I have a server which is able to relay the WebRTC media data from A to B. For the video conferencing, if we go with P2P approach then a mesh network is created. Whenever P2P doesn't work, we can have this relay server.
The main problem is that in the mesh network, the number of upload link is "N - 1" for N participants. Hence the number of connection goes upto N * (N-1). Usually mesh network allows 5-6 stable connections.
Many online sources suggest to implement SFU. If SFU decrypts the media data and then re-encrypts for every peer then that virtually requires a WebRTC component on the server side.
Are there any lightweight C/C++ based library which helps in this regard?
Any better alternative strategy for the same?
BTW, I tried to share the same offer to all the peers with their own answer, but as expected it didn't work. The peer gets disconnected after receiving few chunks.
I have referred below related posts:
WebRTC - scalable live stream broadcasting / multicasting
Multi-party WebRTC without SFU
Can I re-use an "offer" in WebRTC for mulitple connections?
There are quite a few free and open source projects that implement an SFU:
Jitsi is probably the best known, but it is written in Java, and might therefore be unsuitable in some deployments;
Janus is written in C; it is small, efficient, and well supported, but might not be the easiest to understand;
Ion-SFU and Galène are written in Go, and might be easier to adapt to your needs.
Related
How does ZeroC ICE compare to 0MQ? I know that 0MQ/Crossroads and DDS are very similar, but cant seem to figure out where ICE comes in.
I need to quickly implement a system that offloads real-time market-data from C++ to C#, as a first phase of my project. The next phase will be to implement an Event Based architecture with an underlying Pub/Sub design.
I am willing to use TCP.. but the the system is currently running on a single 24 core server.. so an IPC option would be nice. From what I understand ICE is only TCP, while DDS and 0mq have an IPC option.
Currently ,I am leaning towards using Protobuf with either ICE or Crossroads IO. Got turned off from the OpenSplice DDS website. Ive done lots research on the various options, was originally considering OpenMPI + boost:mpi, but there does not seem to be MPI for .NET.
My question is:
How does ICE compare to 0MQ? I cant wrap my head around this. Was unable to find anything online that compares the two.
thanks in advance.
........
More about my project:
Currently using CMAKE C++ on Windows, but the plan is to move to CentOS at some point. An additional desired feature is to store the tic data and all the messages in a "NoSql" database such as Hbase/Hadoop or HDF5. Do any of these middleware/messaging/pub-sub libraries have any database integration?
Some thoughts about ZeroC:
Very fast; Able to have multiple endpoints; Able to load balance on the endpoints; Able to reconnect to a different endpoint in case one of the node goes down. This is transparent to the end user; Has good tool chain (IceGrid, IceStorm, IceBox, etc); Distributed, high availability, multiple failover, etc
Apart from that, I have used it for hot swapping code modules (something similar to Erlang) by having the client create the proxy with multiple endpoints, and later on bring down each endpoint for a quick upgrade one by one. With the transparent retry to a different endpoint, I could have the system up and running the whole time i did an upgrade. Not sure if this is an advertised feature or an unadvertised side-effect :)
Overall, it is very easy to scale out your servers if need be using ZeroC Ice.
I know ZeroMQ provides a fantastic set of tools and messaging patterns and I would keep using it for my pet projects. However, The problem that i see is that it is very easy to go overboard and lose track of all your distributed components. This is a must have in a distributed environment. How will you know where your clients/server are when you need to upgrade? If one of components down the chain does not receive a message, how to identify where the issue is? the publisher? the client? or any one of the bridges (REP/REQ, XREP/XREQ, etc) in between?
Overall, ZeroC provides a much better toolset and ecosystem for enterprise solutions.
And it is open source :)
Jaybny,
ZMQ:
If you want real good performance and the only job for Phase 1 of your job is to move data from C++ to C#, then Zmq is the best option.
Having a pub/sub model for event driven architecture is also something that Zmq can help you with, with its in-built messaging pattern.
Zmq also supports your IPC requirements in this case. Eg: you can have one instance of your application that consumes 24 cores by multithreading and communicating via IPC.
ZeroC Ice:
Ice is a RPC framework very much like CORBA.
Eg.
Socket/ZMQ - You send message over the wire. Read it at the other end, parse the message, do some action, etc.
ZeroC Ice - Create a contract between client and server. Contract is nothing but a template of a class. Now the client calls a proxy method of that class, and the server implements/actions it and returns the value. Thus, int result = mathClass.Add(10,20) is what the client calls. The method, parameters, etc is marshalled and sent to the server, server implements the Add method, returns the result, and the client gets 30 as the result. Thus on the client side, the api is nothing but a proxy for a servant running on a remote host.
Conclusion:
ZeroC ICE has some nice enterprisy features which are really good. However, for your project requirements, ZMQ is the right tool.
Hope this helps.
For me.. the correct answer was Crossroads I/O . It does everything I need.. but still unable to pub/sub when using protobufs... im sure ZeroC ICE is great for distributed IPC, but 0MQ/Crossroads, gives you the added flexibility to use Inter-Thread-Communication.
Note: on windows, 0mq does not have IPC.
So, all in all, the crossroads fork of 0mq is the best. but you will have to roll your own windows/ipc (or use tcp::127..) , and publisher side topic filtering features for pub/sub.
nanomsg, from the guy who wrote crossroads and 0mq (i think).
http://nanomsg.org/
I am implementing a messaging system using C++ and Qt. After much thought, I have determined that multicasting or a multicast-style technique will work best to solve my problem. However, I have learned about UDP's unreliability and believe it to be unacceptable.
My requirements are as follows:
Messages are to be sent in a binary serialized form.
From any given node on the network, I must be able to send messages to the other nodes.
Message delivery must be insured.
I have heard of OpenPGM and NORM as alternatives for UDP. If anyone has experience with either of these, could you please share?
I am also open to the possibility of implementing "reliable" multicast by myself, in the application layer, but I would prefer not to if there is a library that already implements this.
I am using C++ and Qt, therefore .NET or Java-based solutions are not acceptable unless they are open-source and I may port them to C++.
Thank you very much.
EDIT 20120816T1853 MDT: An additional question: would either PGM or NORM have to be implemented at the hardware/IP level? Or can they be overlayed on top of existing protocols?
We have implemented our own reliable multicast protocol over UDP called RSP, since we needed something cross-platform and at the time couldn't find a good solution between Linux and Windows. The Windows PGM implementation disconnects slow clients which leave the send window, whereas our implementation throttles the sender similar to TCP. Afaik OpenPGM can be configured to do the same.
The open source NORM at http://cs.itd.nrl.navy.mil/work/norm can be built as a library and has C++ API with Python and Java language bindings. If you ping the developer (me) via the mailing list, I can help get you started.
There is a large RFC division of reliable multicast protocols, and many implementations out there. It's a long time since I looked at this but there are TRAM, LRMP, ...
In a few weeks I'm thinking of helping a project finish a pretty major aspect of a C++ world with 3D characters: networking. I will help with the server's transfer of information from/to clients. I already know C++ well enough. I just need to know what specifically I should know to do this and resources from which I could find this information. Thanks :)
as RageD said, it's a big difference in networking between different types of games. A FPS server typically sends complete game state to all clients regularly (e.g. 60Hz) over UDP. Other game types can use TCP (tuned a bit like TCP_NODELAY and forcing immediate ACK packets) or reliable UDP (raknet lib or others). Network protocol can become really wide so you'll need to think how to make it easily extendable. I'd recommend you to start from here: http://www.gamedev.net/community/forums/showfaq.asp?forum_id=15
In my application we will present the video stream from a traffic camera to a client viewer. (And eventually several client viewers.) The client should have the ability to watch the live video or rewind the video and watch earlier footage including video that occurred prior to connecting with the video stream. We intend to use wxWidgets to view the video and within that we will probably use the wxMediaCtrl.
Now, from the above statements some of you might be thinking "Hey, he doesn't know what he's talking about." And you would be correct! I'm new to these concepts and I confused by the surplus of information. Are the statements above reasonable? Can anyone recommend a basic server/client architecture for this? We will definitely be using C++ wxWidgets for the GUI, but perhaps wxMediaCtrl is not what I want... should I be directly using something like the ffmpeg libraries?
Our current method seems less than optimal. The server extracts a bitmap from each video frame and then waits for the single client to send a "next frame" message, at which point the server sends the bitmap. Effectively we've recreated our own awkward, non-standard, inefficient, and low-functionality video streaming protocol and viewer. There has to be something better!
You should check out this C++ RTMP Server: http://www.rtmpd.com/. I quickly downloaded, compiled and successfully tested it without any real problems (on Ubuntu Maverick). The documentation is pretty good if a little all over the place. I suspect that once you have a streaming media server capable of supporting the typical protocols (which rtmpd seems to do), then writing a client should fall into place naturally, especially if you're using wxWidgets as the interface api. Of course, it's easy to write that here, from the comfort of my living room, it'll be a different story when you're knee deep in code :)
you can modify your software such that:
The server connect, server grabs an image, passes it to ffmpeg establishing stream, then copy the encoded data from ffmpeg stream and send to client via network, if the connection drops, close the ffmpeg stream.
Maybe you can use the following to your own advantage:
http://www.kirsle.net/blog.html?u=kirsle&id=63
There is a player called VLC. It has a library for c++ and you can use it to embed the player in your GUI application. It supports a very wide range of protocols. So you should leave connecting, retrieving and playing jobs to VLC and take care of the starting and stopping jobs only. It would be easy and probably a better solution than doing it yourself.
For media playing facility, both music and audio, you can a look on GStream. And talking about the server, I think Twisted (A network library in Python) should be good option. The famous live video social website justin.tv is based on Twisted. Here you can read the story from here. Also, I built a group of server for streaming audio on Twisted, too. They can serve thousands of listener on line in same time.
Closed. This question needs to be more focused. It is not currently accepting answers.
Want to improve this question? Update the question so it focuses on one problem only by editing this post.
Closed 8 years ago.
Improve this question
I have one computer that is running a c++ program to control a robot and another computer that is running a computer vision system (also written in c++). I would like for these two programs to be able to talk to one another. The communication would not have to be complex, I would just need the robot computer to be able to tell the vision computer when a trial begins and ends (when to start and stop a data collection sequence). Do any of you have advice on how to approach this problem? Rs232 communication between the computers? Some kind of networking solution? smoke signals? Any suggestions would be welcome.
thank you in advance
(edit) In case you think the statement above is vague:
I need to pass a binary (go/don't go) signal from one computer to another. Unfortunately I can't be more specific about what this will look like because (obviously) I don't know what is available. Both computers are on a network, and both computers are running windows. The goal is to syncronize data collected by the computer vision system with actions performed by the robot. The communication does need to be fast enough that it will not slow down either the robot or the computer-vision program. a "good" solution would be 1) easy to implement 2) fast. I do not know much about networking and I am looking for a place to start looking.
thank you again for your assistance
You might use a simple UDP protocol - the advantage being that if you understand the concepts of simple packet protocols on RS232 you'll find it easy to transfer that knowledge to sending the packets via UDP.
If you want a reliable (as in, other parts of the system will worry about errors and retries) stream of bytes between the two PCs, then TCP/IP is not much more complicated to use than UDP.
Both UDP and TCP are accessed through 'sockets'. I'm afraid you'll find that from C++ there is rather a lot of tedious boilerplate to getting that working, but there are lots and lots of examples around.
If they are network-connected you could just use sockets.
The best option will be to use network communication. The easiest way to approach this should be to look at the networking examples in Qt.
You basically will create a client and a server application. You decide what the client does when it sees a certain message from the server. That's all. Qt should take care of the rest of the stuff.
Other answers suggests TCP/IP, UDP, RS232, ... All those things are just options when you use QtNetwork module. I assume that since you ask your question, you don't know about the difference between those. So the safest bet will be to use the highest level (free) library, hence the suggestion to look into Qt.
Another option is to use Boost.Asio. I tend to prefer Qt solution since their API is nicer.
That sounds like a fairly good use for the network socket. If both your machines are on Windows you can even use named pipes.
For Windows, you will need to open the COM n port as a file to communicate over a serial port[1]. I don't have access to my code now, I can look it up when I get home.
RS232 is easy and I like it. However, it it is slow. You need to consider that in your design.
[1] For C++.
Most modern computers have Ethernet capability, so get yourself a cheap hub or switch and look at networking APIs. There's usually some fairly easy socket stuff. One advantage of this is that, if you want to increase communication ability later, such as having your vision software provide instructions and guidance to your robot, you've got the basics set up.
Alternately, set up your vision program so you can start and stop it by hitting random keys. When you're going to use it, put the keyboard in front of the robot computer's CD drive, and eject at the start and end of the robot run.
This may be overkill in your situation, but if I were in your shoes I would probably implement it using the HTTP protocol. The vision computer would run a HTTP server and the robot computer would communicate the state changes using POST requests. The Poco C++ Net library provides you with the facilities required to do this.
I would use a TCP/IP socket for communications. TCP guarantees that the data will make it. So, all you need to do is parse the data.
RS232 is an easy option to program for, however modern PCs don't tend to have RS232 ports. You may need to get USB-RS232 adapters or install a PCI card.
The other problem with RS232 is that you have an additional wire to worry about which can be a nusiance. Also RS232 cables can be limited in length (5-15m) unless you invest in some clunky RS232 repeaters or bluetooth connectors, etc.
On top of all that you're also adding one more item to your project that can go wrong and cost you time in deploying and debugging.
IMO, an elegant engineering solution would be to utilise the hardware that you have and use TCP/IP sockets to communicate.
The web is awash with examples on passing messages between servers and clients:
If you're using Linux:
http://www.linuxhowtos.org/C_C++/socket.htm
Using Windows:
http://www.adp-gmbh.ch/win/misc/sockets.html
I also might look at something like 0MQ to make the connection more robust. It will transmit and reassemble messages regardless of the transport, and handle buffering in the case of temporary loss of connectivity.
But the bottom line is that I would use TCP/IP, but depending on the nature of the robot you may want a slightly more robust connection system than TCP sockets. UDP is nice because it's connectionless-- if the robot temporarily travels out of range/sight/etc you wont have to rebuild the socket and context.