Making a Call using SIP
When you ‘dial' to make a VoIP call, an invitation is issued to the requested VoIP User Agent. The sending UA can either enter a URI [e.g. sip:user1@domain.com] or a plain old phone number, to send the invitation. The invitation is handled by the SIP proxy server which returns a ‘100' message to the calling UA, indicating that the call sequence has been initiated, while it looks for the receiving UA.
If successful, the invitation is answered by a ‘200' response from the remote User Agent which verifies the receiver is ready and that parameters are established for a communication session. At this point the SIP proxy hands off the session to the two endpoints which establish a Realtime Transport Protocol stream to carry the conversation.
H.323
An alternative VoIP architecture, H.323 is a standard defined by the ITU, thus originating within the telecommunications industry rather than the Internet industry. Originally designed to accommodate multimedia, multi-point conferencing, H.323 is intended to provide a complete description of many sophisticated multimedia functions and is used in many high-end telecommunications switching devices.
SIP and H.323 compete in some respects, and are interoperable in others. SIP is gaining ground particularly among companies originating from the computer and internet industries, while H.323 is the more mature of the two and is well-represented in the telecommunications sector.
SIP enjoys some distinct advantages in terms of simplicity, as opposed to the far more complex H.323 standard, and in fact there is a ‘format war' going on between these two with advocates from either side involved in sometimes acrimonious debate. Overall, the minimalist approach of SIP attempts to build only what is missing from existing communications technologies while synergising with other elements of the IP suite, whereas H.323 attempts, perhaps more ambitiously, to provide a detailed description of every facet of IP-based communications.
The VoIP Gateway
Basic VoIP can work as a peer-to-peer application between IP endpoints – meaning you and your friends can download a softphone and communicate with no intermediaries being required. But additional facilities are needed to make VoIP a real alternative in telecommunications environment still dominated by the PSTN systems, and also to integrate VoIP equipment with the PSTN world.
To interface with PSTN networks and existing equipment, a device called a media gateway (often called simply a ‘gateway') is required. The gateway has a data network interface on one side, and a PSTN interface on the other, and its role is to intermediate between the two network types.
Gateways vary in size and scope. Industrial-strength, exchange-level gateways are used to intermediate large-scale networks, while single-port gateways, called ‘Terminal Adaptors' or ATAs, are used to connect a single device or single PSTN services through an ADSL or cable modem. It is these household-level ATA devices that are set to become very popular in the next 12-18 months as broadband-equipped households and offices begin to use VoIP in earnest.
Summary
This article has introduced some of the technologies and terminologies used in consumer-level VoIP technologies, in particular what is used to encode data for transmission across the Net, the steps and terminologies involved in initiating and managing a VoIP session, and some of the leading protocol suites.
Subsequent versions will provide more detail on the details involved in the business side of VoIP, which is still very much in development.