What is H.323?
H.323 is one of the oldest standards, which are used for VoIP telephony and video conferencing. This is an entire system of protocols and elements, which allows to transmit media data via packet networks with unlimited carrying capacity. The structure of H.323 recommendation provides various communication possibilities – starting from simple telephony and up to video conferencing with media data transfer.
One of H.323 advantages is its interoperability which allows to collaborate between devices from different manufacturers.
Evolution and further development of the standard
Many years have passed since the introduction of H.323, it being improved with the introduction of every new version. Currently there are seven versions of the standard starting from 1996.
The first version was quite poor because it was introduced with only one purpose – to enable communication between endpoints from different manufacturers. This version was not about reliability, safety and good connection quality. Moreover, previously incompatible endpoints could “communicate” only within a corporate network.
The second version introduced two years later became a breakthrough. It was aimed at the active use of VoIP telephony in multipoint conferencing. This time the key purpose became the reliability – endpoint (conference participants) authentication, fixed packet data transfer, protection from unauthorized data hack and oddly enough – absence of ability to reject incoming calls. Moreover, the process of connection between endpoints was accelerated and the feature of call redirection was added.
The third version provided signaling transport for a bigger number of calls via single TCP connection. Inter-network gateways, which could provide thousands of simultaneous calls, benefited from the new standard.
The changes in the fourth version were about increasing H.323 endpoints capacity. The launch of the fifth version was aimed at the general standard stabilization. By the way, the TrueConf solutions use the fourth version of H.323 protocol.
In June 2006 the sixth version of the standard was approved including changes in H.225 and H.245 transport protocols. Assigned Gatekeeper support became available where the endpoint from the list of alternative gatekeepers is registered. Besides, a number of applications and documents were supported, allowing to use GSM and H.264 codecs in H.323 solutions.
The final, seventh version of H.323 was launched in November 2009. It is worth to mention two most important new features for the users:
- Transfer of user information in different languages (which allowed easy collaboration between employees from various international organizations);
- Automatic data delivery about a group conference held on MCU server for all H.323 endpoints (which allowed users to join the conference without entering any data).
The H.323 standard is based on the four components of point-to-point or multipoint video conferencing:
- Zone controllers (gatekeepers)
- Multipoint Control Unit (MCU)
The endpoint - is the instrument for H.323 device management, sort of a user interface. Endpoints can connect with each other in VoIP telephony or video conference mode. Gateways are used to connect endpoints from various networks – for example H.323 and ISDN. They perform the following functions:
- Establishing connection between the endpoints.
- Conversion of sound formats.
- Data exchange.
If endpoints work within one H.323 network, the gateways are not used.
Zone controller or gatekeeper – is a central point of H.323 network, because it is responsible for addressing calls, it manages bandwidth and defines authenticity of endpoints and gateways during connection. However, the H.323 recommendation does not set a gatekeeper as obligatory. However, it is impossible to use many modern features without gatekeeper, which are implemented in VoIP applications and video conferencing solutions.
MCU (Multipoint Control Unit) is used to connect three and more endpoints. All endpoints, participating in a conference first connect to MCU server and then MCU server distributes video streams to all endpoints.
Every H.323 endpoint or device supporting H.323 protocol has its own IP address. H.323 packets routing in the network is carried out through the H.323 endpoint. UDP protocols are used to connect endpoints with gateways and gatekeepers and for media traffic transfer. TCP protocols are used only to establish a call between the endpoints and for exchange of extra features.
Media data transfer according to H.323 recommendation is divided into five basic steps:
- Gatekeeper detection and gatekeeper registration.
- Establishing a connection between two and more endpoints.
- Video and voice exchange through transport protocols.
- Multimedia exchange - transfer of different graphical and text documents and collaboration while using them.
- Ending the call.
The detection process is needed for endpoints to find a gatekeeper by its network address and to register on the endpoint. This procedure can be carried out automatically (multicasting – message exchange between the endpoints and the gatekeeper. If there are several gatekeepers, the endpoint chooses which to register) or manually (when the gatekeeper network address is known in advance during device configuration). The first option of gatekeeper detection is preferred, because the endpoint will be able to switch automatically to other gatekeeper without entering the endpoint configuration in case of any faults during the work of endpoint.
The registration procedure is needed to let endpoints send their address to gatekeeper and to enter into gatekeeper control area.
The following protocols are used to set a connection between endpoints and for media traffic exchange:TCP:
- H.225 - setting up a connection between H.323 devices.
- H.245 - information exchange about new features (for example codecs supported). One endpoint “informs” another about supported features (codecs) and chooses a codec for sending from the possibilities of the other endpoint.
- RAS - is used between endpoints, gateways and gatekeepers. It is responsible for registration, allows calls and statuses.
- RTP - is used for real time transfer of media traffic.
To complete the connection, endpoints send messages to gatekeeper and then the channel closes, and the connection is terminated.
The H.323 standard defines the function of audio data exchange as its main feature (as H.323 has originally been used in VoIP telephony), that is why every endpoint should support at least one codec from G.7XX codec family. Video conferencing was positioned as secondary in respect of H.323 and as a consequence, video codecs support was not mandatory. Today in time of video conferencing and its integration into multiple endpoints, support of video codecs is obligatory. H.323 uses video codecs from H.26X codec family for video coding.
There is a list of technical requirements for audio codecs in H.323, as sound is the key element in VoIP telephony. The requirements are as follows:
- Low level of delay.
- Possibility to restore lost packets.
- High sound quality.
- Narrow bandwidth (no more than 8 kbit/s).
The codecs from G.7XX family meet all these requirements. However, if we talk about the bandwidth, only some of G.7XX codecs meet this requirement.
By default, H.323 standard uses G.711 codec, which has a quite high bandwidth rate – 64kbit/s. Although, not every internet user has the same channel. Moreover, today G.711 is considered as an outdated codec, because its sampling frequency (converting analog signal into digital) is 8kHz, while the more advanced codec G.722.1 has twice as high frequency (16kHz). The low frequency codecs such as G.723 (5.3/6.3kbps) and G.729 (8kbps) are used in internet connection.
As for video codecs, every modern H.323 endpoint should support H.264 encoding, because this video codec is able to deliver the best picture quality.
Comparison of H.323 and SIP protocols
In this article we have compared H.323 and SIP protocols.