Verizon is tapping into advanced analytics and its ThingSpace Internet of Things (IoT) platform to craft applications that are unencumbered by the various technology environments that form IoT habitats.
In this Q&A, Telco Transformation spoke to Faraz Shafiq, an associate managing director leading the global IoT practice of Verizon Enterprise Solutions. Shafiq is responsible for Verizon's long-term IoT portfolio strategy, analytics, customer experience and associated professional services. He supports the product roadmap and go-to-market initiatives for Verizon Enterprise Solutions.
Open source technologies are key to distributed applications like analytics based on IoT data. Container-based microservices help adoption across the sprawl of a diversity of geographical, IT, devices and network environments that characterize IOT analytics.
Telco Transformation: The early adopters of Verizon's IoT platform and analytics have been in the agriculture and smart cities industries. 725115 Has the practice expanded into other sectors and how does application enablement explain the wider adoption?
Faraz Shafiq: Our IoT practice has expanded beyond agriculture and smart cities into solutions for medium to large enterprises in broad verticals from utilities to healthcare. Many clients that we serve are convinced of the potential benefits of IoT and analytics, but they want to better understand what it can do specifically for them. Our teams of dedicated IoT consultants examine the situation of each client and craft a solution tailored to their needs. The solutions focus on what is best for the client and can be achieved by using different platforms, systems and applications. Our ThingSpace IoT platform leverages open APIs and microservices that help to quickly customize solutions and drive IoT adoption. Verizon removes the complexity from IoT deployments through its services and solutions portfolio and broad ecosystem.
TT: How much have the solutions you provided matured from descriptive to predictive and then to prescriptive analytics in the processing of IoT data? How does this impact the services that you provide to your customers?
FS: It is still early for prescriptive analytics. A survey we completed recently with Harvard Business Review showed that 61% of the companies are only doing descriptive analytics or lower for their IoT data. Less than 30% can claim experimentation or adoption of using prescriptive analytics for IoT. Predictive analytics has seen wider adoption because of the readily available open source machine learning tools, data science and AI platforms. For prescriptive analytics to be adopted, customers need to have a relatively high degree of confidence, based on a proven probability of success with prescribed courses of action, before they entrust machines to take automated decisions.
Predictive analytics needs an infrastructure capable of an automated feedback loop using many algorithms and data sources to know what is working and what is not. It does not take a great deal of in-house talent to get started with predictive analytics given the availability of many open-source models. By contrast, prescriptive analytics requires lot more dedicated and skilled talent for clients to understand what will work in their environment.
Regarding the infrastructure, capability requires a shift from only relational databases to Hadoop or other unstructured databases, either on the cloud or on premise. It is important to have the ability to parse unstructured data and correlate it with other sources of time series and relational data. A master data management and governance component is also needed to link these sources of information, effectively use the relationships between them and properly manage the various data sources. Additionally, full potential of prescriptive analytics requires a high degree of automated decision taking, such as responding to network outages without human intervention. Only a handful of enterprises have achieved this level of analytic and infrastructure maturity.
FS: ThingSpace is our organically developed platform for IoT. ThingSpace mainly provides data and connectivity management, security and application enablement. A lot of additional features are on the roadmap. ThingSpace also provides the APIs and micro-services required to develop and manage IoT solutions.
TT: The geographical spread of the network matters when you gather data for IoT. LTE CAT-M1 has been much touted. How are you looking to expand your network to meet the data needs of your clients?
FS: We launched LTE CAT-M1 in December 2016 with nationwide expansion planned for this year. The devices that have the supporting CAT-M1 chipsets, available now from vendors like Altair, Sequans and others, can start using the CAT-M1 network immediately. CAT-M1 offers the same reliable coverage as our LTE network and is specifically designed with IoT in mind. It also has QoS (quality of service) capability that alternative proprietary technologies like LoRA cannot offer. We expect CAT-M1 to drive momentum for IoT device deployments during 2017.
TT: What are the additional use cases that become possible with CAT-M1?
FS: Since CAT-M1 is optimized for IoT traffic, it makes longer battery life possible which is critical for remote applications. Hence the maintenance requirement of devices goes down. Furthermore, the cost of the CAT-M1 chipset is lower which makes it more economical. Finally, since CAT-M1 utilizes licensed spectrum, it can offer QoS and is more attractive for critical applications.
TT: IoT needs a distributed network with diverse technologies that need to be managed. What do you need to do to manage it effectively?
FS: Yes, IoT is an ecosystem play. For service providers, it means effectively using IoT in conjunction with other technologies such as 5G, SDN and NFV. We are actively working on advancing the date for the launch of 5G. SDN and NFV are additional key technologies in the mix. The agility and flexibility to control network and traffic through SDN and NFV will allow us to launch new services rapidly. For a global footprint like ours, SDN and NFV can quickly enable global rollouts of new services. All these technologies will be effectively used for network and data orchestration that will directly benefit IoT deployments.
Related posts:
— Kishore Jethanandani, Contributing Writer, Telco Transformation