Big Data, Quality of Service, and Telecom 2.0
Quality-of-service (QoS) is one of the most paradoxical metrics in the telecommunications industry.
In the abstract, QoS is an innately qualitative measure of subscriber experience with a carrier’s services. On first glance, you’d think that customer experience would be measurable primarily through structured customer surveys and logged subscriber feedback, which, of course, most carriers collect.
But you would be wrong. Telecommunications QoS is also inherently quantitative. QoS is tracked within carrier operational support systems (OSS) by analyzing streams of machine data that are continuously generated by switches, routers and other infrastructure. In addition, QoS-relevant machine data is often extracted from regular engineering tests of landline and mobile connections.
If you’re a carrier, you can optimize subscriber QoS in a more granular fashion if you factor in a wide range of relevant metrics, update them continuously, and use them to drive critical OSS functions, such as service planning, resource provisioning, network optimization and customer engagement. The relevant QoS metrics often include call setup times, call attempts, handoffs, dropped calls, signal strength and interference, response times and packet transmission latencies. In addition, call-detail-record analysis, a staple of carrier operational support systems everywhere, contributes to tracking of trends in call-abandonment rates, which may reflect underlying network congestion, reliability and other QoS issues.
Big data is a key enabler for this vision of continuous cross-service QoS management. The sheer size of QoS-relevant data in storage will continue to climb into the petabytes and beyond. The low latencies in end-to-end data movement and application execution scream for stream computing and complex event processing. And the range of QoS-relevant data sources, ranging from structured to free-form, continue to grow.
Competitive forces will push communications service providers to more extreme scales in the big-data substrate of their OSS platforms. Carriers continue to converge diverse service offerings into larger “Telecom 2.0” bundles with common billing, pricing, infrastructure and customer service. These bundles are in response to customer demands for more seamless integration of landlines, VoIP, wireless, Internet, television, pay-per-view and other online services.
What sort of carrier big-data infrastructure is necessary to support this vision of continuous QoS management? In the world of Telecom 2.0, the carrier’s OSS must leverage the following big-data infrastructure:
- Data warehousing platform(s) for storage, analysis, reporting, dashboarding, query and governance of QoS-relevant data linked to office subscriber systems of record
- Hadoop or NoSQL platform(s) for discovery, extraction, collection, transformation, cleansing, integration and preprocessing of multistructured QoS-relevant data
- Stream computing or complex event processing platform(s) for low-latency consolidation, filtering and correlation of QoS-relevant events
- In-memory data platform(s) for real-time interactive data modeling, visualization and exploration of QoS scenarios
- Graph analytics platform(s) for tracking of QoS-relevant user, component, application, system and network behaviors
- Next best action, decision automation, or recommendation engine platform(s) for executing predictive models, business rules, orchestrations and other business logic needed to respond to QoS-relevant events
- Identity resolution platform(s) to facilitate linking of diverse subscriber identifiers—at the application, device and network levels—in order to correlate QoS metrics across heterogeneous networks
- Archival platform(s) for logging, time-series analysis and regulatory reporting of QoS-relevant historical data
If implemented effectively within the carrier OSS, this big-data infrastructure will enable carriers to maintain continuously optimized service levels. In addition, it will enable carriers to differentiate in the marketplace by providing consolidated QoS reports, dashboards, diagnostic tools and other decision-support front-ends. Ideally, these consolidated QoS metrics should be accessible to subscribers directly through self-service portals or indirectly through carriers’ customer support ecosystems.
In all of these ways, big data can have a substantial impact on the quality of experience for subscribers.
- To find out more about managing big data, join IBM for a free Big Data Event
- Visit the IBM big data for telecommunications web page
- Download the whitepaper "Big data analytics for communications service providers"
- Watch this video on how telecommunications provider Sprint Nextel is using big data to gain valuable insights
- Read about new Accelerators for telecommunications event data analytics and social data analytics