I'm working on a project, which I am not at liberty to discuss the core, but I have reached a stumbling block. I need data to be transferred from C++ to some other language, preferably Java or Python, in realtime (~10ms latency).
We have a sensor that HAS to be parsed in C++. We are planning on doing a data read/output through bluetooth, most likely Java or C# (I don't quite know C#, but it seems similar to Java). C++ will not fit the bill, since I do not feel advanced enough to use it for what we need. The sensor parsing is already finished. The data transferring will be happening on the same machine.
Here are the methods I've pondered:
We tried using MatLab with whatever the Mex stuff is (I don't do MatLab) to access functions from our C++ program, to retrieve the data as an array. Matlab will be too slow (we read somewhere that the TX/RX will be limited to 1-20 Hz.)
Writing the data to a text, or other equivalent raw data, file constantly, and opening it with the other language as necessary.
I attempted to look this up, but nothing of use showed in the results.
It seems like you are looking for a IPC (Inter-process communication). The easiest ones to implement are the socket and Pipes. I have added links which will help you decide which one to use. I have implemented named pipes and socket to transfer data at every 1ms. I did these implementations in python, c and pascal.
Here is the descriptions of all the IPC types
What's the difference between pipes and sockets
We had same issue where we had to share sensor data between one Java app to other multiple apps including Java,Python and R.
First we tried Socket connections but socket communication were not fault tolerant. Restarting or failure in one app affected other.
Then we tried RMI calls between them but again we were unhappy due to scalability.
We wanted system to be reliable, scalable, distributed and fault tolerant. So, finally we started using RabbitMQ where we created one producer and multiple consumers. It worked well for 2 years. you may consider using Apache Kafka.
You have options like Socket pipes, RMI calls, RabbitMQ, Kafka, Redis based on your system requirements now and in near future.
Related
I have the following situation:
I have 2 JVM processes (really 2 java processes running separately, not 2 threads) running on a local machine. Let's call them ProcessA an ProcessB.
I want them to communicate (exchange data) with one another (e.g. ProcessA sends a message to ProcessB to do something).
Now, I work around this issue by writing a temporary file and these process periodically scan this file to get message. I think this solution is not so good.
What would be a better alternative to achieve what I want?
Multiple options for IPC:
Socket-Based (Bare-Bones) Networking
not necessarily hard, but:
might be verbose for not much,
might offer more surface for bugs, as you write more code.
you could rely on existing frameworks, like Netty
RMI
Technically, that's also network communication, but that's transparent for you.
Fully-fledged Message Passing Architectures
usually built on either RMI or network communications as well, but with support for complicated conversations and workflows
might be too heavy-weight for something simple
frameworks like ActiveMQ or JBoss Messaging
Java Management Extensions (JMX)
more meant for JVM management and monitoring, but could help to implement what you want if you mostly want to have one process query another for data, or send it some request for an action, if they aren't too complex
also works over RMI (amongst other possible protocols)
not so simple to wrap your head around at first, but actually rather simple to use
File-sharing / File-locking
that's what you're doing right now
it's doable, but comes with a lot of problems to handle
Signals
You can simply send signals to your other project
However, it's fairly limited and requires you to implement a translation layer (it is doable, though, but a rather crazy idea to toy with than anything serious.
Without more details, a bare-bone network-based IPC approach seems the best, as it's the:
most extensible (in terms of adding new features and workflows to your
most lightweight (in terms of memory footprint for your app)
most simple (in terms of design)
most educative (in terms of learning how to implement IPC). (as you mentioned "socket is hard" in a comment, and it really is not and should be something you work on)
That being said, based on your example (simply requesting the other process to do an action), JMX could also be good enough for you.
I've added a library on github called Mappedbus (http://github.com/caplogic/mappedbus) which enable two (or many more) Java processes/JVMs to communicate by exchanging messages. The library uses a memory mapped file and makes use of fetch-and-add and volatile read/writes to synchronize the different readers and writers. I've measured the throughput between two processes using this library to 40 million messages/s with an average latency of 25 ns for reading/writing a single message.
What you are looking for is inter-process communication. Java provides a simple IPC framework in the form of Java RMI API. There are several other mechanisms for inter-process communication such as pipes, sockets, message queues (these are all concepts, obviously, so there are frameworks that implement these).
I think in your case Java RMI or a simple custom socket implementation should suffice.
Sockets with DataInput(Output)Stream, to send java objects back and forth. This is easier than using disk file, and much easier than Netty.
I tend to use jGroup to form local clusters between processes. It works for nodes (aka processes) on the same machine, within the same JVM or even across different servers.
Once you understand the basics it is easy working with it and having the options to actually run two or more processes in the same JVM makes it easy to test those processes easily.
The overhead and latency is minimal if both are on the same machine (usually only a TCP rountrip of about >100ns per action).
socket may be a better choice, I think.
Back in 2004 I implement code which do the job with sockets. Until then, many times I search for a better solution, because socket approach triggers firewall and my clients worry. There is no better solution until now. Client must serialize your data, send and server must receive and unserialize.
It is easy.
I am working in a desktop application, where application is deployed in both windows and mac platforms. As part of the application, it should communicate with native layer. Currently the communication between native layer and Java layer is done using sockets. Recently some one in the team suggested to use zeroMQ. Can any one of you guys please clarify my doubts.
How zeroMQ better than sockets
Is it possible to install zeroMQ library as part the Desktop client installation
I gone through the link 'https://github.com/zeromq/clrzmq4', it given libraries specific to amd64 and i386 processor family. Do I need to build it separately from the source code for different processors?
Do I still require .dll files to use zeroMQ in Java?
Do I require Visual studio to build zeroMQ libraries in windows (Since my native layer written in C#, my C# application communicate with zeroMQ socket socket written java)?
How zeroMQ better than sockets
http://zeromq.org/topics:omq-is-just-sockets
Is it possible to install zeroMQ library as part the Desktop client installation?
Yes, you need to build the libraries depends on the processor and embed them in your application.
Do I need to build it separately from the source code for different processors?
Yes, you need to build the libraries from source. zeroMQ is processor centric.
Do I still require .dll files to use zeroMQ in Java?
Yes, Following link may help you
Exception in thread "main" java.lang.UnsatisfiedLinkError: ... \jzmq.dll: Can't find dependent libraries
Do I require Visual studio to build zeroMQ libraries in windows?
Yes
This link may help you to get basic examples.
Regarding ZeroMQ in a desktop application on Windows talking to another process on the same machine, bear in mind that zmq_ipc is not supported (see zmq_ipc(7)). Or at least, that's the last I heard. This is because it's fundamentally impossible to implement anything like quite like select() or epoll() for named pipes in Windows. Just use zmq_tcp instead.
The same basic problem plagued the development of the select() implementation in Cygwin and its derivatives. They got round the problem by starting a thread for every non-socket file descriptor (i.e. named pipes, serial ports, etc) being selected, with each thread polling the HANDLE for whether any data had arrived (or whatever events were being set in select()). Not very efficient. Yeurk.
Proactor vs Reactor
Windows is proactor (which can do only proactor), everything else (*nix, VxWorks) is reactor (which can also be used to implement a proactor). The development of the boost.asio library for C++ was influenced by this, and is a proactor design as a result so that it could be run on Windows. RabbitMQ is proactor too.
ZeroMQ with zmq_poll() is reactor.
Proactor - you pro-actively start up an asynchronous routine to handle whatever turns up in the future.
Reactor - you react to the arrival of whatever has turned up by starting a synchronous call to whatever routine you wish to handle it knowing that it will complete very quickly because the data is already there.
The difference is key. In a proactor design, once you have started up that asynchronous routine to read a message, you cannot (easily) stop it or change it until it has done its thing. That is very annoying if you change your mind, for example as a result of reading some message from somewhere else.
Small caveat - Windows does support select() for network sockets (thus reactor programming is possible with network sockets, but only network sockets), and is the only reason why ZMQ is supported to any extent whatsoever on Windows.
Mixing ZMQ with Desktop Application Event Callbacks
Anyway, proactor means that Windows and C# fundamentally expects everything to be served by callbacks. This basically means you won't be using the zmq_poll() call to tell you if new messages have arrived if you also have callbacks handling GUI events. Instead you'd most likely be making asynchronous calls to zmq_revcmsg(). Trying to mix zmq_poll() in with callbacks is asking for trouble (you'd be blending proactor and reactor).
Message Formats
ZeroMQ and sockets both transfer bytes (as discrete messages with ZeroMQ, as a byte stream with sockets). One still has the challenge of deciding what the bytes mean to applications.
I can recommend using something like Google Protocol Buffers to serialise messages for transport by ZeroMQ. It's available for both C# and Java, but it doesn't demarcate message boundaries. Fortunately, ZeroMQ does. [Using GPB over a socket stream can be painful, you have to demarcate the message boundaries oneself]. So you can serialise a message to a buffer, hand the buffer over to ZeroMQ as a message, the recipient receives the message and knows for absolute certain that there is one single solitary GPB within. If you like you can use GPB's "oneof" to smuggle arbitrary message types across, which can be very liberating. You can accomplish the same with other serialisation technologies too of course, my personal favourite being ASN.1.
I am currently decide what kind of communication method/network protocol I am going to use for a new project.
What I can tell you about this project is that:
- It is Android/java based, using X amount of Android devices
- These devices should be able to send strings to each other over a local network. We are talking about small strings here. Small as in less than 100 characters.
- The amount of packages/transmissions being sent can vary "A LOT". I can't say how much unfortunately, but the network protocol needs to be as scalable as possible.
I have researched different kinds of possible solutions and is now deciding wether to use "Sockets" or "RMI"
As I have understood about RMI:
It is easier than Java sockets to implement and maintain (smaller amount of code)
It is "a bit slower" than sockets, as it is a new "layer" build on top of Sockets
There may be some scalability issues (if this is true, how "serious" is it?) as it creates a lot of new sockets, resulting in Exceptions.
Obviously the system needs to run as smooth as possible, but the main objective is to make it scalable so it can handle more Android devices.
EDIT: The system the system is not "peer-to-peer". All of the android devices should be able to be configured as the server.
None of your concerns are the real issue, in my view.
RMI has a pre-defined protocol, raw sockets do not.
If you use raw sockets, you have to do all the work to define what messages and protocols are exchanged by client and server.
There are so many good existing protocols (RMI, HTTP, etc.) that I'd wonder why you feel the need to invent your own again.
Android devices communicating over HTTP - tell me why it won't be fast or scalable enough. HTTP is good enough for the Internet - why not you and your solution?
I would suggest you to expose some kind of webservice (SOAP or REST) in your application server. For example, people frequently expose their data to mobile devices as a REST webservice API returning some kind of JSON format in order to make it easier to marshal it again in the client device.
This way you take profit of the underlying implementation of HTTP communication in every application server; any other way, you would have to write your own worker thread pool using nio primitive operations in order to achieve performance... Not a thing to be done in a real production environment - maybe in an academic one?
I'd like to somehow get control over active sockets on my computer with java/scala. For example, if a program has stablished a tcp connection, i want to be able to retrieve and listen/write onto this socket.
So, the first question is: is that possible?
And the second: how?
You can do this by writing a custom Socket factory. This is fairly complicated and I don't suggest you do this unless you are trying a to hack a program you have no control over (again a very bad idea)
Instead I suggest you monitor your own usage. When you read/write from a Socket you also keep any information you need. i.e Its your program so change it to do what you need.
I found a java wrapper of lipcap http://jnetpcap.com/ that let me do the sort of things i was after. From the web:
jNetPcap is an open-source java library. It contains:
A Java wrapper for nearly all libpcap library native calls
Decodes captured packets in real-time
Provides a large library of network protocols (core protocols)
Users can easily add their own protocol definitions using java SDK
jNetPcap uses a mixture of native and java implementation for optimum packet decoding performance
I'm planning on building a Java server that will handle real time game communications between clients. What is the best type of Java implementation out there that could efficiently and, hopefully, accurately communicate between a client and server at high speeds (say 5-15 packets per second)? I know there are many types of Java networking APIs (ie. ObjectInputStream and ObjectOutputStream, DatagramPacket, KyroNet, etc.), but I'm not sure what is the most effective and/or commonly used implementation for such a scenario. I would assume that most real time games use UDP communication methods, but I understand the reliability issues that come with it. Are there UDP implementations that have some form of flow control? Anyway, thanks in advance!
A few things to consider:
Java NIO is really good, and can handle the kind of throughput/latency you are looking for. Don't use any of the older networking / serialization frameworks and APIs
Latency is really important. You basically want a minimal layer over NIO that allows you to send very fast, small, inidividual messages with minimal overhead.
Depending on the game, you may want TCP or UDP or both. Use TCP for important messages, UDP for messages that aren't strictly necessary for the game to proceed or will be subsumed by a future update (e.g. position updates in a FPS)
Some people implement their own TCP-like messaging protocol over UDP for real time games. This is probably more hassle than it's worth, but be aware of it as an option if you really need to optimise for a specific type of communication
For real time games, you are nearly always doing custom serialisation (e.g. only sending deltas rather than full updates of object positions) - so make sure your framework allows this
Given this, I'd recommend one of the following
Kryonet - lightwieght, customisable, designed for this kind of purpose
Netty - slightly more enterprise-oriented, but very capable, robust and scalable
Roll-your-own based on NIO - tricky but possible if you want really fine grained control. I've done this before, but in retrospect I probably should have picked Kryonet or Netty
Good luck!
Immidiately forget ObjectOutputStream and ObjectInputStream. These are the standard output-input mechanisms of the old standard java serialization, which is slow and produces bloat objects. Some resources to start with:
http://code.google.com/p/kryonet/
http://code.google.com/p/pyronet/