Closed Loop Control Delivers Lean Ops in Mobile Networks
June 12, 2018 | by Chris Neisinger
The mobile industry is in transition as new network architectures such as NFV and radio access technologies including multi RAT, C-RAN and 5G introduce complex operational requirements. Mobile operators are looking for simplified, cost-efficient operational models. Automation of existing static rules-based processes is not sufficient. In order to realize the true power of advanced architectures, service assurance and network reliability solutions need to evolve as well. These new architectures must be augmented with closed loop control to enable transformed operations and to enhance performance. Advanced analytics functions including anomaly detection, root incident prediction, and prescriptive automation can drive closed loop control in mobile networks to deliver lean operations and improve system availability.
Closed Loop Control
Closed Loop Control is not new to telecoms. Whether it is an active/standby failover of a PGW or an east/west reroute of traffic on a metro SONET ring, telecom operators have been effectively using closed loop control for decades. However, the domain has been restricted to single network elements or to relatively simple deterministic sub-systems. Closed loop control of large complex networks makes operators a bit nervous.
One area of major concern of fully automated closed loop control is that race conditions could trigger excessive recovery actions or a cascade of re-convergence events. A segment or element of the NFV infrastructure can fail over and then cause a ripple effect of elements failing again and again. Advanced analytics based solutions with end-to-end network visibility can alleviate this concern.
Advanced Analytics Provide Context and Identify True Problems
Network functions or network segments rarely transition directly from fully functioning to completely failed. Sluggish performance often precedes element or sub-system failure. Even if a device has sluggish performance, the device may be operating at peak efficiency. The performance impact may actually be due to an adjacent device sending too many malformed packets or some other adjacent failure. The actual root cause cannot be determined by looking only at the domain of the specific device; instead, a wider contextual view is needed. The best way to achieve this expanded contextual view is through advanced network analytics that leverage anomaly detection, commonality analysis, incident clustering, and behavioral analysis.
Consider this analogy. View a small area of a painting with limited data points (analogous to looking at the stats of a single network element). It is quite difficult to get the entire picture from the limited context.
Can you tell that: is actually: ?
The Big Picture
Now take an end-to-end view of the entire painting. The expanded context makes it extremely easy to see and understand the environment.
A similar approach can be applied to service assurance of complex telecommunications networks. An expanded network analytics context that includes data center cloud visibility, service layer visibility and service performance beyond the data center, provides the most efficient and reliable method to trigger corrective actions in a NFV environment.
Closed loop control of complex systems based on advanced analytics is best triggered based on end-to-end performance. Service assurance based on limited visibility within the data center results in less than optimal performance and higher costs. Efficiencies are achieved by having a complete view of the network and making the best decision for the whole network rather than the performance of individual elements. A distributed analytics platform with full visibility within the data center and with extensions into the transport and RAN, allow the network operator to make the right decision, at the right location, at the right time.