Audio and video learning (III) -- SIP protocol

1. Concept

SIP agreement Full name “ Session initiation protocol ”, It belongs to the application layer control protocol . It specifies the standard for interactive communication between the terminals of one or more participants , can establish 、 Modify and disconnect the connection to realize video 、 voice 、 im etc. , stay OSI The model is located in the session layer , As shown in the figure below ：

  What kind of agreement is it ？

• SIP It's a Based on text Application layer control protocol , Strictly speaking, it belongs to the conversation layer , Used to create 、 Modify and release the session of one or more participants . These conversations can be Internet Multimedia Conference 、IP Telephone or multimedia distribution . The participants of the session can use multicast （multicast）、 Mesh unicast （unicast） Or a mixture of both ;
• SIP It's a kind of signaling protocols , For initial 、 management 、 Terminate voice and video conferencing on the network , Specifically, it is used to generate 、 Modify and terminate the session between one or more participants ;
• SIP yes Point to point protocol ;
• SIP The terminal of the protocol can be PC machine 、 Intercom terminal 、 Mobile phones with specific software installed 、 Flat 、 Customized devices such as web pages or voice robots . Mainly used for instant messaging ;
• SIP It is a kind of from the Internet IP Voice session control protocol , Have flexibility 、 Easy to implement 、 Easy to expand ;
• It's an application layer protocol , Work with other application layer protocols , adopt Internet Control multimedia communication session ;
• SIP With the reservation agreement responsible for voice quality （RSVP） interoperability . It also works with several other protocols , Including lightweight directory access protocol for location (LDAP)、 The Remote Authentication Dial in user service responsible for authentication (RADIUS) And responsible for real-time transmission RTP Waiting for multiple agreements ;

2. function

• User positioning ：SIP In the protocol, the called party can move in different positions , When the caller user requests to establish a session with the called party , towards SIP The server sends a request to find the user's location ;
• User effectiveness ： There can be more than one session participant , After receiving the call request , You can voluntarily choose whether to participate in the session ;
• User capability description ： The requesting party is at SDP Describes which media data or session parameters it supports ;
• Establish a conversation ： When a user sends a session request , After some necessary media parameters have been described clearly , The called party user accepts the request , And according to... In the request information SDP Description to configure , Need to send “Ringing" A ringing message establishes a session ;
• session management ： add to 、 Modify media stream parameters or other parameters ; Keep the conversation waiting 、 Activate a session and terminate an existing session .

3. Message type

SIP Protocol messages are text-based protocols , The standard message starts with a line 、 Message header and message body , It is divided into request message and response message , The difference is whether the starting line of the message contains status information .SIP The request message defines 6 Methods ：

• REGISTER register ：UAC Tell the location of the registration server itself at startup
• INVITE The invitation ：UA Initiate session invitation
• ACK: After receiving the invitation, the called party , Respond 200 or 180 confirm ,ACK Indicates that the caller has received a confirmation message , Session successfully established
• CANCEL: The called party cancels the session invitation of the calling party
• BYE： One of the session participants wants to exit the successfully established session , launch BYE news
• OPTIONS： Query the capacity and load of the proxy server

The response message is 6 Kind of , Including temporary response 、 The request is successful 、 Location exception notification 、 Request failure notification 、 Server failure or global error .

SIP The protocol supports message extension , Message type 、 Both the message header and the message body can be expanded , Such as the following situations

• Integrate telephone signaling into sip In the news , To solve the problems of public switched telephone network and sip Network and interconnection ;
• Solve cross domain firewall and NAT Communication problems ;
• refer Event expansion , Support call forwarding ;

4. Main components

• End user equipment UA： For creation and management SIP Protocol session device ;
• Location server ： Record and manage UA Location information for ;
• Register the server  ： To determine where the user is ,UA The terminal registers with a registration server , The registration server will assign a specific address to the terminal UA , The address has a life cycle ,UA The registration status needs to be refreshed regularly to keep alive ;
• proxy server ： Accept UA Session request and query the registration server , Get recipient UA Address information for . then , It forwards the session invitation information directly to the recipient UA（ In the same domain ） Or proxy server （UA In another domain ）. According to the processing method of the request, it is divided into stateful proxy server and stateless proxy server , Stateless proxy servers only deliver simple messages , The stateful proxy server will retain the relevant information of the received request and response , Used to process subsequent messages related to this message . Manage the whole life cycle of messages by creating a transaction state machine , Each is handled by the transaction state machine , Until the transaction is complete ;
• Redirecting servers ： After receiving the request, query the location information of the requester , Then create a location list in the reply message. Please return this information , allow SIP The proxy server will SIP The protocol session invitation information is directed to the external domain ;

5. Standard agreement

SIP The protocol is designed to be very simple , Has a limited set of commands . It is also text-based , So anyone can read in SIP Passed between endpoints in a session SIP news .

SIP Our standard protocols are IETF To formulate the , therefore SIP All the standard protocols are passed through RFCXXXX The way to publish , Each agreement sets out some SIP Characteristics of ：

• RFC3261：SIP The basic agreement of , Defined SIP Basic functions of , Characteristics, etc. . We need to SIP Words , This agreement cannot be ignored . Someone on the Internet has translated it into Chinese , Combined with the original English agreement , It should be easier to understand .
• RFC3262：SIP in , How to locate the server . There is not much research on this , General SIP The protocol stack can support , Just let them do it .
• RFC3265： Event notification mechanism , Some event monitoring can be carried out through this protocol .
• RFC3515： Call forwarding .
• RFC3666： And PSTN Description of some features when connecting .
• RFC3911： adopt Join To conduct the meeting .

6. Implementation mechanism

SIP Is a layered protocol , Its behavior is described by a set of equal and independent processing stages , There is a loose coupling relationship between each stage .

 

• SIP The lowest level of is its syntax and coding . Its encoding uses an extended background - Noel form Syntax (BNF) Appoint .
• The second layer is the transport layer . It defines how the client sends requests and receives responses , And how the server receives requests and sends responses over the network . all SIP The elements contain the transport layer .
• Next is the trading floor . Transactions are client transactions ( Using the transport layer ) Requests sent to server transactions , And all responses to the request sent back to the client from the server transaction . User agent client (UAC) Any task completed is performed using a series of transactions . Stateless agents do not contain a transaction layer .
• The layer above the transaction layer is called the transaction user . Except for stateless agents , Every SIP Entities are transaction users .

7. Conversation process

SIP The agreement is the same as HTTP The protocol is similar and adopts some of the same design principles, such as client UAC/ Server side UAS Model , The active initiating requester is UAC End , The receiving end is UAS End , The integrity of a request is called a transaction , End with one or more responses , Multiple temporary responses can be supported in the middle , In the process UAC and UAS Your roles are interchangeable , As in a call transaction A Make a call ,A yes UAC, The other end is UAS; Who hangs up first at the end of the call is UAC, The other end is UAS.

SIP By agreement INVITE request , With SIP URL The method of means to initiate , With BYE Request to end . There are three ways according to different network structures ： Two UA Call directly between 、 Call through proxy server and call through redirection .SIP Generally, the following protocols need to be used to complete the communication at one time ：

• SDP： Session name 、 Conversation time 、 Which one? IP port 、 Types of multimedia supported by session participants 、 Transport protocol type 、 Encoding and decoding algorithm and other information
• RTP： Real time transmission protocol
• RTCP: Real time transmission control protocol , and RTP Control protocol for working together
• TLS： Provide security and integrity protection

7.1 One to one call

Session creation process without proxy server , Direct realization SIP Message initiation and response process .

7.2  Call... Through a proxy server

7.3  By redirecting the call

UAC It's really UAS After the address , The subsequent communication process is the same as the one-to-one call process .

stay 200 OK In the transmission process, if the opposite end does not receive after timeout , Then it means that the lost message needs to be sent again , until UAS received ACK request , Ensure the complete completion of reliable transmission .