This Industry Viewpoint was authored by Prabhu Ramachandran, Director of WebNMS
Like all businesses, optimizing the customer experience for communication service providers (CSPs) helps make everyone happy. Satisfied customers don’t look elsewhere and keep using the service. Low churn and new user referrals help the provider’s profitability, and it lets them focus their resources on developing new services, not managing upset customers. Performance Management (PM) of services plays a critical role in enabling this optimization. Real-time, direct collection of performance data provides critical insight into the present customer experience, and analysis using big data techniques will help prevent problems. If a problem does occur, trusted PM accelerates troubleshooting and keeps the customer informed of resolution status.
Customer experience is a broad term referring to the composite impression of all the touchpoints between the service provider and its user. It includes several factors that are independent from performance such as ordering, pricing, billing and customer service. However, more and more services rely on networked information — often hosted in the cloud — tightening the relationship between network performance and the overall service experience. This dependency increases the value that CSPs can extract from their PM data.
Before working with the data, the CSP must collect it with sufficient fidelity to accurately represent the conditions their customers are experiencing. As input to business decisions, the data must be comprehensive, trusted, up-to-date and cohesive. Among the many issues in data collection, there are two common problems that often plague CSPs — bottlenecks in collection rate at scale and fragmented, isolated data stores resulting from infrastructure silos.
Successful CSPs tend to operate large networks, much larger than a typical enterprise. To accurately reflect the customer’s experience, performance management must collect real-time data as close to the customer as possible — generally the service endpoint. A large consumer or SMB-client network might have millions of endpoints, distributed across a large geographical region. The proliferation of virtualized endpoints, such as mobile apps and virtual network functions (NFV), is dramatically increasing service scale on top of already large physical infrastructures. Of course, the performance data must be connected not only from a single endpoint, but end-to-end across the entire service.
Retrieving performance data from a diverse network of this scale requires purpose-built PM tools. Solutions that don’t scale to this level, such as enterprise-grade tools, suffer from severe lags in keeping databases up to date. These lags cause problems for engineers trying to troubleshoot the network or customer service representatives helping customers understand the situation. In the best case, solutions should deterministically handle the present service load with ample headroom for future business growth.
Unfortunately, even if all the data has been reliably collected in real time, it is often stored in fragments spread across multiple information silos. These silos have grown within the management systems of CSPs operating multi-vendor and multi-layer networks. Common silos boundaries include proprietary vendor applications, organizational domains, and growth through acquisition. Though this approach often keeps services running adequately, the data fragmentation increases complexity and limits opportunities for proactively improving the customer experience. Fragmentation makes service assurance and troubleshooting unnecessarily challenging and frustrating, diverting resources from offering better services. It also inhibits a comprehensive analysis of the data. Therefore, unifying PM systems to centralize the data is a common goal of many CSPs.
Once the data is collected, a variety of processing helps manage the network. Historically, the most common operations are performance monitoring and reporting, often bundled with fault management as part of a service assurance solution. CSPs monitor compliance with service level agreements (SLAs) promised to their customers and track key performance indicators (KPIs) — custom heuristics that hopefully reflect the service performance impact on the customer experience. Reporting provides insight into service performance for customers and other CSP organizations.
Sophisticated businesses, including CSPs, realize that their collected data has value and that value increases when pooled in a central repository, where unforeseen relationships may be discovered through analysis of the unified data and used to improve business decisions. CSPs expect these big data analytics techniques to help them proactively enhance the customer experience. Though data unification raises the bar for data collection performance, the opportunity for cross-correlation across all the network data and between application performance and customer feedback data helps justify taking on the challenge.
Analyzing real-time performance data versus historical trends may reveal patterns that predict impending performance degradation and help CSPs steer resources to address the issue and prevent customer impact. This assumes the data collection and processing time is quick enough to allow an effective reaction — another reason PM bottlenecks must be eliminated. Along a longer timeline, similar analysis will help steer investment in capacity planning to maintain and optimize the customer experience with efficient use of capital expenditures.
CSPs define KPIs with baseline data that correlates service quality metrics against network performance data. Static KPI definitions can be enhanced with analytical techniques to validate baseline KPIs against actual application performance data and actual customer experience feedback. Dynamically adjusting KPIs according to this analysis will help continuously improve the customer experience by helping CSPs more accurately track the value they provide to customers.
In summary, Performance Management is essential to optimizing the customer experience for CSPs. To be effective, PM must unify data collection across the entire end-to-end, multi-vendor network. Real-time customer service requires reliable performance at scale. Integrating the PM platform with a big data analytics engine provides opportunities to proactively improve the customer experience. Platforms with open interfaces will facilitate this integration. In addition to helping improve the customer experience, PM unification and analytics align very well with the ongoing journey towards SDN-like service orchestration, where increasingly automated operations require highly integrated systems.
With over 14 years of experience delivering service provider software solutions, Prabhu Ramachandran directs WebNMS, the service provider division of Zoho Corporation. Prabhu leads strategic marketing, product management, customer support, partnerships and professional services for WebNMS. Leveraging the technology of the corporation’s flagship WebNMS Framework, Prabhu has expanded the business from its longstanding leadership position in multi-vendor network and element management software into vertical solutions for Carrier Ethernet, MPLS, broadband, LTE and satellite networks. In 2012, Prabhu began driving WebNMS into network orchestration, SDN, NFV and IoT/M2M platforms, all critical enablers for service providers to grow profitable businesses. He holds a Bachelor’s Degree in Electronics and Communication from Madras University, Chennai, India.