At TELECOM 95, much will be made of a concept called the 'Intelligent Network' (IN) or the 'Advanced Intelligent Network' (AIN). This is not 'intelligence' in the 'artificial' sense - where abstract rules and 'fuzzy' logic are used in place of procedural programming. Far from it: the code which drives the IN must be extremely robust.
IN is the prize at the end of network operators' efforts to modernize their core networks by introducing fully digital components. In North America this process is complete, and in Europe many national telecom networks are at the final stages. Once digitalization is complete, and IN is in place, operators and independent service providers can launch a wealth of new features and service enhancements.
The real triumph has been the speed at which the enabling digital transformation has taken place - the last 10 or 15 years has arguably seen more technological and cultural change in telecoms than the previous 60, and the starting point was often very primitive. Consider the billing mechanism required for a mechanical 'step-by-step' Strowger exchange.
The Strowger was the original automatic exchange; the steam engine of the telecom network. Even now many examples are still in place.
Configurations vary, but to keep track of customer billing across a Strowger switch each subscriber might be assigned a mechanical meter. In these circumstances, vast banks of these meters could then be arranged in a grid. As calls progress, subscribers' meters are stepped forward by electrical pulses generated on their lines. The regularity of the pulses can then be increased at peak times or when trunk calls are being measured.
The simplest way of calculating the bills is to bring in a photographer at the end of the billing period and take a shot of the bank of meters. Blowing up the photograph then makes it possible to calculate the number of pulses generated on each line since the last bill was sent out. Although this is an extreme case, it illustrates the unwieldy nature of the network, when it was made up of different generations of 'dumb' equipment.
Obviously, this sort of arrangement was not only cumbersome and manpower-hungry, but also made it very difficult to introduce sophisticated services. As long as systems like this remain in place, for example, it is almost impossible for an operator to introduce per-second billing.
IN is the single bound by which operators are able to free themselves from the constraints of tailoring services to the lowest common denominators of technology.
In the Intelligent Network (IN) all of the elements in a digital telecommunications network are arranged into a unified, programmable system. The most coherent way to do this is to run a network of computers alongside the core telecoms elements - the switching and transmission equipment - and use this 'system' to control all the activities that take place. The system does not have to know about the intricacies of the network elements but instead instructs them through a standard language.
The digital switching and transmission equipment that telecoms operators have installed is made up of specialized computers designed to perform highly specific functions. These elements are, in effect, like specialized, ultra-reliable peripherals operating on a very grand scale:
Without an IN, however, they operate in pretty much the same distributed fashion as their mechanical predecessors, by signalling to each other across the network, rather than being controlled by a central processor running the equivalent of a computing application program.
In these traditional circumstances, each call operates as a self-contained process which ripples out from the originating telephone, via the local switch, to the next, and so on, until a circuit is arranged. This distributed operation is usually negotiated through a standard signalling system called SS7 (Signalling System Number Seven).
Each time a call is made the caller actually operates the network directly, by dialling a tiny program for it. When the call is terminated the local switch signals the time spent on the line to the billing system - and that, in a standard telephone call, is about all there is to it.
Such an approach worked perfectly well and formed a very robust framework when all the network did was set up circuits, time the calls and release the circuits when one of the terminals went on hook. But as service providers began to develop more complex applications it became clear that a completely intelligent network would offer a whole range of advantages.
This IN architecture brings all of the switching and transmission functions under the control of a distributed computer system (see IN - How it Works, below) so that a high level of sophistication can be built into services right across the network, instead of being applied only to specific parts. IN, for instance, will enable fixed and mobile network services to be integrated - at present mobile cellular services must operate on completely separate networks because of the 'intelligence' required to operate the service.
To place a call to a cellular phone, the mobile network must log the requested number; look up the identity of the target terminal; identify if it is currently 'logged in' to the network and, if so, which cell it occupies. It then has to allocate a frequency or timeslot (if it's a digital system) and signal the handset to ring and connect.
But its intelligent activities don't end there. As the call progresses it must monitor the movement of the cellphone and assign new frequencies or timeslots as its user leaves one cell and enters another.
Ultimately, operators would like to allow 'fixed' users to have the same sort of mobility. If similar intelligence were applied to the standard network, users could, for instance, log themselves on to different fixed and mobile telephones at different times of the day and have all their inbound calls routed to the appropriate one. Just as importantly, they could take themselves 'off-line' by having most incoming calls automatically call-answered - with the system letting through only specified callers such as immediate superiors or close family.
In meeting the above requirements, the IN effectively changes the relationship between network and caller by interposing a structured system. The caller is no longer programming the network directly, but logging a request with a controlling system: basically a network of computers. Having requested a connection, or initiated a function (telling the service to put on call answering, for instance), by dialling a number in the usual way, the system decides what actions should be taken and generates the appropriate signals to all the network elements that need to know about it.
The non-IN telecom network is the ultimate distributed system, and it has served operators well. There is no single point of failure, so its number one functional priority - to remain constantly available - has historically been met reliably.
The development of the IN has a whole range of operational advantages, and its development is generally felt to be a good thing, but it is undoubtedly susceptible to catastrophic failure - not because there is any single piece of hardware waiting to go wrong, but because of the immense size and complexity of the distributed controlling software. And indeed, where IN is already widely deployed, there have been major 'outages' of the telecom networks.
Digitalization made the network elements 'intelligent', but IN will get them to work together as a team. This approach has several advantages.
All high technology is most manageable when developed in distinct, functional modules, with strong, open standards defining their interfaces. In the computer industry this realization resulted in the OSI concept, and the move to object-oriented programming. IN, along with other telecommunications standards, helps to create a similar environment in the telecoms field, by providing functional divisions between classes of equipment. This way, what would otherwise be impossibly complex development tasks can be broken down into distinct subsets. This in turn allows different teams, or indeed different companies, to concentrate on digestible chunks. It's an approach that creates evolutionary stability, since changes can be made to one set of elements without affecting the integrity of the entire system.
This modular environment, along with the development of other telecoms standards, enables operators to 'manage' services responsively. They are able to connect and disconnect customers, change numbers and institute complex billing criteria, all from a central terminal - whereas before they may have had to program individual exchange switches. This is a major benefit to operators, enabling them both to shed staff, reducing costs, and to improve customer service.
IN also helps to create a competitive market in the business of supplying the network infrastructure. The consumers of this technology - in this case the telecoms operators and service providers - therefore avoid being 'locked in' to their suppliers.
It will make the network itself intelligent and capable of offering a whole range of complex services and extra features to its users in a coherent way.
And last, but certainly not least, the IN concept changes the basis on which networks are managed and services are provided.
Of course it would be possible to use the existing switches and multiplexers to run the distributed IN - they are already powerful computers in their own right, and they tend to be very reliable. And until relatively recently, in fact, this was the favoured approach. But through the 1980s minicomputer vendors worked hard at making non-stop computers (by duplicating or triplicating all their hardware elements) and fault-tolerant systems which could meet the needs of critical on-line applications.
As things now stand, these vendors are well-placed to win business with operators to supply the network's controlling platforms: computers are now considered robust enough for public telecoms network applications.
IN will deliver a new generation of services to both ordinary subscribers and businesses. The possibilities are almost limitless. The IN not only supervises the bread-and-butter placing of calls from one telephone to another and collects information for call billing, but it can also introduce other intelligent equipment into the mix and get it to interact in complex ways with the rest of the system.
Voice messaging is an example. Voice mail has been available as a network service for some time, but in many cases it was presented as a separate service, unconnected to others. IN allows the operator to program the network to redirect calls (under the instruction of the user) to a voice processing system, so that callers can leave messages. This application has proved particularly popular with mobile phone users, who like to make themselves unavailable at times, but who don't want to stay out of touch.
Other services can be offered almost as a by-product of functions the network is already carrying out for its own purposes. For example, the network's signalling system is already keeping track of who's calling who, for billing purposes. Calling line identification simply provides some of this information to the user, by displaying the telephone number of the caller. Technologically, this could be extended to include the name of the owner of the terminal. The IN could then take the number and use its customer database to display the individual or company name on the receiving terminal or associated display device.
By turning the network into a programmable system, service providers will be able to endlessly refine the sophistication, and increase or decrease the complexity of their services, as the market develops and customer needs change and develop.
IN will bring in major changes in the structure of the telecommunications industry - it changes the relationship between the network and its operator by standardizing the commands that cause the network to do things. This is true both for simple things, like setting up a connection between two telephones; and for complex things, like 'hunting' for a subscriber across several numbers.
The effect is to create a commercial environment along the lines of the PC industry, where a standardized and so-called 'open' hardware design and operating system have enabled a diversity of applications programming companies to compete for a huge pool of business.
IN could apply a similar dynamic to the provision of telephone services. The idea is to split the industry into two tiers: network operators and service providers. The network operators will own and operate the physical infrastructure across which services will be provided - under their control will be the exchanges, the cabling, and the trunk network. They will be made up of the current incumbent national carriers, plus a wide diversity of new telecoms operators, including cable TV companies and utility companies.
And the service providers will offer the applications which run on the competitive networks. They will set tariffs and sign up customers to their services. In effect, they will buy wholesale capacity from the network operators and retail it to their customers.
Network operators will probably be able to be service providers as well, but under the new model they will have to open their networks and sell capacity to competing service providers under a 're-regulation' framework.
This new model is deemed to have several advantages. On an immediate level it provides an effective framework for price competition, and could be the engine for dramatic telecoms cost reductions. But - perhaps more importantly - it provides the necessary diversity for the industry to test the market for the wealth of possible service combinations that the technology has thrown up, such as packages which combine fixed and cellular connection.
Click Here to Return to Telecommunications Features and Press
Releases