Projects tagged ‘ipc’

D-Bus is a message bus system, a simple way for applications to talk to one another. In addition to interprocess communication, D-Bus helps coordinate process lifecycle; it makes it simple and ... [More] reliable to code a "single instance" application or daemon, and to launch applications and daemons on demand when their services are needed. [Less]

The HelenOS project is an effort to develop a complete and usable modern operating system, yet offering room for experimenting and research. HelenOS uses its own microkernel written from scratch and ... [More] supports SMP, multitasking and multithreading on both 32-bit and 64-bit, little-endian and big-endian processor architectures, among which are AMD64/EM64T (x86-64), ARM, IA-32, IA-64 (Itanium), 32-bit MIPS, 32-bit PowerPC, SPARC V9 and Xen 3.0. Thanks to the relatively high number of supported architectures and suitable design, HelenOS is extremely-well portable. [Less]

YARP: Yet Another Robot Platform. Libraries and applications to support flexible communication between processes and devices spread across a local network. Especially useful for inhomogeneous ... [More] networks with a mixture of operating systems and compilers. Originally developed for large-scale humanoid robot projects such as COG, Kismet, and RobotCub, where many idiosyncratic devices need to be integrated into a single control system without there being any one OS that supports them all. Uses well-specified protocols across tcp, udp, shared-memory, and multicast, with binary and text-mode variants. [Less]

dbus-cxx is yet another C++ wrapper for dbus. However, it differs from the current implementations in several respects. First, it exposes the C API to allow direct manipulation. Second, it relies heavily on sigc++ to provide an OO interface.

L4Ka::Pistachio is the latest L4 microkernel developed by the System Architecture Group at the University of Karlsruhe in collaboration with the DiSy group at the University of New South Wales ... [More] , Australia. It is the first available kernel implementation of the L4 Version 4 kernel API (currently code-named Version X.2), which is fully 32 and 64 bit clean, provides multiprocessor support, and super-fast local IPC. [Less]

Parabellym is an environment for building modular applications of any kind. It consists of a shared library, which provides module management and message based inter-module communication functions, and a tiny loader.

Fiasco is a new µ-kernel (microkernel) running on x86 PCs intended to be compatible with the L4 µ-kernel for x86. It implements the L4 ABI as defined in the L4/x86 Reference Manual. The ... [More] Fiasco kernel can be used to construct flexible systems. We at TU Dresden use it as a base for our DROPS system which supports running real-time and time-sharing applications concurrently on one computer. However, Fiasco is not only suitable for big and complex systems, but also for small, embedded applications. [Less]

OverviewKryoNet is a Java library that provides a clean and simple API for efficient TCP and UDP client/server network communication using NIO. KryoNet uses the Kryo serialization library to ... [More] automatically and efficiently transfer object graphs across the network. KryoNet runs on both the desktop and on Android. KryoNet is ideal for any client/server application. It is very efficient, so is especially good for games. KryoNet can also be useful for inter-process communication. Running a serverThis code starts a server on TCP port 54555 and UDP port 54777: Server server = new Server(); server.start(); server.bind(54555, 54777);The start method starts a thread to handle incoming connections, reading/writing to the socket, and notifying listeners. This code adds a listener to handle receiving objects: server.addListener(new Listener() { public void received (Connection connection, Object object) { if (object instanceof SomeRequest) { SomeRequest request = (SomeRequest)object; System.out.println(request.text); SomeResponse response = new SomeResponse(); response.text = "Thanks!"; connection.sendTCP(response); } } });Note the Listener class has other notification methods that can be overridden. Typically a listener has a series of instanceof checks to decide what to do with the object received. In this example, it prints out a string and sends a response over TCP. The SomeRequest and SomeResponse classes are defined like this: public class SomeRequest { public String text; }public class SomeResponse { public String text; }Kryo automatically serializes the objects to and from bytes. Connecting a clientThis code connects to a server running on TCP port 54555 and UDP port 54777: Client client = new Client(); client.start(); client.connect(5000, "192.168.0.4", 54555, 54777); SomeRequest request = new SomeRequest(); request.text = "Here is the request!"; client.sendTCP(request);The start method starts a thread to handle the outgoing connection, reading/writing to the socket, and notifying listeners. Note that the thread must be started before connect is called, else the outgoing connection will fail. In this example, the connect method blocks for a maximum of 5000 milliseconds. If it times out or connecting otherwise fails, an exception is thrown (handling not shown). After the connection is made, the example sends a "SomeRequest" object to the server over TCP. This code adds a listener to print out the response: client.addListener(new Listener() { public void received (Connection connection, Object object) { if (object instanceof SomeResponse) { SomeResponse response = (SomeResponse)object; System.out.println(response.text); } } });Registering classesIn order for the above examples to work, the classes that are going to be sent over the network must be registered with the following code: Kryo kryo = server.getKryo(); kryo.register(SomeRequest.class); kryo.register(SomeResponse.class);Kryo kryo = client.getKryo(); kryo.register(SomeRequest.class); kryo.register(SomeResponse.class);This must be done on both the client and server, before any network communication occurs. It is very important that the exact same classes are registered on both the client and server, and that they are registered in the exact same order. Because of this, typically the code that registers classes is placed in a method on a class available to both the client and server. Please see the Kryo serialization library for more information on how objects are serialized for network transfer. Kryo can serialize any object and supports data compression (eg, deflate or delta compression). Remote Method InvocationKryoNet has an easy to use mechanism for invoking methods on remote objects (RMI). This is done by creating an ObjectSpace and registering objects with an ID: ObjectSpace objectSpace = new ObjectSpace(); objectSpace.register(42, someObject); // ... objectSpace.addConnection(connection);Multiple ObjectSpaces can be created for both the client or server side. Once registered, objects can be used on the other side of the registered connections: SomeObject someObject = ObjectSpace.getRemoteObject(connection, 42, SomeObject.class); SomeResult result = someObject.doSomething();The getRemoteObject method returns a proxy object that represents the specified class. When a method on the class is called, a message is sent over the connection and on the remote side the method is invoked on the registered object. The method blocks until the return value is sent back over the connection. Exactly how the remote method invocation is performed can be customized by casting the proxy object to a RemoteObject. SomeObject someObject = ObjectSpace.getRemoteObject(connection, 42, SomeObject.class); ((RemoteObject)someObject).setNonBlocking(true, true); someObject.doSomething();Note that the SomeObject class does not need to implement RemoteObject, this is handled automatically. The first true passed to setNonBlocking causes remote method invocations to be non-blocking. When doSomething is invoked, it will not block and wait for the return value. Instead the method will just return null. The second true passed to setNonBlocking indicates that the return value of remote method invocations are to be ignored. This means the server will not waste time or bandwidth sending the result of the remote method invocation. If the second parameter for setNonBlocking is false, the server will send back the remote method invocation return value. There are two ways to access a return value for a non-blocking method invocation: RemoteObject remoteObject = (RemoteObject)someObject; remoteObject.setNonBlocking(true, false); someObject.doSomething(); // ... SomeResult result = remoteObject.waitForLastResponse();RemoteObject remoteObject = (RemoteObject)someObject; remoteObject.setNonBlocking(true, false); someObject.doSomething(); byte responseID = remoteObject.getLastResponseID(); // ... SomeResult result = remoteObject.waitForResponse(responseID);TCP and UDPKryoNet always uses a TCP port. This allows the framework to easily perform reliable communication and have a stateful connection. KryoNet can optionally use a UDP port in addition to the TCP port. While both ports can be used simultaneously, it is not recommended to send an huge amount of data on both at the same time because the two protocols can affect each other. TCP is reliable, meaning objects sent are sure to arrive at their destination eventually. UDP is faster but unreliable, meaning an object sent may never be delivered. Because it is faster, UDP is typically used when many updates are being sent and it doesn't matter if an update is missed. Note that KryoNet does not currently implement any extra features for UDP, such as reliability or flow control. It is left to the developer to make proper use of the UDP connection. ThreadingKryoNet imposes no restrictions on how threading is handled. The Server and Client classes have an update method that accepts connections and reads or writes any pending data for the current connections. The update method should be called periodically to process network events. Both the Client and Server classes implement Runnable and the run method continually calls update until the stop method is called. Handing a client or server to a java.lang.Thread is a convenient way to have a dedicated update thread, and this is what the start method does. If this doesn't fit your needs, call update manually from the thread of your choice. Listeners are notified from the update thread, so should not block for long. Static wrapper classes are provided on the Listener class to change how a listener is notified, such as ThreadedListener. The update thread should never be blocked to wait for an incoming network message, as this will cause a deadlock. LAN server discoveryKryoNet can broadcast a UDP message on the LAN to discover any servers running: InetAddress address = client.discoverHost(54777, 5000); System.out.println(address);This will print the address of the first server found running on UDP port 54777. The call will block for up to 5000 milliseconds, waiting for a response. LoggingKryoNet makes use of the low overhead, lightweight MinLog logging library. The logging level can be set in this way: Log.set(LEVEL_TRACE);KryoNet does minimal logging at INFO and above levels. DEBUG is good to use during development and indicates the total number of bytes for each object sent. TRACE is good to use when debugging a specific problem, but outputs too much information to leave on all the time. MinLog supports a fixed logging level, which will remove logging statements below that level. For efficiency, KryoNet can be compiled with a fixed logging level MinLog JAR. See MinLog for more information. KryoNet versus ?KryoNet makes the assumptions that it will only be used for client/server architectures and that KryoNet will be used on both sides of the network. Because KryoNet solves a specific problem, the KryoNet API can do so very elegantly. The Apache MINA project is similar to KryoNet. MINA's API is lower level and a great deal more complicated. Even the simplest client/server will require a lot more code to be written. MINA also is not integrated with a robust serialization framework and doesn't intrinsically support RMI. The Priobit project is a minimal layer over NIO. It provides TCP networking similar to KryoNet, but without the higher level features. Priobit requires all network communication to occur on a single thread. The Java Game Networking project is a higher level library similar to KryoNet. JGN does not have as simple of an API. [Less]

1. Use linux, apache and process(prefer c) implements RESTful web service(prefer JSON data structure) 2. Use xhtml, css and javascript implement client side pages.

Create shared variables under perl. Allows the vars to be complex objects (hash of arrays, etc.). Think of it as a combination of "use vars", "Tie::DBI" and "Storable". Works on all platforms. ... [More] Simple syntax: use IPC::Lite qw(%hash); $hash{'foo'} = 'bar'; # this is shared between processes. Timeout syntax: use IPC::Lite Timeout=>'3 days', qw(%hash); $hash{'foo'} = 'bar'; # expires, if unused, after 3 days [Less]