Lyn Cantor is president of Tektronix Communications, the vendor of assurance and big data analytics systems that has emerged as the world’s first Telecoms Intelligence Provider. Here, he tells VanillaPlus that amid all the big data hype, CSPs are generating real savings and new revenues from the intelligent application of big data analytics in specific workflows and use cases.
VanillaPlus: Big data analytics is surrounded by lots of hype and debate. What does it mean to Tektronix Communications?
Lyn Cantor: There certainly has been a lot of discussion about big data and analytics ranging from the practical to hype. Our point of view is that these discussions must focus on analytics with a purpose. There is so much data moving around in the network today that most discussions are technical – about databases. We believe these technologies are enablers but we make use of them to bring big data and analytics into the CSP’s workflow. Data is nothing but data until it is brought together in a usable way. Our view of analytics with a purpose describes how we organise data with right technologies into applications that bring specific users value. Fundamentally you have to bring the data from the network but the application of the technology and how it enables the user to access it is different depending on the point of view of the user. Very few people are linking the technology to the practical applications of the user with the possible exception of IBM which, like us, has applied big data to the use cases of specific users.
VP: Can you give an example of a use case in which Tektronix Communications’ approach delivers the analytics with purpose you describe?
LC: One of our value propositions is to help CSPs make or save money with mobile broadband. A lot of that revolves around the subscriber and how they are cared for effectively. A CSP such as AT&T Mobility handles around 2.5 million calls a month from subscribers. About 500,000 are technical in nature and once a call is triggered, agents have to pull data together from six, seven or eight sources to understand what is going on with the user. This is where the use of analytics comes in, especially from the viewpoints we collect. We can take that data reduce the conversation down from 15 minutes to two or three because of the data we have that can isolate the problem and identify where it is. Once the existence of the problem has been identified, the trouble ticket is opened and big data is used all along the journey to solving the issue. That journey may involve escalations to core operations of radio access network engineering functions, all of which will analyse big data to address the problem. This level of depth requires rapid visualisation or rapid analysis from a database perspective. In the core network you’re dealing with really, really large data such as an Oracle ExaData database, for example. That’s a switch to a different database approach where the data is originated differently and built for scale; a deep archive analysis. The GUI and tools are different here to the customer care database. If the problem escalates to the RAN, the database is different again. The value we deliver is that we provide the workflow to make use of the data we collect. We have the ability to look at the four dimensions of the network, technology subscriber and service. We’re able to identify the subscriber when they call and identify what they’re using, then across those four dimensions we can apply big data to the workflow. That’s what CSPs want so they can take waste out of this [care] workflow. This is the thoughtful use of big data to take care of a customer’s problem.
VP: How is the user profile of big data analytics changing within CSPs?
LC: We capture the data with probing and data collection technologies and aggregate it for workflow so the data we collect, especially when we get closer to the radio domain, is very high value to lots of stakeholders. We then become a very important feed into many parts of the CSP business. Others talk about big data from a data warehousing perspective. The dialogue with the CIO is about having an enterprise data warehouse where all the data is taken and put in the warehouse to be sifted for key learnings. We augment that by giving mobility information and that becomes very valuable for billing aspects – such as disputes – but most commonly within marketing organisations – they get the real-time information that can be used for campaign management.
VP: Does that mean the roles and responsibilities that big data analytics supports are changing?
LC: It feeds into a dynamic that is very subtle – the blurring of the CIO and the CTO strategy. The CIO continues to play an evolving and important role in terms of domain responsibility but the data we collect is all around the network and that gives the CTO a really good view of what’s going on in the network. We have a number of use cases here. We’ve seen the CTO’s office drive churn mitigation because the data is so closely connected to the network, for example.
VP: How big is big when it comes to big data?
LC: To give an idea of the scale, we shipped about 17 petabytes of data last year. That’s close to 20% of the overall volume of Facebook traffic and the equivalent to the capacity of every hard drive shipped globally in 1996. Our systems are pretty big. We touch a lot of data and it’s of high value but only when condensed and extracted to provide our customer with value?
VP: How do you see CSPs’ approaches to big data changing as the telecoms value chain involves increased partnership and third parties? Is there a need for standardisation when it comes to big data analytics so insight can be integrated across multiple businesses?
LC: Standards are going to take a while to evolve because there are so many use cases. We’re working with the TM Forum and we think standardisation could be developed in a way that follows a similar model to the Forum’s ETOM. Standardisation of logical workflow would be a good example because we’ll see the use of big data woven into a series of roles. My CTO and product lifecycle management team are very involved in the choice of technologies by the user and identifying which technologies allow you to scale and deliver the best cost advantage. Our work in the TM Forum is really about giving our view about optimal workflows to get the best business results.
VP: You mentioned the cost advantages that big data analytics can deliver. Do you think CSPs have fully understood the benefits they can accrue through effective big data analytics?
LC: It is still early days and we’re just now seeing new technologies in terms of databases that are applicable to web-based search but we are close to adoption of web-based technologies. Over time, the hype will burn away. The situation reminds me very much of the enterprise CRM market. The promise of SAP and Oracle was to redo the workflow and enable companies to become CRM oriented. In transformations of this type, you can never boil the ocean and be effective. You have to take the tools and be very specific about the problems you solve. If it’s just a big land grab, big projects could just create a lot of pain. Use-case specificity will be the answer and those who adopt that will get the greatest benefit of the data flowing through the network. We’re a big data company because we have the data – others just have the tools. We’re applying our data to use cases that help CSPs make or save money.
