A The Session Initiation Protocol (SIP) is a signalling protocol used to establish multimedia communications sessions across an IP network. This could be anything from a simple two-way telephone call across the Internet to a collaborative multimedia conference session that includes the use of video and text chat between two or more people located in a variety of different locations.
SIP has now largely been adopted as a replacement for the H.323 protocol, which was previously the de facto standard for supporting audio, video and data transmissions as well as internet telephony and voice-over-IP (VoIP) communications across private LANs and the internet.
SIP is defined in the Internet Engineering Task Force’s (IETF’s) RFC 3261 standard, although elements of it are included in more than 100 RFCs and internet drafts. SIP continues to be modified and extended as the technology advances and more SIP products are launched.
Q How does SIP fit into voice over IP (VoIP)?
A Like HTTP and SMTP, SIP is a request-response protocol that works in the application layer of the Open Systems Interconnection (OSI) communications model. It therefore integrates smoothly into web browsers, messaging applications and other internet services, providing features like embedded audio files, click-to-dial calling, instant messaging (IM) and IP centrex services.
It also works well in conjunction with other standards and protocols like Remote Authentication Dial-in User Service (Radius), which authenticates dial-in users; the Lightweight Directory Access Protocol (LDAP), which locates organisations, individuals, and other resources such as files and devices on the net; and the Resource Reservation Protocol (RSVP), which helps with best-effort call routing.
SIP’s flexibility therefore gives service providers a single toolkit with which to build converged voice and multimedia services, while many VoIP vendors have adopted SIP for signalling purposes in IP telephony equipment.
Q What does a SIP architecture consist of?
A There are two basic SIP components: the User Agent Client (UAC) and User Agent Server (UAS). UACs can take the form of embedded software clients in end-user devices such as IP handsets or handheld PCs, or integrate with desktop applications such as contact managers or collaboration tools. The UAS sits in the network device that handles the signalling.
To confuse matters, the UAC itself has both a client and a server element: the client element initiates the calls and the server element answers them, a combination that allows peer-to-peer calls to be made using a client-server protocol. The main function of the UAS is to provide name resolution and user location and to pass on messages to other SIP servers using next-hop routing protocols.
SIP servers operate in either stateful or stateless modes. A stateful server records all of the incoming requests it receives, along with the responses it returns and the outgoing requests it forwards on to other SIP servers. In contrast, a stateless server forgets all information once it has sent a request. Stateless servers usually make up the SIP infrastructure on telecommunications networks, while stateful servers sit on private company LANs to keep track of local user activity.
Q How does SIP work?
A The call set-up procedure follows four steps. First, the caller sends a request to initiate a communications session to the SIP URL of the called party or recipient. If the caller knows the location of the party they are trying to contact, they can send the request directly to their IP address; if not, they can send it to a locally configured SIP network server.
This SIP server then attempts to resolve the recipient’s location. This is done either by searching the network’s domain name service (DNS), accessing contact databases or redirecting calls to additional SIP servers until it discovers one that does know the IP address of the recipient.
Once the address is found, the request for a communications session is forwarded on to the recipient. When the recipient’s telephony client receives the request, it provides the user with the option of taking the call. If this is declined, the SIP server forwards the call on to either a voice mail server or another user; if the call is accepted, the two SIP clients swap information about how they want to communicate, for example using voice only, video, text chat or white board applications, and a connection is established.
Q What SIP features are most likely to appeal to firms?
A A SIP server can split an incoming call so that it rings at several extensions simultaneously and communicates with the first one to answer – useful in office environments where either the intended recipient divides his or her time between two desks, or where the caller wants to talk to any one of a number of people.
SIP can also introduce different types of content half way through a session, by sending a video clip for example, or some form of online document that needs to be displayed or filled in.
As well as user location and availability information, SIP services can provide standard telephony features like call forwarding and personal mobility, whereby the user is contactable on the same number wherever they happen to be.
Because SIP identifies users by an email-like address, like sip:user@company.com for example, it is just as easy to redirect a caller to another phone as it is to redirect them to a web page. When routing calls, SIP servers can create emails related to those calls and forward those on as well, providing the basis for a unified messaging system. The use of Multi-purpose Internet Mail Extensions (Mime) means that SIP calls can also contain Java applets, images, audio files, authorisation tokens or billing data that are copied from or linked directly to specific web pages.
Q How interoperable are the various SIP vendors’ products?
ASIP was created purely as a mechanism to initiate, terminate and modify communication sessions, and is not concerned with the content of those sessions. This focus on delivering a certain basic set of functions enhances the protocol’s simplicity, and lends it a high degree of interoperability and scalability that enables it to be implemented across a wide range of different network and computer architectures.
Due to its relative immaturity, however, SIP is likely to suffer the usual interoperability and deployment challenges that plague early versions of any technology.
Most equipment vendors will only include a small subset of core SIP features in their products. This is partly because they do not see any demand for some of SIP’s more advanced features among potential customers, and partly to keep costs down. They may also want to protect their existing PBX and handset business by extending SIP’s basic functionality with proprietary protocols that allow SIP to communicate with their legacy equipment.
The upshot of all this is that interoperability problems are bound to occur, as demonstrated by testing conducted by Network World magazine earlier this year. Of 1,100 interoperability tests on 75 products from 25 telephony vendors, including IP handsets, softphones, SIP proxies, firewalls and wired and wireless switches, 18 percent proved outright failures and only 43 percent passed the basic call transfer interoperability tests.
Q What are SIP’s limitations?
A By concentrating on delivering only core SIP functionality to avoid interoperability problems, vendors run the risk of dumbing down SIP equipment to levels where it no longer satisfies the demand for advanced telephony features within the enterprise, such as least-cost call routing, non-stop failover, performance monitoring and reporting and centralised call system management.
- www.tinyurl.com/9nv6j
- www.sipcenter.com
- www.tinyurl.com/dmlez
- www.sipforum.com
Ltd
