January 14 2019

AI Analytics Could be the Answer to the Telecom Sector’s Current Challenges

AI Analytics Could be the Answer to the Telecom Sector’s Current Challenges

The technological challenges that the arrival of IoT, 5G or network virtualisation represent makes it impossible to keep managing the telecom networks and services in the traditional way. The growing complexity of the network environment and the need to achieve greater efficiencies impose the design and implementation of automated and more intelligent systems. In such a scenario, the application of advanced data analytics tools becomes crucial.

Although in recent years different use cases in which advanced data analytics play an important role have seen the daylight, organisations in the telecommunication industry have encountered two major barriers that have delayed their adoption. The first one is related to the often siloed structure of and lack of access to relevant data. The second, perhaps even more important, being the technical and cultural readiness of the organisation when it comes to carrying out this type of projects. Moreover, doubts about whether data analytics would bring tangible results to short-term business goals have not helped either.

The Telecom Industry is Now More Mature

But little by little, all these barriers have been overcome. Lately, operators have worked with reorganising their data so that they are more readily useful when deploying analytics solutions. During this period of time, new roles have emerged like that of the Chief Data Officer, heading up a new area that is responsible for defining the optimal organisation and use of data. This includes security aspects and access policies that assure a correct data exploitation. In a complementary way, or as an alternative to companies that lack this figure, it may be useful to consult experts in data analytics that also have a proven track record from the telecom industry. At Blue Telecom Consulting, for example, we have been working with projects and developing specific use cases operators’ networks and systems since 2014. One of our main assets is that we have created our own methodology and gathered best practices for advanced data analytics projects, based on the Data Maturity Model.

This model focuses on auditing and ensuring that the data (information, semantics, quality, repositories, accesses, security, history, etc.) needed to implement or solve the customer use case meets the minimum conditions to make it feasible. Another important aspect of this methodology is to ensure that the impacts on the operations and business processes are correctly identified and that their managers are prepared to manage and exploit the results.

Apart from these assessments, the identification of a real problem that needs solving, both from a business and technical point of view, is key to success. Our recommendation is to start out by choosing a small, visible project that is expected to have a high degree of impact on the business and/or operations and has a clearly defined internal sponsor. We like to call this the “quick wins” that will pave the way for larger projects. The internal data structure, processes and objectives of each organisation and project will influence exactly which techniques and tools that are chosen in each case.

Through the projects that Blue Telecom Consulting has executed, we have identified a number of areas that have shown to be especially valuable to operators. Two main blocks stand out: the first one focuses on the area of network operations in real time. We have developed use cases where we with a 6-12 hours window can predict and thus anticipate alarms of likely upcoming network events. Also via certain types of analyses we can identify the root cause behind the issues and are able to strongly reduce the time it takes to solve them.

Another great opportunity is to work with offline data, both in traditional networks and in virtual or hybrid environments. Thus we have developed use cases for network resource planning, benchmarking of technology, solutions or vendors and related analyses regarding the network performance and forensic analysis of complex  problems. The reason why such a wide range of possibilities has opened up in this area is that by applying the appropriate and most advanced tools and methodologies it is now possible to evaluate a very large number of variables simultaneously. This has resulted in a proven capability to efficiently optimise both the planning, configuration and operations of network resources.

Two Use Cases: Assuring Quality and Guaranteeing Service Delivery

Guaranteeing the continuity of service is a constant priority for any operator. The growing complexity of networks requires new techniques to handle the multitude of events that occur in the network at any given time.

We have thus developed a number of different use cases with the application of advanced analytics techniques in network operations. We can provide tools that enable the system administrators to anticipate potential problems in the network in the near future with high accuracy. This again allows the operator to anticipate those incidents whose negative effect can be minimised through preventive maintenance. This way, the organisation can operate in a preventive and not, as it used to, in a merely reactive or corrective way.

Our second block of use cases focuses on identifying events that are already negatively influencing the quality of network service. Often the system administrators find it difficult to discover the source of a major incident via the RCA process – the Root Cause Analysis -, which delays its solution. 

By correlating algorithms of a massive number of properly defined and optimised variables, it is possible to statistically identify the three or four events, among the hundreds or thousands occurring, that are causing a greater negative impact on the KPI in question. This way, attacking the root of the problem directly, the time of incident resolution is reduced considerably, increasing the efficiency and optimising the technicians’ work. 

Towards Autonomous Systems

The two above groups of use cases are clear examples of the transition period we are going through in terms of the practical application of data analytics in telecoms, which situates us between the traditional and future operations that will be defined by the digitisation and subsequent automation of network operations. In the next generation operations, the intervention of a human being for resolving an incident or validating an action will not be required. The very same algorithm that identifies the root cause of an issue will be able to fix the problem. With the technical advances we will be seeing systems ready to solve more and more complex situations in an autonomous way. These systems will be based on huge data repositories, data lakes, with well-structured and quality data that can be accessed in order to create algorithms aided by artificial intelligence tools.

With regards to how this is implemented, a mixed approach will surely be dominating, where the operators’ centralised data analytics departments are complemented with external experts’ skills and know-how. This includes hands-on experience and best practices from other organisations that can contribute to making projects more agile and increase their positive impact. Whatever the maturity of the organisation with regards to the level of innovation and adoption of data analytics is – not to mention the use of artificial intelligence that will soon be widespread -, any project should always respond to specific needs and real problems. Also, if we are to see this fully adopted it must be in line with the way of working of each company. Data analytics with its developments should therefore never be an end in itself, but rather a means with which any organisation can respond to new challenges that it meets whilst developing.

This article was first published in Spanish in Computing.es on November 27, 2018 


Leave a comment

Notify of
Inline Feedbacks
View all comments
Scroll to Top