DUBLIN, Ireland-The phenomenal growth of cellular services around the world has created problems for operators as their networks become increasingly congested, encouraging companies to develop technologies and systems to increase the volume of traffic that can be carried over existing infrastructure.
Several established infrastructure vendors have been developing systems to increase capacity.
Nokia high-capacity system
Nokia’s latest high-capacity system is designed specifically for GSM networks. The system is designed to increase network capacity, power and efficiency, and the company claims it offers a 10-fold capacity increase over conventional networks, with switching capacity more than doubled to serve 400,000 subscribers and network investment costs significantly reduced.
The system comprises a mobile switching center (MSC), home location register (HLR), base station with integrated access transmission equipment, base station controller, and a range of tools for training services, planning and information management.
“The number of mobile subscribers (worldwide) is expected to reach 300 million by the end of this year, and we anticipate around 1 billion users in 2005,” said Mikko Heikkonen, president of Nokia Telecommunications Network Systems. “We are addressing capacity needs for operators with this new system.”
Nortel frequency hopping
Irish network operator Esat Digifone in early 1998 was the first to use Nortel Networks’ frequency hopping technology.
“We were taking on subscribers faster than we could roll out the network, and because sites were hard to come by, we needed to maximize the capacity of our existing sites,” explained Paul Crain, Esat Digifone’s network technical director. “The Nortel product works by using almost all the frequencies in each cell.”
He estimates Esat Digifone’s network capacity increased roughly five-fold after the carrier implemented the product, which uses one-by-one fractional reuse frequency hopping.
“The technology was particularly suited to our network at the time, particularly since an increase in the number of transceivers in the network was leading to interference issues. It works particularly well with slow-moving traffic, which makes it ideal for just about any city.”
The company rolled it out across 100 sites over a six-week period.
CSIRO frequency assignment
A number of software companies have also identified opportunities created by the massive growth in mobile-phone subscriber levels.
Australian scientific and industrial research organization CSIRO has developed FASE (Frequency Assignment by Stochastic Evolution), a software system that was originally developed with Telstra and has been in use since 1995.
Damian Thomas, technical manager with CSIRO, claims that on Telstra’s analog network in Melbourne, the system increased traffic capacity by 30 percent and on the smaller BCTel analog network in Vancouver, Canada, it increased it 100 percent. CSIRO pitches the increase for digital networks, based on experiences in Australia, at 20 percent.
FASE is an off-line tool used by engineers in two ways: to occasionally re-optimize a whole network or to select the best frequencies for new cells or expansions of existing cells. Carriers can do this on a daily basis to facilitate localized changes.
The system runs on PCs and requires network data, particularly a cell-to-cell signal interference table, that can be prepared from drive survey results. It works by randomly assigning frequencies to cells to meet demand and constraints. It then tests the worth of swapping frequencies one by one for other constraint-obeying frequencies, accepting better assignments and rejecting poor ones.
It always concentrates on those frequencies causing the worst interference in the current assignment. Many hundreds of thousands of swaps are tested, with the solution becoming more refined over time.
For small networks, high-quality solutions are found within minutes. For large networks, it may be necessary to run the program for 10 hours or more.
CSIRO claims, in comparison, that even a rough manual assignment would take many weeks to complete.
“The number of possible ways of assigning the available frequencies to all the cells of a large city network is enormous,” explained Thomas. “No network can test every possible arrangement. Mathematical optimization is used to steer a path through the maze to find good solutions, aiming to minimize signal interference caused by the same frequencies being re-used in cells too close together. At the same time, every cell must have enough frequencies to meet demand, and correct gaps between frequencies in the same and neighboring cells must be ensured.”
MSI optimization tools
U.K.-based MSI (Metapath Software International) has also developed software-based optimization tools, including Planet DMS, which uses relational database technology to manage the flow of essential radio engineering data enterprise-wide; and Planet Maxxer, which monitors the exact performance of the network and enables efficient processes for managing growth while maintaining service levels.
According Mohammed Sheikh, vice president of performance engineering, network operators can use Planet Maxxer to work out the best (and worst) performing cells.
“Operators can accurately pinpoint network congestion and areas where calls are being dropped without having to send out field teams to every site. They can also offer service-level agreements for delivery of guaranteed up-time, with built-in alarms to notify the network of any faults.”
The technology can be implemented on existing networks, with rollout times depending on the size of the network and the number of infrastructure vendors involved. “It will take longer to implement on a network with decentralized monitoring or if a number of different equipment manufacturers are involved and the Maxxer interface needs to be customized to comply with the interfaces on the equipment.”
Implementation time is typically between two weeks and two months.
Robert Jackets, MSI performance engineering services practice manager, claims it is impossible to predict exactly how different optimization tools perform against each other, although he estimates 15 percent is reasonable.
“This saves the network operator from building new sites or at least from building them as soon as they would have to do otherwise,” said Jackets. “The system can also be used to plan future traffic growth. The level of improvement depends on how efficient the network already is and how efficiently it uses the information provided.”
Because the data required for Maxxer is commercially sensitive, networks tend to implement most of the technology themselves, with MSI providing advice on how processes can be improved or how the information can be used more effectively.
