A TCP communication library based on middleware components. Read the official docs to learn more.
The following are projects that extend the functionality of this library.
The concept behind this library is to apply a middleware pipeline to data coming in and out from the open communication channels.
For data coming through the channel input, two middleware components can be applied: adapters and handlers.
graph LR;
channelInput((Input)) --> a1[/Adapter/]
subgraph adapters
a1 --> a2[/Adapter/]
end
a2 --> h1[/Handler/]
a2 --> h2[/Handler/]
For data going through the channel output, only adapters are applicable. Whatever comes out from the pipeline is delivered to an internal handler that writes the data to the channel's underlying socket.
graph LR;
openChannel((Channel)) --> a1[/Adapter/]
subgraph adapters
a1 --> a2[/Adapter/]
end
a2 --> channelOutput([Output])
Unless you have very specific needs, middleware components should inherit from the abstract classes provided instead of implementing the interfaces directly. The base class for all middleware components (adapters and handlers) does a few things for us that won't be available when implementing the interfaces directly. This includes
-
Type Checking - Ensures the data type is suitable for the middleware component. If it's not, the middleware is not executed. If the middleware is an adapter, the data is automatically forwarded to the next middleware in the pipeline. This behaviour can be changed by overriding the
OnDataNotSuitable
method. -
Type Mutation - The capacity to convert the data type, when compatible with the expected middleware data type. All middleware components already deal with
IByteBuffer
<-->Byte[]
andT
<-->IEnumerable<T>
mutations, but they also provide an opportunity to change/extend this behaviour by overriding theConvertType
method.
An adapter is a middleware component that can be executed at any point in the pipeline and it has a single conceptual purpose: to adapt data.
graph LR;
in[/Data In/] --> Adapter --> out[/Data Out/]
An adapter is expected to forward data to next component in the pipeline, although that is not always the case. If an adapter doesn't forward any data, the pipeline is interrupted.
Unless you have very specific needs, you should inherit your adapter from the ChannelAdapter<T>
abstract class instead of implementing the IChannelAdapter
interface directly.
We also need to indicate whether the adapter is meant for the input or/and the output pipelines. We do that by adding the interfaces IInputChannelAdapter
or/and IOutputChannelAdapter
respectively.
Here's an example of how to implement an adapter that adapts from an IByteBuffer
(or Byte[]
). This adapter can only be added to the input pipeline, since it only implements the IInputChannelAdapter
interface.
public class MyChannelAdapter : ChannelAdapter<IByteBuffer>, IInputChannelAdapter
{
public override Task ExecuteAsync( IAdapterContext context, IByteBuffer data )
{
// adapt/transform data
var adaptedData = ...
// forward adapted data
context.Forward( adaptedData );
}
}
In addition to the abstract ChannelAdapter<T>
adapter, you have a few ready-made adapters that you can use.
Adapter | Target | Description |
---|---|---|
AnonymousChannelAdapter | Input/Output | A quick way to implement an anonymous adapter |
BufferLengthAdapter | Input | Ensures the input buffer doesn't exceed in length |
Although handlers are very similar to adapters, their conceptual purpose is different: to handle data. That means that business logic should be applied here and not on an adapter. Handlers are executed at the end of the pipeline and as such, they don't forward data. Unlike adapters, if more than one handler exists for a given data type, all are executed.
graph LR;
in[/Data In/] --> H1[Handler]
in ----> H2[Handler]
Similarly to the adapters, unless you have very specific needs, you should inherit your handler from the ChannelHandler<T>
class and not implementing the IChannelHandler
interface directly.
public class MyChannelHandler : ChannelHandler<MyData>
{
public override Task ExecuteAsync( IChannelContext context, MyData data )
{
// implement your handler here
}
}
Because adapters and handlers are so similar, there might be a temptation to do everything with adapters. And while that's feasable, it's not recommended. Adapters should be used to adapt data and handlers to handle data (business logic).
graph LR;
channelInput((Input)) --> a1[/Adapter/]
a1 --> a2[/Adapter/]
a2 --> h1[/Handler/]
a2 --> h2[/Handler/]
Adapters | Handlers |
---|---|
Adapt and forward data | Handle data and business logic |
Run at any point in the pipeline | Run at the end of the pipeline |
Single adapter for forwarded data type | Multiple handlers for forwarded data type |
Before version 0.10, the base middleware split the execution of IEnumerable<T>
<--> T
spreads into multiple tasks. The reason behind this was to improve the speed of execution, however, this also meant that the order of execution was not guaranteed.
Starting from version 0.10, the base middleware executes IEnumerable<T>
<--> T
spreads sequentially, guaranteeing the order of execution. This decision was made because on most cases, the performance gain was negligible and the task spawning caused some confusion to implementors.
At any point, within an adapter or handler, we can write data to the channel output; this will trigger the output pipeline and at the end of it, send the data to the other party. However, there are two distinct ways of doing this, both with a distinct behaviour.
The middleware context gives us access to an output buffer that we can write to. This IS the recommended method. Writing to the output buffer doesn't immediately trigger the output pipeline. Instead, it is only triggered at the end of the (input) pipeline, after all adapters and handlers have executed (fully). If the pipeline is interrupted, because an adapter didn't forward any data or a handler crashed, the data in the buffer will be discarded and never written to the channel.
public override async Task ExecuteAsync( IAdapterContext context, IEnumerable<Message> data )
{
// ...
context.Output.Write( replyData );
}
This is the most straightforward method and it will immediately trigger the output pipeline, however, it is NOT the recommended way, unless you need the data to reach the other party, no matter what happens next (current or next middleware component). This is an asynchronous process.
public override async Task ExecuteAsync( IAdapterContext context, IEnumerable<Message> data )
{
// ...
await context.Channel.WriteAsync( replyData );
}
Install the package from NuGet
dotnet add package Faactory.Channels
The first step is to add the library to the DI container and configure the channel pipelines. These configurations are always named, which means that we can have multiple channel configurations for different purposes. Nonetheless, if we only need one channel pipeline, we can do it all at once by configuring a *default` configuration.
IServiceCollection services = ...;
// add our hosted service
services.AddChannels( channel =>
{
// set up input pipeline
channel.AddInputAdapter<ExampleAdapter>()
.AddInputHandler<ExampleHandler>();
// set up output pipeline
channel.AddOutputAdapter<ExampleAdapter>();
} );
If we need to configure multiple channel pipelines, we use the parameterless method, which returns a builder that allows us to configure named channels.
IServiceCollection services = ...;
services.AddChannels()
.Add( "channel1", channel =>
{
// set up input pipeline
channel.AddInputAdapter<ExampleAdapter>()
.AddInputHandler<ExampleHandler1>();
// set up output pipeline
channel.AddOutputAdapter<ExampleAdapter>();
} )
.Add( "channel2", channel =>
{
// set up input pipeline
channel.AddInputAdapter<ExampleAdapter>()
.AddInputHandler<ExampleHandler2>();
// set up output pipeline
channel.AddOutputAdapter<ExampleAdapter>();
} );
After the channels are configured, we need to add the listener services. The library provides listeners for TCP and UDP channels. When adding a listener, we can specify the channel name and the options, or we can use the default channel configuration.
IServiceCollection services = ...;
services.AddTcpChannelListener( 8080 ); // TCP listener with default channel configuration
// services.AddTcpChannelListener( "channel1", 8080 ); // TCP listener with named channel configuration
// services.AddUdpChannelListener( 7701 ); // UDP listener with default channel configuration
We can use multiple listeners in the same application, each with its own configuration.
Although raw data handling in the adapters can be done with Byte[]
, it is recommended to use a IByteBuffer
instance instead, particularly for reading data. You can read more about it here.
Data received in the adapters that is not read will remain in the channel's input buffer. When more data is received, it is delivered again along with the newly received data. If an adapter uses Byte[]
instead, the data in the input buffer is automatically marked as read and discarded.
Every channel instance (client or service) uses its own IServiceScope
. This means that if you add a scoped service to the DI container and use it in an adapter or handler, you'll have an unique instance per channel.
In some cases, you might need to monitor channel events. This can be useful for logging, statistics or any other scenario where this information is needed. The following events are available
- Channel Created
- Channel Closed
- Data Received
- Data Sent
To receive channel events, you'll need to create a class that implements IChannelMonitor
interface and then add it to the DI container. You can add multiple implementations and whether they are transient, scoped or singleton depends entirely on your needs.
public class MyChannelMonitor : IChannelMonitor
{
// ...
}
// ...
IServiceCollection services = ...;
services.AddSingleton<IChannelMonitor, MyChannelMonitor>();
A channel service is a background service that is executed when a channel is created and stopped when it closes, sharing the same lifetime and scope as the channel. This is useful for long-running services that need to be executed within the channel scope.
The easiest way to create a channel service is to inherit from the ChannelService
abstract class and override the ExecuteAsync
method.
public class MyService : ChannelService
{
protected override async Task ExecuteAsync( CancellationToken cancellationToken )
{
while ( !cancellationToken.IsCancellationRequested )
{
// insert code...
/*
here we have access to the channel instance through the Channel property
*/
// await Channel.WriteAsync( ... );
await Task.Delay( 1000 );
}
}
}
If you have other specific needs, you can also implement the IChannelService
interface directly.
public class MyService : IChannelService
{
// ...
public Task StartAsync( IChannel channel, CancellationToken cancellationToken )
{
// Invoked when a channel is created
}
public Task StopAsync( CancellationToken cancellationToken )
{
// Invoked when a channel is closed
}
public void Dispose()
{ }
}
The service is added by using the builder's AddChannelService
method.
IChannelBuilder channel = ...;
chanel.AddChannelService<MyService>();
It is possible to store custom data on a channel instance. The IChannel
interface exposes a Data
property, which is essentially a case-insensitive string dictionary. This can be useful for storing data that is used later by other adapters and handlers.
public class SampleIdentityHandler : ChannelHandler<IdentityInformation>
{
public override Task ExecuteAsync( IChannelContext context, IdentityInformation data )
{
if ( !IsAuthorized( data ) )
{
return context.Channel.CloseAsync();
}
/*
store the UUID on the channel data for later use
*/
context.Channel.Data["uuid"] = data.UUId;
return Task.CompletedTask;
}
}
By default, channels are initialized with an idle detection mechanism that closes the channel if no data is received or sent after a certain amount of time, which defaults to 60 seconds. This mechanism can be disabled or customized through the channel options.
IServiceCollection services = ...;
services.AddChannels( channel =>
{
channel.Configure( options =>
{
// this is the default setting; added here just for clarity
options.IdleTimeout = TimeSpan.FromSeconds( 60 );
} );
// ...
} );
To disable the idle detection mechanism, set the IdleTimeout
property to TimeSpan.Zero
.
Tip
The idle detection mechanism is available for all channel types: TCP, UDP and WebSockets.
The library also provides a TCP/UDP client that can be used to connect to a server. This client automatically connects to the server and creates a channel instance when the connection is established. Connection drops are automatically handled and the client will attempt to reconnect.
Clients use the same channel configuration as the listeners, but they require additional configuration.
IServiceCollection services = ...;
/*
this registers a default client with the default channel configuration
*/
services.AddChannelsClient( "tcp://example.host:8080" );
/*
we could also register the default client with a named channel configuration
*/
// services.AddChannelsClient( "channel1", "tcp://example.host:8080" );
/*
when we need to create a client, we only need to inject the `IChannelsClientFactory` interface
*/
public class MyClient
{
private readonly IChannelsClient client;
public MyClient( IChannelsClientFactory factory )
{
client = factory.CreateClient();
}
public Task ExecuteAsync()
{
// ...
}
}
If we need to configure multiple clients with different channel configurations, we need to register them as named clients instead.
IServiceCollection services = ...;
/*
this registers a named client (client1) with the default channel configuration
*/
services.AddChannelsNamedClient( "client1", "tcp://example.host:8080" );
/*
this registers a named client (client2) with a named channel configuration (channel1)
*/
services.AddChannelsNamedClient( "client2", "channel1", "tcp://example.host:8080" );
Support for web sockets is available through the Faactory.Channels.WebSockets
package. It provides ASP.NET Core routing and middleware for easy integration with this library. Read more about it here.