I have a WCF Service with HTTP bindings which returns dataset on 500k size.
When using WCF default logging I can see the messages and data being transfered with each message
<system.serviceModel>
<!-- add trace logging -->
<diagnostics wmiProviderEnabled="true">
<messageLogging
logEntireMessage="true"
logMalformedMessages="true"
logMessagesAtServiceLevel="true"
logMessagesAtTransportLevel="true"
maxMessagesToLog="3000"
/>
</diagnostics>
....
<system.diagnostics>
<sources>
<source name="System.ServiceModel.MessageLogging">
<listeners>
<add type="System.Diagnostics.DefaultTraceListener" name="Default">
<filter type="" />
</add>
<add initializeData="c:\nettcpTestLOCALToTEST.xml" type="System.Diagnostics.XmlWriterTraceListener"
name="messages">
<filter type="" />
</add>
</listeners>
</source>
</sources>
</system.diagnostics>
Point is, I am looking for a way to reduce the traffic between server and client, and I have been told that NetTCP is transferring the data binary? Is that correct?
I have set up a test scenario with a NetTCPBinding and when I read the WCF on the client side, the Response Message includes the whole dataset schema and data in XML format. Is it just serialized so the can be written to a log, or was this message transfered binary?
Is the amount of data being transfered with a NetTCP binding smaller than with HTTPBinding? Is it text or binary?
thanks in advance
yes the message will be transfered binary but the Serializer (Datacontractserializer I assume) will serialize the data in XML format:
Use the DataContractSerializer class to serialize and deserialize instances of a type into an XML stream or document
DataContractSerializer
From the docu:
The NetTcpBinding generates a run-time communication stack by default, which uses transport security, TCP for message delivery, and a binary message encoding. This binding is an appropriate system-provided choice for communicating over an Intranet.
NetTcpBinding MSDN
If you opt to implement ISerializable you can use WCF too but you have to implement an DataContractResolver to resolve the types: if the client "knows" the Types (for example you put them into a dll and add them to the client-app) you can use the following example-code (sorry I only have this in F# around but you should find it easy to translate)
This should yield the serialization in more compact form.
type internal SharedTypeResolver() =
inherit System.Runtime.Serialization.DataContractResolver()
let dict = new Xml.XmlDictionary()
override this.TryResolveType(t : Type, declaredT : Type, knownTypeResolver : System.Runtime.Serialization.DataContractResolver, typeName : Xml.XmlDictionaryString byref, typeNamespace : Xml.XmlDictionaryString byref) =
typeNamespace = dict.Add(t.Assembly.FullName)
typeName = dict.Add(t.FullName)
true
override this.ResolveName(typeName : string, typeNamespace : string, declaredType : Type, knownTypeResolver : System.Runtime.Serialization.DataContractResolver) =
let res = knownTypeResolver.ResolveName(typeName, typeNamespace, declaredType, null)
if res = null then Type.GetType(typeName + ", " + typeNamespace) else res
PS: found the same in C#:
public class SharedTypeResolver : DataContractResolver
{
#region Overrides of DataContractResolver
///
/// Override this method to map a data contract type to an xsi:type name and namespace during serialization.
///
///
/// true if mapping succeeded; otherwise, false.
///
/// The type to map.The type declared in the data contract.The known type resolver.The xsi:type name.The xsi:type namespace.
public override bool TryResolveType(Type type, Type declaredType, DataContractResolver knownTypeResolver, out XmlDictionaryString typeName, out XmlDictionaryString typeNamespace)
{
if (!knownTypeResolver.TryResolveType(type, declaredType, null, out typeName, out typeNamespace))
{
var dict = new XmlDictionary(); // nice trick to get the right type for typeName
if (type != null)
{
typeNamespace = dict.Add(type.Assembly.FullName);
typeName = dict.Add(type.FullName);
}
else
{
typeNamespace = dict.Add("noAss");
typeName = dict.Add("noType");
}
}
return true;
}
///
/// Override this method to map the specified xsi:type name and namespace to a data contract type during deserialization.
///
///
/// The type the xsi:type name and namespace is mapped to.
///
/// The xsi:type name to map.The xsi:type namespace to map.The type declared in the data contract.The known type resolver.
public override Type ResolveName(string typeName, string typeNamespace, Type declaredType, DataContractResolver knownTypeResolver)
{
return knownTypeResolver.ResolveName(typeName, typeNamespace, declaredType, null) ??
Type.GetType(typeName + ", " + typeNamespace);
}
(Please note: stackoverflow don't like the assignmentoperator "<-" from F# and i don't know how to circumvent - therefore I used "=")
oh well - I guess I have to say how to add those resolvers to your host:
private static void AddResolver(OperationDescription operationDescription)
{
if (operationDescription == null)
throw new ArgumentNullException();
var serializationBehavior = operationDescription.Behaviors.Find();
if (serializationBehavior == null)
{
serializationBehavior = new DataContractSerializerOperationBehavior(operationDescription);
operationDescription.Behaviors.Add(serializationBehavior);
}
serializationBehavior.DataContractResolver = new SharedTypeResolver();
}
use this with:
var contrDescription = _host.Description.Endpoints[0].Contract;
var description= contrDescription.Operations.Find("MyServiceMethod");
AddResolver(description);
replacing "MyServiceMethod" by the name of your service-method (on call per method or you iterate over all of them)
Related
I am trying to impelemnt content based routing using XPath in wcf.
I have create class library which contains service contract and data contract as following.
[ServiceContract(Namespace = "http://orders/")]
public interface IService5
{
[OperationContract]
string GetData(int value);
}
[DataContract]
public class Quantity
{
[DataMember]
public int value1 { get; set; }
}
I created one service as follows:
public class Service5 : IService5
{
public string GetData(int value)
{
return string.Format("You entered in service 5: {0}", value);
}
}
And I am trying to implement routing based on 'value'
In app.config (inside router project) i hav added following lines for namespace and XPath filter
<namespaceTable>
<add prefix="cc" namespace="http:orders/Quantity/"/>
</namespaceTable>
<filters>
<filter name="All" filterType="XPath" filterData="cc://value1 > 500 " />
But whenever i run the code i get an exception for ' cc://value1 > 500 ' as invalid qualified name exception.
How can i resolve this ?
There are multiple things wrong here:
The class Quantity on which you appear to want to apply the filter does not feature in your service contract at all, so will be entirely absent in the XML for filtering purposes.
The namespace in your router config starts http:orders, when the service contract namespace starts http://orders.
The namespace in your router config contains /Quantity, when the service contract namespace does not.
The filter xpath cc://value1 is not a valid xpath
---
Hey The problem is in the line
filter name="All" filterType="XPath" filterData="cc://value1 > 500 "
It should be
<filter name="All" filterType="XPath" filterData="//cc:value1 > 500 " />
observe cc:// in ur code.
This will solve ur problem
I am using the NetDataContractSerialiser with WCF. This is working well with all our types being serialised. However, a service I am calling is generating the following exception
The formatter threw an exception while trying to deserialize the message:
There was an error while trying to deserialize parameter http://tempuri.org/:xmlServiceObjIN. The InnerException message was 'The deserializer cannot load the type to deserialize because type 'System.Collections.Generic.EnumEqualityComparer`1[[GrantEd.Common.DomainModel.Code.Enums.enumFundingParameterContextKey, GrantEd.Common.DomainModel.Code, Version=14.0.71.0, Culture=neutral, PublicKeyToken=null]]' could not be found in assembly 'mscorlib, Version=2.0.0.0, Culture=neutral, PublicKeyToken=b77a5c561934e089'.
Check that the type being serialized has the same contract as the type being deserialized and the same assembly is used.'. Please see InnerException for more details.
The class being serialised has a property defined as
public IDictionary<enumFundingParameterContextKey, string> Context { get; set; }
the declaration of the enum is
[Serializable]
public enum enumFundingParameterContextKey
{
[EnumMemberAttribute()]
ClientId = 0,
[EnumMemberAttribute()]
EntitlementDefinitionId = 1
}
which is defined in another assmebly.
When I replace the enumeration with int the class deserialises with no problems. Any ideas why using the enum would result the exception?
The reason for using NetDataContractSerializer was to for type information to be available and avoid having to use KnownType
Make the base type as non integer for your enum. e.g. a byte.
Example:
[Serializable] public enum enumFundingParameterContextKey : byte
{
[EnumMemberAttribute()]
ClientId = 0,
[EnumMemberAttribute()]
EntitlementDefinitionId = 1
}
Details are in my following blog post: dotnet-35-to-dotnet-40-enum
Try to mark enum with [DataContract] instead of [Serializable]. But it looks like the enum type is unknown at all on the client.
I get this error when calling my service:
Server Error in '/' Application.
--------------------------------------------------------------------------------
Configuration Error
Description: An error occurred during the processing of a configuration file required to service this request. Please review the specific error details below and modify your configuration file appropriately.
Parser Error Message: There was an error while trying to serialize parameter http://DSD.myCompany.net/DsdWebServices/2011/05/:config. The InnerException message was 'Type 'System.OrdinalComparer' with data contract name 'OrdinalComparer:http://schemas.datacontract.org/2004/07/System' is not expected. Add any types not known statically to the list of known types - for example, by using the KnownTypeAttribute attribute or by adding them to the list of known types passed to DataContractSerializer.'. Please see InnerException for more details.
Source Error:
Line 130: passwordAttemptWindow="10"
Line 131: passwordStrengthRegularExpression=""
Line 132: type="DsdWebsite.Providers.DsdMembershipProvider, DsdWebsite.Providers" />
Line 133: </providers>
Line 134: </membership>
Source File: C:\Development\DSD Website\WebUI\web.config Line: 132
--------------------------------------------------------------------------------
Version Information: Microsoft .NET Framework Version:2.0.50727.5444; ASP.NET Version:2.0.50727.5420
The service is a data service for a membership provider. I created a MembershipUser DTO to move data back and forth across the service. It uses only standard classes: string, int, DateTime. I use Guid instead of object for the providerUserKey.
The interface for the service looks like this:
[ServiceContract(Namespace = "http://DSD.myCompany.net/DsdWebServices/2011/05/")]
[ServiceKnownType(typeof(MembershipUserDTO))]
[ServiceKnownType(typeof(NameValueCollection))]
[ServiceKnownType(typeof(Guid))]
[ServiceKnownType(typeof(DateTime))]
public interface IDsdMembershipProviderService
{
[OperationContract]
void Initialize(string name, NameValueCollection config);
[OperationContract]
MembershipUserDTO CreateUser(string username,
string salt,
string encodedPassword,
...
The DTO looks like this
namespace DsdWebsite.Services.Providers
{
[Serializable]
[DataContract]
[KnownType(typeof(Guid))]
[KnownType(typeof(DateTime))]
public class MembershipUserDTO
{
public MembershipUserDTO(string providerName, string userName, Guid providerUserKey, string email,
string passwordQuestion, string comment, bool isApproved, bool isLockedOut,
DateTime creationDate, DateTime lastLoginDate, DateTime lastActivityDate,
DateTime lastPasswordChangedDate, DateTime lastLockoutDate,
string firstName, string lastName, string cellPhone, string officePhone,
string brokerId, bool isAdmin, bool mustChangePassword)
{
ProviderName= providerName;
UserName = userName;
ProviderUserKey= providerUserKey;
Email= email;
PasswordQuestion= passwordQuestion;
Comment= comment;
IsApproved=isApproved;
IsLockedOut= isLockedOut;
CreationDate= creationDate;
LastLoginDate= lastLoginDate;
LastActivityDate= lastActivityDate;
LastPasswordChangedDate = lastPasswordChangedDate;
LastLockoutDate=lastLockoutDate;
...
Finally, my web.config looks like this:
<membership
defaultProvider="DsdMembershipProvider"
userIsOnlineTimeWindow="15"
hashAlgorithmType=""> <providers>
<clear/>
<add
name="DsdMembershipProvider"
connectionStringName="DsdMembershipConnectionString"
enablePasswordRetrieval="true"
enablePasswordReset="true"
requiresQuestionAndAnswer="true"
applicationName="/DsdWebsite/"
requiresUniqueEmail="true"
passwordFormat="Encrypted"
maxInvalidPasswordAttempts="5"
minRequiredPasswordLength="7"
minRequiredNonalphanumericCharacters="0"
passwordAttemptWindow="10"
passwordStrengthRegularExpression=""
type="DsdWebsite.Providers.DsdMembershipProvider,
DsdWebsite.Providers" />
</providers> </membership>
How can I determine what type or object is causing the error?
Thanks
Use following ServiceKnownTypeAttribute constructor to specify type of class (declaringType) containing the static method methodName that will return service known types:
public ServiceKnownTypeAttribute(
string methodName,
Type declaringType
)
Inside the aforementioned static method add all service known types that are already added (although I think you would do good without DateTime and Guid), and add System.OrdinalComparer as well.
The catch is that System.OrdinalComparer is internal class so you will have to get the type via reflection.
EDIT:
System.OrdinalComparer is part of mscorlib assembly. Basically you can get its type in a following manner:
Type[] types = typeof( string ).Assembly.GetTypes();
and then you can retrieve the wanted type by name (using Linq, add necessary using statements).
Type type = types.Where( x => x.FullName == "System.OrdinalComparer" );
Previous two line can be combined in one, for simplicity done using two lines.
If you need more details, just say.
I need to have 2 families of classes (one on server and one on client side) which are identical in data structure but differs in behavior. Also I suppose that these fmailies will be enough big, thus I don't want to implement intermediate level of DTO and transformations into and from it.
I decided to move in following manner: declare shared assembly with declaration of data and services interfaces like these ones:
public interface ITest
{
string Title { get; set; }
int Value { get; set; }
}
public interface IService
{
ITest GetData();
}
Having these declarations I can implement these interfaces on server side for example basing on Entity Framework (data) and WCF (services). On the client side I can use for example Dependency Properties (data) and WCF (service).
When I started trying to implement this, I met several troubes.
First one was about server side of WCF - it simply do not want to work with interfaces as return parameters. Thanks to StackOverflow this issue was resolved like here
.
Next problem is that XML rendered by server side includes qulified assembly name of serialized on the server class.
<s:Envelope xmlns:s="http://schemas.xmlsoap.org/soap/envelope/">
<s:Body>
<GetDataResponse xmlns="http://tempuri.org/">
<Test z:Id="1" z:Type="Dist.Server.Model.Test" z:Assembly="Dist.Server, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null" xmlns="http://schemas.datacontract.org/2004/07/Dist.Server.Model" xmlns:i="http://www.w3.org/2001/XMLSchema-instance" xmlns:z="http://schemas.microsoft.com/2003/10/Serialization/">
<Title z:Id="2">Test</Title>
<Value>123</Value>
</Test>
</GetDataResponse>
</s:Body>
</s:Envelope>
Thus during deserialization on client side there was an attempt to load this type. As this type is inaccessible on client side, I had to implement some kind of type mapping. I found that this is quite easy as NetDataContractSerializer used for serialization supports Binder property. Thus I override this property on client side and return correct value (hardcode in meantime, but it's OK for tests).
public class NetBinder : SerializationBinder
{
public override Type BindToType(string assemblyName, string typeName) {
var type = Type.GetType("Client.Test, Client, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null");
return type;
}
}
Now I have following picture:
- Server uses NetDataContractSerializer to serialize response. It uses actual value (calss) during serialization instead of type used in declaration of service (interface).
- Client side receives XML and starts deserialization. To resolve type, NetDataContractSerializer calls my Binder that returns correct type.
- NetDataContractSerializer creates instance of correct type and starts loading of its properties.
And here I got a trouble that I don't know how to resolve. Values of properties are not deserialized. It means that instance of class is created correctly (uninitialized instance created through reflection services), but all properties are in their default values (0 or null).
I tried to play with declaration of class on client side: mark it as [Serializable], implement ISerializable, etc., but nohing is helpful. NetDataContractSerializer requires class to be marked as [DataContract] or [Serializable]. First option leaves properties empty, second one causes exceptions like "</Test> is unexpected, expected is bla-bla-bla_Value_Ending_bla-bla-bla".
Does anybody have any suggestions on how to resolve this last step?
I can provide full sources for better understanding, but I don't know ifI can attach them here...
Thanks in advance.
You could have a look at frameworks like AutoMapper that would take care the transformation to and from DTO. This would make your life much easier.
Instead of using an interface why not create a base class containing only the data and inherit it on both sides by sharing the assembly containing this class. Playing around with the ServiceKnownType should help you fix the last issues.
You may also share the same base classes on both sides and implement the specific logic as extension methods.
Seems that problem was solved enough easily. I created own serializer and used it instead of NetDataContractSerializer. Code is quite simple:
public class MySerializer: XmlObjectSerializer
{
public override void WriteStartObject(XmlDictionaryWriter writer, object graph) {
}
public override void WriteObjectContent(XmlDictionaryWriter writer, object graph) {
var formatter = new XmlSerializer(graph.GetType());
formatter.Serialize(writer, graph);
}
public override void WriteEndObject(XmlDictionaryWriter writer) {
}
public override object ReadObject(XmlDictionaryReader reader, bool verifyObjectName) {
var realType = Type.GetType("Client.Test, Client, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null"); //temporary solution
var formatter = new XmlSerializer(realType);
return formatter.Deserialize(reader);
}
public override bool IsStartObject(XmlDictionaryReader reader) {
return true;//temporary solution
}
}
I checked SOAP that goes from server to client and it's almost the same as NetDataSerializer renders. The only difference is in attribute xmlns="".
Kai, Johann thanksfor your tries.
Using the WCF web programming model one can specify an operation contract like so:
[OperationContract]
[WebGet(ResponseFormat = WebMessageFormat.Xml, UriTemplate = "SomeRequest?qs1={qs1}&qs2={qs2}")]
XElement SomeRequest1(string qs1, string qs2);
Now if we had to make a contract that accepts an array of parameters with the same name (in this case qs1) contract like so...
[OperationContract]
[WebGet(ResponseFormat = WebMessageFormat.Xml, UriTemplate = "SomeRequest?qs1={qs1}&qs1={qs2}")]
XElement SomeRequest2(string qs1, string qs2);
We get the error message at run time when we make the invocation to the method:
the query string must have 'name=value' pairs with unique names. Note that the names are case-insensitive. See the documentation for UriTemplate for more details.
How does one define an HTTP service that exposes a resource with an array of parameters without resorting to a loosey-goosey interface?
I've implemented a simple custom QueryStringConverter so that you can make qs1 an string[] then have the query string variable be comma delimited (e.g. http://server/service/SomeRequest?qs1=val1,val2,val3,val4)
[OperationContract]
[WebGet(ResponseFormat = WebMessageFormat.Xml,
UriTemplate = "SomeRequest?qs1={qs1}")]
XElement SomeRequest2(string[] qs1);
First you need a class that inherits from WebHttpBehavior so that we can inject our custom QueryStringConverter:
public class CustomHttpBehavior : System.ServiceModel.Description.WebHttpBehavior
{
protected override System.ServiceModel.Dispatcher.QueryStringConverter GetQueryStringConverter(System.ServiceModel.Description.OperationDescription operationDescription)
{
return new CustomQueryStringConverter();
}
}
Then our CustomQueryStringConverter that handles string[] parameters:
public class CustomQueryStringConverter : System.ServiceModel.Dispatcher.QueryStringConverter
{
public override bool CanConvert(Type type)
{
if (type == typeof(string[]))
{
return true;
}
return base.CanConvert(type);
}
public override object ConvertStringToValue(string parameter, Type parameterType)
{
if (parameterType == typeof(string[]))
{
string[] parms = parameter.Split(',');
return parms;
}
return base.ConvertStringToValue(parameter, parameterType);
}
public override string ConvertValueToString(object parameter, Type parameterType)
{
if (parameterType == typeof(string[]))
{
string valstring = string.Join(",", parameter as string[]);
return valstring;
}
return base.ConvertValueToString(parameter, parameterType);
}
}
The last thing you need to do is create a behavior configuration extension so that the runtime can get an instance of the CustomWebHttpBehavior:
public class CustomHttpBehaviorExtensionElement : System.ServiceModel.Configuration.BehaviorExtensionElement
{
protected override object CreateBehavior()
{
return new CustomHttpBehavior();
}
public override Type BehaviorType
{
get { return typeof(CustomHttpBehavior); }
}
}
Now we add the element to our configuration extensions so that our CustomWebHttpBehavior is used, we use the Name of that extension instead of <webHttp /> in our behavior:
<system.serviceModel>
<services>
<service name="NameSpace.ServiceClass">
<endpoint address="" behaviorConfiguration="MyServiceBehavior"
binding="webHttpBinding" contract="NameSpace.ServiceClass" />
</service>
</services>
<behaviors>
<endpointBehaviors>
<behavior name="MyServiceBehavior">
<customWebHttp/>
</behavior>
</endpointBehaviors>
</behaviors>
<extensions>
<behaviorExtensions>
<add name="customWebHttp" type="NameSpace.CustomHttpBehaviorExtensionElement, MyAssemblyName" />
</behaviorExtensions>
</extensions>
<serviceHostingEnvironment aspNetCompatibilityEnabled="true" />
</system.serviceModel>
You can now also extend your CustomQueryStringConverter to handle other types that the default one doesn't, such as nullable value types.
NOTE: There is a bug logged at microsoft connect that directly relates to this code. The code does not actually work in almost all circumstances where you attempt to Query Convert different types.
http://connect.microsoft.com/VisualStudio/feedback/details/616486/bug-with-getquerystringconverter-not-being-called-by-webservicehost#tabs
Please make sure you read this carefully before wasting hours of your time creating custom REST query string converters that cannot possibly work. (Applies to Framework 4.0 and below).
To respond to your comment on my other answer:
You can do a wildcard parameter at the end of the querystring like
[WebGet(ResponseFormat = WebMessageFormat.Xml,
UriTemplate = "SomeRequest?qs1={*qs1}")]
XElement SomeRequest2(string qs1);
This way the qs1 string parameter will be the whole raw querystring after the qs1=, you could then parse that manually in your code.
The QueryStringConverter relies on the formatting of the querystring so doing something exactly how you want is not possible without possibly rewriting QueryStringConverter instead of the little overrides we did in the other answer.
From MSDN:
Wildcard segments must follow the following rules:
There can be at most one named wildcard segment for each template string.
A named wildcard segment must appear at the right-most segment in the path.
A named wildcard segment cannot coexist with an anonymous wildcard segment within the same template string.
The name of a named wildcard segment must be unique.
Named wildcard segments cannot have default values.
Named wildcard segments cannot end with “/”.
Be aware that in WCF 3.5 you must specify the full qualified assembly name in:
<extensions>
<behaviorExtensions>
<add name="customWebHttp" type="NameSpace.CustomHttpBehaviorExtensionElement, MyAssemblyName, NOT SUFFICIENT HERE" />
</behaviorExtensions>
</extensions>
Just like this: SampleService.CustomBehavior, SampleService, Version=1.0.0.0, Culture=neutral, PublicKeyToken=null
Otherwise you will get exception:
Configuration Error
Description: An error occurred during the processing of a configuration file required to service this request. Please review the specific error details below and modify your configuration file appropriately.
Parser Error Message: Invalid element in configuration. The extension name 'CustomWebHttp' is not registered in the collection at system.serviceModel/extensions/behaviorExtensions.