I started a new ‘WCF Workflow Service Application’ project. First I define my TimeEntry model class:

using System;

namespace QueuedWorkflowService.Service {
    public class TimeEntry {
        public Guid TimekeeperId { get; set; }
        public Guid MatterId { get; set; }
        public TimeSpan Duration { get; set; }

        public override string ToString() {
            return string.Format(“Timekeeper: {0}, Matter: {1}, Duration: {2}”, TimekeeperId, MatterId, Duration);

Then I defined a code activity to write to the time entry provided into the event log:

using System;
using System.Diagnostics;
using System.Activities;

namespace QueuedWorkflowService.Service {
    public sealed class DebugLog : CodeActivity {
        public InArgument<string> Text { get; set; }

        protected override void Execute(CodeActivityContext context) {
            string message = string.Format(“Server [{0}] – Queued Workflow Service – debug :{1}”, DateTime.Now, context.GetValue(this.Text));
            EventLog.WriteEntry(“Queued Service Example”, message, EventLogEntryType.Information);

All the C# code is now written and I create my workflow:

I need a variable to hold my time entry so I define one at the scope of the service:

The project template creates the CorrelationHandle for me but we won’t be using it.

The receive activity is configured as follows:

With the Content specified as:

This is such a simple service that I don’t need any correlation between messages, it just receives and processes the message without communicating back to the sender of the message. Therefore I also cleared out the CorrelatesOn and the CorrelationInitializer properties.

Finally I set up the Debug activity to write the time entry to the event log:

That’s it! I’m done. This now runs using the default binding introduced in WCF4 (http and net.tcp). Starting up the project launches my service and the WCF test client is also opened pointing to my new service. The service is running in Cassini, the local web server built into the Visual Studio debugging environment.

15 minutes, or there about, to build a workflow service. What follows a summary of the steps discovered over the next 4 hours try to convert this sample from using an http endpoint to an msmq endpoint.

Default Behaviour
One of the key messages Microsoft heard from the WCF 3 community was that configuration was too hard. To even get started using WCF you had to understand a mountain of new terms and concepts including: channels, address, binding, contract, behaviours,… the response to this in .NET 4 is defaults. If you don’t specify an endpoint, binding etc then WCF creates a default one for you based upon your machine configuration settings. This makes getting a service up and running a very straightforward experience. BUT as soon as you want to step outside of the defaults, you need the same knowledge that you needed in the WCF 3 world.

Here’s the web.config I ended up with after a couple of hours, the MSMQ settings were a voyage of personal discovery… (

<?xml version=”1.0″ encoding=”utf-8″?>
    <compilation debug=”true” targetFramework=”4.0″ />
      <service name=”TimeEntryService”>
            contract=”ITimeEntryContract” />

            contract=”ITimeEntryContract” />
            contract=”IMetadataExchange” />
          <serviceMetadata httpGetEnabled=”true” />
          <security mode=”None”>
            <message clientCredentialType=”None” />
                msmqProtectionLevel=”None” />

          <security mode=”None”>
            <message clientCredentialType=”None” />
                msmqProtectionLevel=”None” />
                <add relativeVirtualPath=”TimeEntryService.xamlx” />

The two important sections are the endpoint and the netMsmqBinding sections. A single service is defined that exposes two MSMQ endpoints, a transactional endpoint and a non-transactional one. This was done to demonstrate the changes required in the netMsmqBinding to support a transactional queue over a non-transactional queue; namely the durable and exactlyOnce attributes. In both cases no security is enabled. I had to do this to get the simplest example to work. Note that the WCF address for the queue does not include a $ suffix on the private queue name and matches the Uri of the service.

We still have some way to go to get this to work, we need a number of services to be installed and running on the workstation:

• Message Queuing (MSMQ)
• Net.Msmq Listener Adapter (NetMsmqActivator)
• Windows Process Activation Service (WAS)

I also ensured that AppFabric was running as this is the easiest way to start the debugging process:
• AppFabric Event Collection Service (AppFabricEventCollectionService)

If you don’t have these services registered on your workstation you will need to go into the ‘Programs and Features’ control panel, then ‘Turn Windows features on or off’ to enable them (Windows 7).

With the services installed and started you need to create a private message queue to map the endpoint to (see : ). The queue name must match the Uri of the service.


The sample is configured to run without security on the queues, i.e. the queues are not authorized. You must allow the anonymous login ‘send’ rights on the queues. If you don’t, the messages will be delivered but the WAS listener will not be able to pick up the messages from the queue.


If you have problems and do not see the message delivered to the correct queue, have a look in the system Dead Letter queues.


You also need to change your VS2010 project to use IIS as the host rather than Cassini. On the project properties dialog, open the Web tab:

As I wanted events from this service to be added to my AppFabric monitoring store, I also added a connection string to the mapped web application and then configured AppFAbric monitoring to use that connection.

And in AppFabric configuration:

Finally you also need to enable the correct protocols on the web application (Manage Application… | Advanced Settings):

I’ve added in net.msmq for queuing support and also net.pipe for the workflow control endpoint.

Make sure that the user the application pool is running as has access to read and write to the queue.

With all the server configured I then wrote a simple WPF test application that used a service reference generated by VS2010, this creates the appropriate client side WCF configuration. The button click handler called the service proxy directly:

private void submitTimeEntryButton_Click(object sender, RoutedEventArgs e) {
    using (TimeEntryContractClient proxy = new TimeEntryContractClient(“QueuedTimeEntryContract”)) {
        TimeEntry timeEntry = new TimeEntry {
                                            TimekeeperId = Guid.NewGuid(),
                                            MatterId = Guid.NewGuid(),
                                            Duration = new TimeSpan(0, 4, 0, 0)

        string message = string.Format(“Client [{3}]- TimekeeperId: {0}, MatterId: {1}, Duration: {2}”,

        EventLog.WriteEntry(“Queued Service Example”, message, EventLogEntryType.Information);

And the awesome UI:

Click the button and you get entries in the event log, a client event and the server event:


I made no changes to the code to move from an http endpoint to a MSMQ endpoint, but it’s not as simple as tweaking the config and you’re good to go. I’d love to see some tooling in VS2010 or VS vNext to take some of the pain away from WCF config, similar to the tooling AppFabric adds into IIS. Until that happens, there are plenty of angle brackets to deal with.

Creating a generic ping proxy

In the previous post we walked through the steps required to implement a service monitoring ‘Ping’ operation as a WCF endpoint behavior. This allows us to add the ping functionality to an endpoint using WCF configuration alone. Following on from this, here’s a generic implementation of a Http proxy class to call Ping on any service.

Last time we left off having created a proxy to a test service using the WcfTestClient application.

Ping operation from metadata in the WCF Test Client

Switching to the XML view lets us see the SOAP request and reply message for a ping.

<s:Envelope xmlns:s="">
    <Action s:mustUnderstand="1" xmlns=""></Action>
    <Ping xmlns="" />

and the response:

<s:Envelope xmlns:s="">
  <s:Header />
    <PingResponse xmlns="">

The namespace highlighted in green is the namespace we gave to our service as part of its ServiceContract attribution:

public interface IService1 {

The SOAP message varies from service to service according to the namespace of the contract and the type of the contract. If we know these two values then we can construct the appropriate SOAP message for a generic client.

private static void CallWcfServiceUsingGenericProxy() {
    const string address = @"http://localhost/TestService/Service1.svc";
    const string serviceContractType = "IService1";
    const string serviceContractNamespace = @"";
    Console.WriteLine("Pinging WCF service using the generic proxy...");
    DateTime utcStart = DateTime.UtcNow;
    string response = PingService(serviceContractType, serviceContractNamespace, address);
    DateTime utcFinished = DateTime.UtcNow;
    DateTime serverPingTimeUtc = ProcessPingResponse(response, serviceContractNamespace);
    WriteTimingMessage(utcStart, serverPingTimeUtc, utcFinished);

The calling of the generic proxy is separated from the processing of the response so that a timing can be made around just the communication.

private static string PingService(string serviceContractType, string contractNamespace, string address) {
    string pingSoapMessage = string.Format(@"", contractNamespace);
    if (!contractNamespace.EndsWith("/")) { contractNamespace = contractNamespace + "/";}
    WebClient pingClient = new WebClient();
    pingClient.Headers.Add("Content-Type", "text/xml; charset=utf-8");
    pingClient.Headers.Add("SOAPAction", string.Format(@"""{0}{1}/Ping""", contractNamespace, serviceContractType));
    string response = pingClient.UploadString(address, pingSoapMessage);
    return response;

To Ping the service, we construct the SOAP message and use a WebClient to make the call. The web client requires headers to be added for the content type and the SOAPAction which tells the WCF Dispatcher which method we want to call.

To process the SOAP message returned from the Ping message we use:

private static DateTime ProcessPingResponse(string response, string contractNamespace) {
    XDocument responseXml = XDocument.Parse(response);
    XElement pingTime = responseXml.Descendants(XName.Get("PingResult", contractNamespace)).Single();
    DateTime serverPingTimeUtc = DateTime.Parse(pingTime.Value).ToUniversalTime();
    return serverPingTimeUtc;

Now we have the UTC DateTime from the server when it processed the response.

All good, if we know the contract namespace and interface type for the service we can ping it. As we saw, this information is attributed on the service contract class. To simplify the proxy code a little, we can use reflection to determine this information given just the contract type.

public class WcfServicePinger<TServiceContract> where TServiceContract : class {
    public DateTime Ping(string addressUri) {
        string @namespace = string.Empty;
        ServiceContractAttribute attribute = typeof(TServiceContract)
            .GetCustomAttributes(typeof(ServiceContractAttribute), true)
            .FirstOrDefault() as ServiceContractAttribute;
        if(attribute == null) {
            throw new ArgumentException("The specified type {0} is not a WCF service contract.", typeof(TServiceContract).Name);
        if(string.IsNullOrWhiteSpace(attribute.Namespace)) {
            @namespace = "";
        } else {
            @namespace = attribute.Namespace;
        return new WcfServicePinger().Ping(addressUri, typeof(TServiceContract).Name, @namespace);

The non-generic WcfServicePinger calls our previous code, as above:

public class WcfServicePinger {
    public DateTime Ping(string addressUri, string serviceContractTypename, string serviceContractNamespace) {
        string response = PingService(serviceContractTypename, serviceContractNamespace, addressUri);
        DateTime serverPingTimeUtc = ProcessPingResponse(response, serviceContractNamespace);
        return serverPingTimeUtc;

So in the end we would use:

DateTime serverTime = new WcfServicePinger<IService1>().Ping(“http://localhost/TestService/Service1.svc”);

Note that we have constructed a generic proxy that calls an Http endpoint. I did try to construct a net.tcp generic proxy class too but it broke my time box. The Ping method can be called via net.tcp using a proxy generated from the endpoint metadata.

Checking WCF Service Availability using an Endpoint Behavior

A common operational requirement for an SOA is the ability to determine if a service is available. Just as it is common to ping a machine, we also want to be able to ping an individual service to determine that it is running and able to respond to messages.

Our first pass at solving this issue was to introduce an IPingable interface that each of our service facades would implement. The code was pushed into our service base class and the software factory updated as appropriate. However, this didn’t feel quite right. Other service extensions such as the metadata endpoint were not so invasive, it can be established by a simple addition to the service configuration file. We felt we wanted the same configurable nature for a ping mechanism and so set out to figure out how to create a custom WCF behavior.

This post walks through the creation of a WCF Endpoint Behavior which will add a Ping() method to a service. The method returns the DateTime, in UTC, stating when the server processed the request. This can be used to establish how long a basic roundtrip to the service is taking for both the request and the reply. The use of an endpoint behavior allows us to add this functionality to any service, therefore we can add it retrospectively to services that we have already created without needing to change them.

The source code for this post is available from my DropBox.

Let’s start by looking at how we would configure such an endpoint behavior. For an existing service, we would copy the assembly containing the endpoint behavior into the bin folder for the service and then add some entries into the web.config:

                <add name="Ping" type="CustomWcfBehaviors.PingEndpointBehavior, CustomWcfBehaviors" />

We register the type [CustomWcfBehaviors.PingEndpointBehavior in assembly CustomWcfBehaviors] that is our custom behavior as a behavior extension and then declare that we want to use the extension. The configuration above sets up a default behavior that will be applied to all endpoints since we haven’t explicitly named it. That’s it, fetch the metadata for the service it will now contain a Ping() method.

The configuration of the behavior is reasonably straight forward, let’s look at the implementation…

namespace CustomWcfBehaviors {
    public class PingEndpointBehavior: BehaviorExtensionElement, IEndpointBehavior {
    private const string PingOperationName = "Ping";
    private const string PingResponse = "PingResponse";

The behavior needs to derive from BehaviorExtensionElement which allows it to be declared in the service configuration as an extension element. We need to introduce two overrides:

// factory method to construct an instance of the behavior
protected override object CreateBehavior(){
    return new PingEndpointBehavior();
// property used to determine the type of the behavior
public override Type BehaviorType{
    get { return typeof(PingEndpointBehavior); }

Now onto the real work, the IEndpointBehavior implementation:

public void Validate(ServiceEndpoint endpoint) {}
public void AddBindingParameters(ServiceEndpoint endpoint, BindingParameterCollection bindingParameters) {}
public void ApplyClientBehavior(ServiceEndpoint endpoint, ClientRuntime clientRuntime) {}

public void ApplyDispatchBehavior(ServiceEndpoint endpoint, EndpointDispatcher endpointDispatcher) {
    if(PingOperationNotDeclaredInContract(endpoint.Contract)) {
    UpdateContractFilter(endpointDispatcher, endpoint.Contract);
    AddPingToDispatcher(endpointDispatcher, endpoint.Contract);

There are four methods that we need to implement but only one that we do any work in. The ApplyDispatchBehavior is where we add our code to manipulate how the endpoint invokes the service operations.

Endpoints map to the ABC of WCF [address, binding and contract]. What we want to do is extend the contract with a new service operation. A service host may allow multiple endpoints to expose the service contract on different protocols. For example, we may choose to host our service using both http and net.tcp. If we use IIS as our service host, this is configured via the Advanced Settings… of the Manage Application… context menu option.

The ‘Enabled Protocols’ lists the protocols we want to use. If we have more than one protocol listed then we have to take care that we don’t attempt to add the new Ping operation multiple times to our contract – once per binding type. To check that we haven’t already added the operation we call a small Linq query…

private bool PingOperationNotDeclaredInContract(ContractDescription contract) {
    return ! contract
        .Where(operationDescription =>

If the Ping operation is not found then we need to add it:

private void AddPingToContractDescription(ContractDescription contractDescription) {
    OperationDescription pingOperationDescription = new
    OperationDescription(PingOperationName, contractDescription);
    MessageDescription inputMessageDescription = new MessageDescription(
        GetAction(contractDescription, PingOperationName), MessageDirection.Input);
    MessageDescription outputMessageDescription = new MessageDescription(
        GetAction(contractDescription, PingResponse), MessageDirection.Output);
    MessagePartDescription returnValue = new MessagePartDescription("PingResult",
    returnValue.Type = typeof(DateTime);
    outputMessageDescription.Body.ReturnValue = returnValue;
    inputMessageDescription.Body.WrapperName = PingOperationName;
    inputMessageDescription.Body.WrapperNamespace = contractDescription.Namespace;
    outputMessageDescription.Body.WrapperName = PingResponse;
    outputMessageDescription.Body.WrapperNamespace = contractDescription.Namespace;
    pingOperationDescription.Behaviors.Add(new DataContractSerializerOperationBehavior(pingOperationDescription));
    pingOperationDescription.Behaviors.Add(new PingOperationBehavior());

Here we are creating an OperationDescription to add to our ContractDescription, this adds the operation specification for our Ping operation to the existing contract. The code to execute is encapsulated by the PingOperationBehavior().

public class PingOperationBehavior : IOperationBehavior {
    public void ApplyDispatchBehavior(OperationDescription operationDescription, DispatchOperation dispatchOperation) {
        dispatchOperation.Invoker = new PingInvoker();
    public void Validate(OperationDescription operationDescription) {}
    public void AddBindingParameters(OperationDescription operationDescription, BindingParameterCollection bindingParameters) {}
    public void ApplyClientBehavior(OperationDescription operationDescription, ClientOperation clientOperation) {}

Similar to the endpoint behavior, we need to implement an interface in this case the IOperationBehavior. The method signatures are similar and we need to fill out the ApplyDispatchBehavior method to call an IOperationInvoker to execute our Ping implementation:

internal class PingInvoker : IOperationInvoker {
    public object[] AllocateInputs() {
        return new object[0];
    public object Invoke(object instance, object[] inputs, out object[] outputs) {
        outputs = new object[0];
        return DateTime.UtcNow;
    public IAsyncResult InvokeBegin(object instance, object[] inputs, AsyncCallback callback, object state) {
        throw new NotImplementedException();
    public object InvokeEnd(object instance, out object[] outputs, IAsyncResult result) {
        throw new NotImplementedException();
    public bool IsSynchronous {
        get { return true; }

The ping operation is so simple that there is no asynchronous implementation. All we do is return the current date time on the server in UTC.

So where are we? Well, we’ve added the Ping method to our existing service contract and mapped it to an IOperationBehavior which will dispatch an IOperationInvoker to call the code.

Next up, we have to update the endpoint dispatcher so that it knows what to do if it receives a Ping request message. The endpoint dispatcher maintains a list of actions that it knows how to action. We need to update this list so that it includes our new ping action. To do this we just refresh the action list from the contract description operations:

private void UpdateContractFilter(EndpointDispatcher endpointDispatcher, ContractDescription contractDescription) {
    string[] actions = (from operationDescription in contractDescription.Operations
                        select GetAction(contractDescription, operationDescription.Name)
    endpointDispatcher.ContractFilter = new ActionMessageFilter(actions);

Finally we need to add a new dispatch operation to the endpoint dispatcher so that it calls our PingOperation when the Ping action is received.

private void AddPingToDispatcher(EndpointDispatcher endpointDispatcher, ContractDescription contractDescription) {
    DispatchOperation pingDispatchOperation = new DispatchOperation(endpointDispatcher.DispatchRuntime,
        GetAction(contractDescription, PingOperationName),
        GetAction(contractDescription, PingResponse));

    pingDispatchOperation.Invoker = new PingInvoker();

private string GetAction(ContractDescription contractDescription, string name) {
    string @namespace = contractDescription.Namespace;
    if(!@namespace.EndsWith("/")) { @namespace = @namespace + "/"; }
    string action = string.Format("{0}{1}/{2}", @namespace, contractDescription.Name, name);
    return action;

Well, we are now done on the server side. We’ve created a couple of new classes:
• PingEndpointBehavior
• PingOperationBehavior
• PingInvoker

These three classes allow us to add a Ping method to a service by adding the Ping behavior via the service configuration file.

To test this, you can use the WcfTestClient application. The sample code demonstrates this by calling Ping on a standard WCF and WF service created by Visual Studio, see my DropBox.

In the next post I’ll discuss how we create a generic service proxy to call Ping.

UPDATE: the code works for the basicHttpBinding but not for wsHttpBinding.

UPDATE: Code now available from

Securing WF & WCF Services using Windows Authentication

To finish off the DEV404 session Pete and I presented at TechEd NZ, I gave a brief run through of the steps required to get Windows Authentication working in a load balanced environment using kerberos. Given the number of camera phones that appeared for snaps I’m going to assume this is a common problem with a non-intuitive solution…

The product I work on is an on-premise enterprise solution that uses the Windows Identity to provide an authenticated credential against which to authorize user requests. We host our services in IIS/Windows Server AppFabric and take advantage of the Windows Authentication provided by IIS. This allows one of two protocols to be used: kerberos and NTLM, which have quite separate characteristics.

Why Use Kerberos?
There are two main reasons we want to use kerberos over NTLM:

1. Performance: NTLM uses a challenge response pattern for authentication which leads to a high network utilization. During performance testing we saw a high volume of NTLM challenges which ultimately throttled our ability to serve requests. Kerberos uses tickets which can be cached permitted a better performing protocol.

1. Double hops: NTLM does not flow credentials – the canonical example is a user requesting serviceA on server1 to access a secured resource on server2. Server1 cannot flow the users identity to server2.

Kerberos and Load Balancing
We want to run our services within a load balanced cluster to avoid single points of failure and to be able to grow resources to meet demand as required, without having to adopt bigger tin. The default configuration of IIS does not encourage this… the Application Pools run as a local machine account. This is a significant issue for Kerberos because of the manner in which the protocol encrypts the tickets passed between client, TGS and target server. The password of the account running the service is used to encrypt tickets so that only a process running under that account can decrypt the message. The default use of a machine specific account prevents a ticket granting access to serviceX on server A also being used to access serviceX on server B.

The following steps are required to fix this:

1. Use a common domain account for the applications pools.

We use a DOMAIN\ account to run our services. This domain account is granted log on as a service and log on as a batch job rights on each of the application servers.

2. Register an SPN mapping the service class to the account.

We run our services on HTTP and so register the load balancer address with the domain account used to run the services:

>setspn -a HTTP/clusteraddress serviceAccount

We are using the WCF BasicHttpBinding which does not require the client to ensure the service is running as a particular user (to prevent man in the middle attacks). If you are using any other type of binding then the client needs to state who it expects the service to be running as.

3. Configure IIS to use the application pool account rather than a machine account

system.webServer/security/authentication/windowsAuthentication useAppPoolCredentials must be set to true.

4. Configure IIS to allow kerberos authentication tokens to be cached

system.webServer/security/authentication/windowsAuthentication authPersistNonNTLM must be set to true.

See also

5. Ensure the cluster address is considered to be in the Local Intranet zone

Kerberos tokens are not supported in the Internet zone, therefore the URL for your services must be considered to be trusted. The standard way to implement this is to roll out a group policy that adds your domain to the local intranet zone settings.

The slide deck for the talk is available from

A Tale of Two Services

Now back in New Zealand after two weeks in the US, first week at TechEd and then a week in our US development centre. I finally feel free of jet lag and so it’s time to make good on a promise to write up a couple of samples I didn’t show at TechEd. The first is a quick introduction to authoring services…

The source code to accompany this post can be downloaded from from the TechEd2010 folder. The sample code is in the archive

A service is simply a piece of software that provides some functionality, access to this functionality is formalized into a contract. A service is often hosted in a separate process and utilized by a number of different consumers. The service does not know anything about the consumer, it just performs some work on their request. Between the consumer and service is most likely a process, machine and possibly a network boundary, therefore any data to be exchanged must be serializable. For the consumer to call the service, it must know where it lives, therefore the service has an address. The consumer must also be able to understand and be understood by the service, the supported communication protocols are captured as bindings. So there we have the ABC of Windows Communication Foundation; the Address, the Binding and the Contract.

Services in Code

With each release of Visual Studio, the key use cases that Microsoft is targeting with its tooling become easier to perform. In VS2010 the ease of service authoring and hosting has taken a leap forward and the code line count required to implement a service dropped. Let’s look at a very simple service that provides a random answer to a question, a Magic Eight Ball service. The contract for the magic eight ball is very simple and is captured as the following class:

using System.ServiceModel;

namespace MagicEightBall.CodedService {
    public interface MagicEightBallContract {
        string AskQuestion(string question);

There is a single method that takes a string containing a question and returns a string containing the answer. The System.ServiceModel namespace is the hint that we are going to use WCF to take care of our service. To provide an implementation of the service we have the following code.

using System;

namespace MagicEightBall.CodedService {
    public class MagicEightBallService : MagicEightBallContract {
        public string AskQuestion(string question) {
            return EightBall.Shake();

    internal sealed class EightBall {
        private readonly static Random random = new Random();
        private readonly static string[] answers = { "Yes", "No", "Ask again", "Definitely", "Bad idea", "Perhaps", "Unsure" };

        public static string Shake(){
            return answers[random.Next(0, answers.Length)];

The eight ball is captured as a simple class with a Shake method, the service is not enforcing any validation such as ensuring a question is asked to keep things simple. Note that there is no System.ServiceModel using statement, this is vanilla .NET. We have a service contract and an implementation, our coding is complete. The next step is to host the service and allow our consumers to call it. The service host can be implemented in a number of ways, for this example we are going to use WAS (Windows Process Activation Service) which uses the IIS infrastructure to host the service – we don’t need to write a host, we’ll just use one that Microsoft provides. To access the service, the host exposes an endpoint, the endpoint is composed of the address, binding and contract. One of the criticisms of WCF in .NET 3 was the steep initial learning curve required to get a service hosted and configured. In .NET 4, the idea of defaults has been introduced which greatly reduces the amount of WCF configuration required to get up and running (to the point where it is possible to have no explicit configuration). In the example below we have a little configuration due to a slightly non-standard approach.

<?xml ="1.0"?>
        <add relativeAddress="MagicEightBall.svc" service="MagicEightBall.CodedService.MagicEightBallService"/>
          <serviceMetadata httpGetEnabled="True"/>
          <serviceDebug includeExceptionDetailInFaults="False"/>

Here we are using the element to specify the last part of the address of the service rather than having a separate .svc file. Personally I think this is quite a tidy approach rather than having separate .config and .svc files. The section states that we want to publish metadata about this service and that we want to hide any exception details from consumers of our service. By publishing metadata about our service we allow tooling to generate a proxy class for us that allows our service to be easily called. Visual Studio provides such tooling, from within a project you can add a Service Reference:

The service reference needs to know the address of the service and then from the metadata it creates a class, the proxy, that allows the project to make use of the service. After clicking on OK, the service reference is listed as part of the project, in the sample below the MagicEightBall client is making use of two separate services.

I’m jumping a little bit ahead though, since we haven’t got the service host set up yet. We want to publish the service which we can do from within VS2010 by choosing Publish… from the context menu for the project:

A dialog pops up asking from a location to publish to, I used http://localhost/MagicEightBall which set up a new web application in IIS. By default the web application is set up to support the http protocol. If you want to change this you need to alter the ‘Enabled Protocols’ in the Advanced Settings dialog which is available from the web application context menu in IIS Manager [Manage application | Advanced Settings…].

In the example above I added the net.tcp protocol in addition to http. Note that there is no space between the comma and net.tcp. Putting a space in here will break the enabled protocols! Now we have created and published a WCF service, to test it, point your browser to http://localhost/MagicEightBall/MagicEightBall.svc. You should see the standard metadata page for your service instructing how to create a proxy class and consume it.

[Note that I have .NET 4 registered as the default framework version for IIS and so the default app pool uses .NET 4. The command C:\Windows\Microsoft.NET\Framework64\v4.0.30319\aspnet_regiis.exe -i registers .NET 4 as the default for IIS.]

To test the service, create a console application, add a service reference called MagicEightBallService using the http url. Code to call the service is as follows:

using System;
using System.Text;

using MagicEightBall.Client.MagicEightBallService;

namespace MagicEightBall.Client {
    class Program {
        private const string CodeEndpointNameHttp = "BasicHttpBinding_MagicEightBallContract";
        string question = "Will you answer my questions?";
        string answer = string.Empty;

        using (MagicEightBallContractClient client = new MagicEightBallContractClient(CodeEndpointNameHttp)) {
            answer = client.AskQuestion(question);


In total there is less than 30 lines of code required for us to write to define, implement, host and consume a WCF service.

Services as Workflows
There is an alternative way to author services which uses a workflow to define the service implementation. A functionally equivalent Magic Eight Ball service can be developed as a workflow service as follows…

First create a new project in VS2010 that is a ‘WCF Workflow Service Application’ which sets up the basic send / receive service template. We need to set up a couple of variables within our workflow so click on the variables button at the bottom left having selected the outer scope:

The handle is created by the template so we need to add in the question and answer strings. The variables are used to pass data into and out of activities, the activity is the equivalent of a program statement and acts on the data. In workflow it is possible to author new activities such as the EightBall in the example above. The code for the activity is as follows:

using System;
using System.Activities;

namespace MagicEightBall.WorkflowService {
    public sealed class EightBall : CodeActivity {
        private static Random random = new Random();
        private static string[] answers = { "Yes", "No", "Ask again", "Definitely", "Bad idea", "Perhaps", "Unsure" };

        public InArgument Question { get; set; }

        protected override string Execute(CodeActivityContext context) {
            string question = context.GetValue(this.Question);
            string answer = answers[random.Next(0, answers.Length - 1)];

            return answer;

This activity is essentially the same code as the Eightball class in the original service. The question is captured as an InArgument to the activity and the result is a string, specified as a CodeActivity. Note the use of the CodeActivityContext to get the value of the question from the workflow runtime at execution time.

After compiling the project we get an EightBall activity in our toolbox and this can be dragged into the service workflow. The completed implementation looks as follows with the addition of the EightBall activity:

The EightBall activity needs to have its arguments mapped to variables. The properties of the activity are defined as follows:

In the receive activity, the operation name is changed to AskQuestion and the content is changed to:

Here the receive activity expects to get a string parameter called question which is mapped to the question variable we created earlier. The receive/send activity pairing is analogous to the AskQuestion method in our coded service.

The send activity returns a string and is paired with the Receive Question send activity as shown in the Request field.

Here we are returning the answer that we got from the EightBall activity. This workflow is now functionally equivalent to our original coded example: a string containing a question is passed in, a string containing an answer is returned.

To host the workflow service, the same steps are taken as before. You simply choose to publish the service from Visual Studio into IIS. The service exposes metadata in the same way as the coded service, therefore you can as Visual Studio to generate a service reference for you and then consume the service in the same way as we did for the coded service.

So we have two ways to solve a problem – which is better? It depends on the work that the service is performing. If the service is co-ordinating work across multiple services then a workflow makes sense as it can be easier to visualize the intended flow of control. If the service co-ordination is long running and needs to be persisted then again a workflow makes sense as this long running, durable capability is built right into the workflow service host that Microsoft ships out of the box.

The sample code contains some additional concepts not discussed such as a separate activity library and instrumentation options for service code. The code is small and so hopefully this does not clutter the examples too much.

Migration from .NET 2/3/3.5 to .NET 4

During the TechEd session, the question was asked:

“How do I migrate my services from WCF3 to WCF4?”

The simple as answer is that you recompile your source under .NET 4 and you should be done. .NET 4 is backwards compatible with .NET 2/3.X but you need to recompile for the new CLR (common language runtime).