I am implementing a TURN server specifically for WebRTC usage and have some questions regarding not supporting certain attributes (send an error response if the attribute is received) or simply ignore them or other doubts. Here they are:
EVEN-PORT If my SDP always signals a=rtcp-mux, will this attribute ever be used? And if so, would it be an error if it appears?
RESERVATION-TOKEN Does this play any role when TURN server is used in the WebRTC context?
SOFTWARE As in STUN, can this be safely ignored without any processing?
DONT-FRAGMENT Is there a preferred and well-accepted norm for this attribute in the WebRTC context?
What is the ideal length of NONCE in the WebRTC context?
Different issue. Are there any statistics available for use of TURN server for transports other than UDP? I am thinking of supporting only UDP for now.
webrtc typically requires rtcp-mux, at least in chrome so I would not care about even-port.
no
yes. It is FYI only.
no. WebRTC implementations typically don't do path-mtu discovery but assume 1200 bytes.
You mean the expiration? https://medium.com/confrere/gone-in-1100-seconds-hunting-bugs-on-the-edge-of-webrtc-132a186c45dd
see https://medium.com/the-making-of-whereby/what-kind-of-turn-server-is-being-used-d67dbfc2ff5d
Related
As a follow-up to my previous post (ApiRTC - Behaviour with meshModeEnabled and meshOnlyEnabled)
Hello,
You say that SFU is necessary for any activity that requires centralizing all the streams (recording, bandwidth optimization,...). However, in MESH mode, the files/media exchanged still manage to be recorded on the Apizee media server even though I don't go through the SFU. How is this possible ?
Can this behaviour be disabled so that the exchanged documents never leave the MESH stream ?
I have not found anything about this in the documentation.
By the way, the documentation often mentions the term "MCU", does this mean that ApiRTC also uses an MCU server in addition to the SFU ?
Thanks in advance.
apirtc
Can this behaviour be disabled so that the exchanged documents never
leave the MESH stream ?
Concerning a recording of all the streams in the conversation (via the startRecording method of the Conversation object see https://apirtc.github.io/references/apirtc-js/Conversation.html#startRecording__anchor):
--> The composition of multiple streams into one video file is done server-side by the SFU (v4.4.8).
Concerning the files (through conversation.pushData method):
--> We manage the file transfer through uploading the file on a storage and share the URI to all parties of a conversation. P2P transfer is not available (v4.4.8)
To exchange data in a P2P mode, you can use the Conversation.sendData method to send raw data across all participants.
Regarding your question about the MCU, no, ApiRTC doesnt use any MCU server to date (v. 4.4.8). The document refers to MCU for very specific on-premise deployment, not supported for ApiRTC users.
Cheers,
Romain
I was reading TURN server RFCs. All related RFCs ( 5766 and the more recent 8656) support Channel mechanism to avoid the 36 bytes overheads of STUN headers (Section 2.5 of RFC 5766) required for the send/data approach:
For some applications (e.g., Voice over IP), the 36 bytes of overhead
that a Send indication or Data indication adds to the application
data can substantially increase the bandwidth required between the
client and the server. To remedy this, TURN offers a second way for
the client and server to associate data with a specific peer.
For WebRTC, clearly there is no point in using the send/data mechanism. How do browsers choose between the two mechanisms for relaying? Is send/data a fallback? Will support for Channels alone in a TURN server be sufficient for WebRTC use-case?
They will usually do SendIndications while waiting for the Channel to be created.
SendIndications also are important if you get something on the relay before the Channel is created. Some clients only create the Channel when they send and not right when the permission is created.
Firefox doesn't support TURN channels: https://bugzilla.mozilla.org/show_bug.cgi?id=857736
Chrome also uses send/binding indications until ICE is done (presumably to avoid the overhead of creating channels which are not used later)
Don't rely on partial implementations of a spec, that won't work.
I've read a lot about WebRTC, but there's one question that still remains. I hope you can help me with that:
Does WebRTC allow me to create a one-to-many connection? I don't mean "being able to have multiple connections to different computers", I really talk about having one connection that multicasts its data to multiple endpoints without the need to "upload" the data once for each endpoint. Will it be possible to send one single package to the web, that, when it reaches the web, magically splits itself into multiple packages with different targets?
I hope you get what I'm looking for :)
Until now, I've only seen one-to-one connections, or solutions that have one connection to a central server that does the multicast for them (which usually results in twice the ping).
But to me, one-to-one connections don't seem to be really useful (due to low upload-bandwith of clients), and solutions with a central server are also possible without WebRTC (using WebSockets), so the only real use case for WebRTC would be one-to-many connections.
So.. is this something that will be possible in the future? Or is it already possible today?
Three things:
IP multicast in the Internet is not possible at the moment (multicast addresses are not routed by ISPs)
WebRTC fits many use cases beyond one-to-many communication, just have a look at this document: https://datatracker.ietf.org/doc/html/draft-ietf-rtcweb-use-cases-and-requirements-06
WebRTC connections between browsers are always encrypted (using SRTP for A/V data and DTLS for generic data) and the encryption parameters (session keys etc.) are negotiated for every connection separately. How would you do that in a multicast environment (think of it as a distribution tree)?
So no, WebRTC cannot be used with IP multicast.
I would answer "It doesn't for now", because as a programmer, I can tell you, that there are number of ways browser devs to make it work if we (users) insist on it's importance. But how ? Since there's encryption, they could allow sharing of the session's encryption keys to the group of 'registered' (multicast) users. But how ? Well, Web was created for sharing. The most obvious way is through web server mediation and JS WebRTC API function (to load the user keys). Since multicast is most often used for efficient video distribution, you have a RTP/SRTP video server. The web server can coexist at the same machine. If they decide to extend it to web browsers - then just the "server" role can be done by the Web browser who created the multicast stream (the sender). The clients need to know who is it.
Again: In December 2013, this is still not possible. And multicasts are allowed on the Internet only in:
some experimental WAN nets
some internet+video ISP nets
LANs (when enabled at switch level, cheap switches transmit it to all ports). But you can be an ISP, researcher or LAN user, so it's necessary.
We are running a course in robotics and Xbee is the most favorite communication protocol for the student. In last two years we helped them build around 62 various projects (40 more in pipeline).
All most all the projects involve sending different kind of data to the bot. Sometimes it is a 1 byte command where as sometimes it is a long string to be interpreted. Sometimes we face the issue of addressing a bot when one xbee is used in broadcast mode to send messages to a particular bot among several. Students use their creativity to address this issue each time.
I personally feel this is reinvesting the wheel. I wonder if any higher level protocol proposals exist for serial port communication and if there isn't any specific protocol design I wonder if if the worth designing one for the student needs.
Do you mean internal only protocol of your system? If yes, often embedded software engineers incline to roll their own protocols. Most of them talks that it lets them make most optimal system.
It is not ideal approach. I agree with you that it's good for students to learn good examples.
Unfortunately I don't know any protocol stack fitting well robotics application. But I advice you to try google's protocol buffer system, its able to simplify most efforts of building protocols engines, and it works with plain c too.
You can implement Modbus ASCII if you want to go with a standard protocol that's already open.
Comli is a master/slave protocol that is used in some older devices or when it is not possible to use ethernet. You can probably get the specification from ABB if you ask - it's no secret.
That said you can put an OPC server/client architecture on top of that to get a bit more powerful communication e.g.
+--------------+ +--------------+ +--------+
| OPC UA Client| -- | OPC UA Server| -comli- | Device |
+--------------+ +--------------+ +--------+
This would make your OPC UA client protocol indepedent which makes things a bit easier down the road.
Modbus is another serial protocol that is used a lot
I believe OPC will give you the highlevel operation that you want.
see
www.opcfoundation.org
www.abb.com
PS. OPC UA is not the same as the old OLE version and thus has nothing to do with COM/DCOM
Like mjh2007 said, Modbus is standard, open and easy. The only problem I can see is if you want the robot to respond "quickly" to a command, since serial Modbus uses timeouts to detect the end of a packet. You can get around this by ignoring the timeout requirements and calculating the expected size of a packet based on it's function code and parameters as you are receiving it, then you can start processing the command immediately upon receiving the last byte and verifying any checksums. This page has some more details on implementing such a scheme.
Be sure to make use of the XBee module's "Transmit Explicit" frame (type 0x11) running in API mode with ATAO set to 1. You can unicast to a particular bot on your network, instead of always broadcasting frames. On a mesh ZigBee network, you want to avoid broadcasts as much as possible.
I'm guessing you're either using "AT mode" for sending raw data, or using "API mode" with ATAO set to 0 (sometimes referred to as "transparent serial").
If you look at that frame type (0x11), you'll see that the recipient gets an 0x91 frame that contains multiple fields already (source/destination endpoint, cluster, profile ID). You can re-purpose those fields since you're not trying to do ZigBee networking.
I'm building an Objective-C app that has both a server and a client. The client can send updates to the server, and the server needs to be able to send updates to each connected client. I've been thinking about how best to implement this system, but am asking for your suggestions.
Currently, I'm thinking that when new updates are available, the server will use threads to send the update to each client in turn. If a client times out, they are disconnected.
I have very little networking experience, so am asking your insight.
Do you think that this system would work well?
If so, do you have any suggestions about how to do the threading? Any NS classes you can point me at? There's got to be some kind of queue I can use, I'm thinking.
Any other thoughts?
EDIT: I do not expect the client count to get much above 50 or so, at the max.
As long as both client and server are OS X apps and can both be written in Objective-C using the Cocoa frameworks, I would highly recommend you take a look at the Distributed Objects (DO) technology in Cocoa. I won't try to give a tutorial in Distributed Objects here, just explain why it might be useful...
DO handles asynchronous network details for you (all your client updates could happen on a single thread). In addition the semantics of communication with a remote object (client to server or visa versa; DO is bidirectional once the connection is established) are very similar to in-process communication. In other words, once you have a reference to the remote object (really an NSDistantObject which acts as a proxy to the object on the other end of the connection), your client code can send messages to the remote object as if it were local:
[remoteServer update:client];
from the client or
[[remoteClientList objectAtIndex:i] update:server];
from the server. I'll leave the details of setting up the connection and for getting the remoteServer or remoteClient reference to you after reading the Distributed Objects programming guide.
The downside of using DO is that you are tied to Cocoa; it will be very difficult to write a non-Cocoa client or server that communicates using Distirbuted Objects. If there's a chance you may want to have non-Cocoa client or server implementations, you should not use DO. In this case, I would recommend something simple with a lot of cross-platform and language support. A REST-style API over HTTP is a good option. Have a look at the Cocoa URL Loading System documentation for info on how to implement HTTP requests and responses. Have a look at Apple's CocoaHTTPServer example code or a code.google.com project of the same name for info on implementing an HTTP server in your Cocoa code.
As a very last option, you can take a look at the Cocoa Stream Programming Guide if you want to implement your own network protocol. NSStream's subclasses will let you listen on a network socket and handle asynchronous reads/writes to/from that socket. A lot of people use AsyncSocket for this purpose. It wraps the (lower-level) CFStream and CFSocket and makes writing network code somewhat easier.
When the server sends updates to the clients, it would probably be easier to just have one thread handle them all, and just use async sockets. Of course this would depend on how many clients you had to deal with too.
There's several networking examples in the apple developer side.
One I would recommend that you check out is the URLCache, which can be downloaded.
Quoting from the Apple's documentation for this example:
URLCache is a sample iPhone application that demonstrates how to download a resource off the web, store it in the application's data directory, and use the local copy of the resource. URLCache also demonstrates how to implement a couple of caching policies:
An interesting option is the BLIP protocol from Jens Alfke. It's like a stripped down version of BEEP: a message oriented networking system. It basically provides the low-level abstractions for a bidirectional message pipe so you can concentrate on layering your communication protocol on top of it.
It has some worthy followers such as Marcus Zarra (author of the CoreData bible) and Gus Mueller of Flying Meat software.
I don't know how you plan to design you system, but usually a server cannot connect to a client; the client must initiate the communication. With a low limit of 50 clients, you may not be looking at a web-server/client-like implementation...
That said, there are basically two ways to handle client server communication:
1. The client polls the server periodically to get updates
2. The client keeps a connection open to the server and the the server responds with a well known (as in both sides understand it) protocol.