[This post is part of a series of posts I produced for http://www.conker.io , a predictive analytics startup. Conker’s people can now be found at http://www.piratedashboard.com/ ]
Back to business after the big startup news of last week.
Looking further at the applicability of predictive analytics across industry and how it can best be utilised, this week we’re looking at telecommunications, more easily understood as network optimisation.
I won’t put a date on it but over the last two to three years the market leaders of the telecoms industry, Ericsson, Nokia-Siemens and Cisco for example have undertaken considerable research into optimising network performance. This eager research has been fuelled by an exponential growth in wireless data traffic on global networks. The increased growth has until now been tamed by technology that was designed for a pre Big Data time.
The current solution of adding more cells to an already congested network of paco and macro cells is not only lacking innovation, but does nothing to address operational costs; the more data traffic the more labour needed.
An innovative solution has been sought by market leaders to ensure network performance can manage the new levels of data traffic. In a world of heterogeneous networks (HetNets) an intelligent system is required to deal with the mounting traffic and a need to delegate network capacity coverage.
Research and market solutions have unanimously arrived at a common solution, a combination of 3:
- Data mining; find patterns, peak times, regions etc.
- Predictive Analytics; predict network usage
- Recommender Systems; offer best case solution to HetNet issues
When incorporated with intelligent algorithms a solution known on the market as a Self Organising Network (SON) is realised; a network that can self analyse, self diagnose and ultimately recommend a solution to network issues such as capacity during peak times. Combining such techniques, and the the monitoring of related progress is seen as a processed research step for Artificial Intelligence.
If you’re hearing this for the first time then you’re behind.
Nokia Siemens have rolled out their solution, iSON. iSON affords its users (network operators) a portfolio of features to best align strategy; implement new network elements with plug and- play ease; automate main operational tasks, transmission capacity and traffic balancing; and automatically optimize some of the network’s vital quality parameters.
Cisco, looking to optimise their cloud services performance have acquired Intucell this year. Intucell being a network management and optimisation company were acquired for over $400million, no small statement. Intucell’s RAN.20 product will seek to optimise and automate issues relating to; network coverage, QoE, QoS, and lower operational costings.
Other players of the Original Equipment Manufacturer market such as Alcatel-Lucent and Ericsson have also rolled out solutions and/or are partaking in ongoing research to fortify their self organising networks. First mover advantage is gone, but it’s not too late to get smart. If history is anything to go by my money is on predictive analytics and self organising networks eventually becoming part of what we term infrastructure technology (a bread and butter technology, like the electrical supply system, or railroads).
Of larger interest to the analytics world should be the combination of 3; data mining, predictive analytics and recommendation systems. Some may get put off by the notion of “steps” to artificial intelligence, but what cannot be denied is the commonality of data, prediction and recommendation that features in most any industry you can think of. It’s with little wonder (thankfully) that IBM and Dublin City University announced this week the launch of a MSc in Big Data and Smart Cities, an announcement that should see Ireland remain ahead of the curve in its adoption of data analysis intelligence.
[This post was part of a series of posts I prepared for http://www.conker.io ]
#PredictTheFuture
Discussion
No comments yet.