Carrier wholesale divisions are facing a rich bouillabaisse of challenges as they strive to keep multiple stakeholders onside. They have responsibility for delivering an ever improved and consistent customer experience across several geographies, they must maintain a sterling quality of service, ensure maximum reliability and security, and work hard to present a flexible commercial model to a complex and competitive market.
Big data has a potential supporting role in all these areas. The right analytics platform can play a central part in delivering a consolidated view across various elements of a carrier business. It allows for a clear vision of the network that lies at the heart of the show, but also bestows greater control over the carrier’s service portfolio, its operational divisions, its relations with customers and its billing arm. Big data at its best can generate actionable insights that benefit the carrier internally as well as deliver dividends to its service provider and enterprise customers, often as a chargeable service.
Big data as a crystal ball
“Proactive and predictive analytics on networks can enable wholesalers to improve SLA adherence, and act on incidents before they occur, helping to reduce truck rolls and improve the customer experience,” believes Ravi Palepu, head of telco solutions for consulting firm Virtusa-Polaris.
“Big data analytics can also enable correlation analysis on the network, on equipment, on services, on applications. By fusing all these data points, the wholesaler can quickly provide better insights on the business impact in case of an event or incident.”
Big data analytics, says Palepu, can be applied to machine learning, playing a part in identifying the root cause of operational problems with a high level of accuracy: “This can then be used to further automate and self-remediate processes,” he adds. “It can also be used for revenue and leakage analytics to provide a comprehensive view on ROI based on customer contract, billing and network data.”
Some carriers still regard the exploitation of big data as a challenge that they’d be better able to meet, if only they weren’t quite so heavily engaged on other fronts. They are already battling with a complicated competitive landscape and the incursion of over-the-top companies into crucial revenue streams, at the same time as striving to reinvent their networks for a rapidly virtualising world.
Christian Michaud, SVP of strategy, product and business development for the service provider group at Tata Communications, believes that carriers are making a mistake if they are not giving at least some mindshare to big data analytics, and believes that most are probably doing so in some fashion or other.
Doing pioneering work
Tata Communications has been an important pioneer in the field, showing how analytics can work on a global scale and to numerous simultaneous ends.
“The way we first used big data was internally,” he explains. “We looked at how we could use it to improve some of our key operating systems, and see where it fitted in with the improvement of churn management and renewal of contracts. This is in place now, and we’ve already seen gains in areas like this.”
Another important analytics dividend for Tata has been in the field of better management of demand for its extensive multi-continent network through predictive analysis.
“A lot of the network is of course cloud-based these days,” points out Michaud. “It’s not a circuit you put in place for a year, but is based instead around much more flexible usage. Even at the IRU level when you are buying capacity on a subsea cable, now you buy a chunk of capacity over multiple cables which you can flexibly deploy. This necessitates a very good predictive tool to see what you are going to have to manage. All of these areas create great efficiencies.”
With this first phase of internal analytics in place, Tata is now turning to its customer base to look into the possibility of external monetisation: “We see big data as a market opportunity,” says Michaud.
“Can we use our capabilities aligned with the fact that we have so much data running over our network, leveraging that to bring solutions to our customers? This is obviously a different challenge, bringing in issues like data ownership.”
The field of churn management is one where Tata has been highly active for some time on behalf of customers. He says data analytics can deliver major results for these customers, although the very best results are in instances where the customer is depending on Tata for the supply of most its connectivity.
“Our analytics ability works well where we are acting almost as an outsourcer, where there is no other supplier,” he claims.
“We can then look at the complete portfolio and make a value analysis and offer churn predictability. But if we only seeing a part of their traffic, there’s a challenge there.”
PCCW Global is another wholesaler with an extensive footprint and lengthy experience of data analytics.“It’s no longer good enough to look at data that shows you about network performance at the end of the month,”says Carlos DaSilva, VP for business development at PCCW Global. “We need to go beyond that for our customers and give them information on what is really happening on the network, not just tell them it’s doing OK.”The instance of roaming, says DaSilva,means that PCCW Global can look beyond its own network. “Something maybe happening that is not our responsibility,but we need to be sure what it is so we can notify our customers what the issue might be,” he explains.
Patterns of information
“On the IP network side, we are also able to look at security with our threat management system. We look at patterns that might indicate a problem, providing that information immediately to the customer so they can make an immediate decision. I’d compare what we do to filtering emails, so we only give the customer the few that really matter so they don’t have to sift hundreds and hundreds.”
One area where carriers need all the intelligent analytics they can get is in the field of fraud prevention. Every year carriers lose billions of dollars due to criminal activity. Until recently, carriers had no serious means of combating fraud head on in real time, as they had a reliance on historical data analysis. Combining big data and real-time detection enables the necessary historical versus real time correlation, adding context and driving key insights for detecting and most importantly predicting fraudulent patterns.
Having a global fraud footprint
Wholesalers have developed a series of tools that can analyse fraud potential in real time and near-real time, speeding up reactivity once dubious activity is spotted on the network. This is both about offering a better service to customers, and creating new revenue possibilities.
Those carriers with a truly global network can leverage their footprint to counteract the threat of fraud faced by regional carriers,service providers, ISPs and enterprises. Operating without a global fraud prevention partner is akin to sitting back and waiting for the attacks to happen.“Telecoms wholesalers can use big data to create new services in fraud prevention,particularly within the financial services sector,” believes Laurent Michel, CMO for Intersec, a developer of analytics software.
“We all know the story – you are on holiday, enjoying some cocktails, then your card gets declined because you forgot to notify the bank of your travels – cue frustrating and often expensive calls to your provider. Wholesalers can work with banks to help prevent this by providing verification that you are in fact in the Caribbean, so it is likely to be you supping on that piña colada and not a fraudster.”
So how does carrier analysis of big data develop from here? In the direction of increasingly sophisticated data mining that develops and enriches daily, acquiring learning experience as data is collected and after it is process, believes Jay Jayasimha,CEO of Cataleya, a vendor of real-time analytics solutions and wholly owned by Singapore-based carrier Epsilon.
The role of machine learning
“Every day we transform the collected data into new perspectives that clearly show the carrier service performance both in terms of quality, reliability and security,” he says.“As the amount of data increases, the data mining becomes increasingly efficient and translates into network behaviour visibility and anomalies detection such as fraud.”More powerful even than having visibility into the business performance is the ability to predict future behaviours based on past data, he argues: “This is where machine learning plays an important role,” adds Jayasimha.
“Byparsing the results of the data mining,machine learning can transform them into clusters of behaviours and based on real-time data, it can predict what will happen next. Combining meaningful insights and machine learning prediction puts the carriers in full control of their business and their network.”
Looking ahead, we can be sure that big data will remain a critical element of intelligent networking, used to offer guaranteed end-to-end quality of service and quality of experience.Network analytics will be key to gaining new visibility into the network and delivering more complex services like voice over LTE and voice over Wifi, and will be critical to the success of such services as they are rolled out.“The next wave of analytics will be around things like advanced churn management, and further improvements in network efficiency,” says Tata’s Michaud.
“We want to be able to extend better usage of network resources to our customers. We’re not there yet – that’s more the next phase.Some of the next wave of big data benefits will happen when the industry is done with the process of virtualising and software-enabling the network.”
