|Issue:||Asia-Pacific III 2015|
|Topic:||How Big Data analytics is transforming the communications industry|
|Title:||President & CTO|
Mounir Ladki joined MYCOM in 2001 to start and lead the software business and since then he has led the company’s product strategy, development R&D, go-to-market. Mounir has 18 years international experience in technical and business leadership roles. An industry visionary with expertise in Mobile Telecoms, IT and Software, he is actively participating in shaping the future of OSS and Networks Management, and has been sitting at the advisory board of TM Forum. Mounir was recognized twice by ‘Global Telecoms Business’ as one of the leading 40 under 40 executives in the Telecoms industry worldwide.
Prior to joining MYCOM, Mounir worked at Nortel where he set up and managed the advanced engineering services group and played a key role in the definition and market delivery of Nortel’s advanced mobile solutions including a number of world firsts. During this period he designed solutions for more than 20 mobile operators worldwide in APAC, EMEA and US, and obtained many distinctions.
Mounir gained a master’s degree in Telecom engineering with distinction from Supélec, a business degree from Sorbonne University and executive education at INSEAD.
Big Data analytics is set to revolutionize the future networks based on NFV and 5G, and the nascent world of IoT (Internet of Things). Big Data Analytics is evolving towards ‘real time’, i.e. it will have the ability to process huge amount of data and deliver actionable intelligence within a few milliseconds and less than one millisecond in some extreme cases. It is also coupled with predictive analytics that rely on Artificial Intelligence to learn from the data and detect new patterns. At the same time, these processes are now operating not only in the cloud but also at the edge, i.e. close to where the action needs to take place. Firstly, this will enable analytics-driven virtual networks (NFV) orchestration, where the network configuration will be changed in real time to serve a particular customer experience requirement.
Big Data has been a buzzword for the last few years, and has certainly peaked on Gartner’s hype cycle. Data has been dubbed the new ‘oil’, and the digital society in which we now live coupled with advances in hardware, cloud and computing software, have raised considerable expectations about the endless possibilities offered by Big Data analytics and how it can transform and enrich our lives, while providing an incredible return on investment.
So, first things first, what is big data?
Big Data is widely defined by the three Vs (Volume, Velocity and Variety). At its simplest, Big Data ensues when volumes of data to store and manage exceed hundreds of terabytes; though worth nothing that we are now evolving towards the Petabyte era. The data is usually very diverse and comes in various structured (files, tables, etc.) and unstructured (e.g. video, social media feeds) formats.
Today’s digital society, with the development of social media and smartphones, is generating huge amounts of digital content and various types of unstructured data (as noted above). This data explosion is set to amplify in the coming years; as an example, 90% of the data in the world is now believed to have been generated in the last two years, with 200 million emails sent every minute and 72 hours of video uploaded to YouTube alone.
Given that the price of storage hardware has gone down significantly, the concept of retaining and storing data to identify trends and patterns can lead to successful business outcomes such as better serving a customer or increasing business efficiency. In most cases, the data will be analysed for various purposes other than the primary purpose it was collected for. For example, a CSP can use the billing records to better understand customer preferences and even detect unsatisfied customers who are about to churn. This approach has started to revolutionize many industries from healthcare to retail, energy, transport and insurance.
The key challenges lie in collecting and storing vast amounts of complex and diverse data and having the capability to analyse this data, detect patterns and trends and predict future outcomes.
The development of Big Data has been enabled by advances in three key areas:
• The cost of hardware and storage systems has gone down significantly together with the emergence of open source software platforms for distributed storage and processing of data (Hadoop, Cassandra, etc.)
• The development of broadband communications and cloud computing that allows large centralized systems to process data from remote data centres in near real-time.
• The development of Analytics software and Artificial Intelligence algorithmic to cut through the data and deliver the analytical or predictive output.
Communication Service Providers (CSPs) have always collected large amounts of data related to their customers for billing purposes (call duration, location/mobility, web browsing, content etc.) Previously, this data was erased after bills were produced but now this ‘Big Data’ is stored for further use. In addition, CSPs own valuable information regarding the profile of the customer: social profile, spending, type of handset, type of subscription, address, etc. Now that they have started using and correlating this mass of data to understand the consumption patterns and preferences of their customers, it enables them to build targeted bundles (for example offer heavy Facebook users unlimited access to Facebook on mobile phone against an additional fixed monthly fee). Some CSPs are also packaging and selling this data to third party advertisers who can use it to build targeted advertising and commercial offerings.
Also, mining the above data, and applying predictive analytics on top allows CSPs to detect the customers with changing habits and patterns, such as those who have faced quality issues or have significantly reduced their spending. Crossing this with historical data, the CSP can then detect ‘risky customers’ i.e. those about to terminate their contract, enabling proactive engagement to turn these customers around (e.g. by offering them attractive commercial terms). This will always be a winning proposition for the CSP given that the cost to attract a new customer or to bring back a lost one is 10 times the cost of retaining an existing customer.
Operators own massive data related to their networks, covering areas such as traffic, load, capacity, performance, quality of service, etc. By crossing this information with the above customer data, the operator can become smarter at managing its network, increasing its efficiency and offering a better customer experience now referred to as ‘Network’ or ‘Structured’ Analytics. For example, the operator can analyse the location of its iPhone 6 customers who are subscribed to 4G to decide where to invest in 4G sites as a priority. Alternatively, it can analyse the type of content consumed by customers (video streaming that requires low latency and jitter Vs heavy downloads that require high throughput) to design the network in a way that maximizes the Quality of Experience.
Big Data and future networks
Beyond these innovative approaches, Big Data analytics is set to revolutionize the future networks based on NFV and 5G, and the nascent world of IoT (Internet of Things).
Big Data Analytics is evolving towards ‘real time’, i.e. it will have the ability to process huge amount of data and deliver actionable intelligence within a few milliseconds and less than one millisecond in some extreme cases. It is also coupled with predictive analytics that rely on Artificial Intelligence to learn from the data and detect new patterns. At the same time, these processes are now operating not only in the cloud but also at the edge, i.e. close to where the action needs to take place. Firstly, this will enable analytics-driven virtual networks (NFV) orchestration, where the network configuration will be changed in real time to serve a particular customer experience requirement.
Secondly, this will also enable new revolutionary services such as driverless cars where data will be generated continuously and in real time by both the car and the road infrastructure (traffic lights, traffic signals, police monitors, radars, etc.) This data will be transmitted, processed, and the actions taken (such as activating the breaks of the car) in less than one ms.
Ultimately, real-time Big Data analytics will allow CSPs to dynamically offer in parallel, multiple contextual on-demand networks with targeted Quality of Service to various third parties. For example we might have one real-time ultra-low latency 5G network slice servicing connected cars, while another high throughput network slice will be servicing UHD video broadcasting.
Big Data privacy and security concerns
Despite the enthusiasm around the use of big data in the telecom industry, some real concerns around privacy and security have recently arisen.
Privacy – The key issue resides in the big gap between the regulatory constraints facing CSPs and those facing the Internet content providers (OTT). While CSPs are heavily regulated and can only use anonymized customer data and are not free to sell it to third parties, OTTs are less regulated and have more freedom in making commercial usage of that data. As a work around, we are seeing more examples where the decision on sharing data is placed directly with the customer who can decide to share his data if he sees a return, for example, a comparison of his spending pattern versus that of his neighbours.
However, these concerns do not apply in the area of Network (or structured) analytics, where customer-related Big Data is anonymized and aggregated and then combined with network infrastructure data to enable better Customer Experience and more efficient network planning and operations.
Security – The same issue applies; while big data collected by OTTs can reside anywhere in the cloud (we all accept this when signing up to services such as iCloud), the storage of big data collected by CSPs has to comply with strict national security policies. In many cases, this data is not allowed to be stored out of the country, and certain access to data must be provided to local authorities for security and national defence purposes.
Also, recent large scale hacking cases have shown the many security flaws and vulnerabilities exposing the big data stored in the cloud. This will be a growing concern as big data related to Smart Cities and vital systems such as health, energy and transport that will be stored and managed in the cloud. The networks will be migrated to the cloud, and big data will be used in real time to drive the orchestration of these virtual networks (NFV).
In conclusion, Big Data Analytics is transforming the communications industry, and is opening endless possibilities that will enhance our lives and the productivity of our businesses. It will allow all industries, and in particular the communication industry, to move from a reactive approach to a proactive, then predictive and ultimately a prescriptive one. The result will be far more intelligent, contextual networks that will enable a much greater customer experience and will unleash the potential of the Internet of Things transforming the way we live, work and communicate.