Traditional capacity planning is mostly about investing. Systems are monitored and collected data is analyzed to identify the impact of trends. Particular attention is given to those situations expected to result in resource exhaustion. Necessary steps are taken to mitigate such conditions to keep service quality on track.

But the business is becoming more dynamic. Emerging services are allowing customers to request and remove features on-demand, some of which are provided over ecosystems involving third-party service providers.
Technology is responding to the situation. Network Function Virtualization (NFV) makes it easier to assign resources on-demand. More importantly, NFV also makes it easier, and faster, to release resources once they are not needed, thus making resources elastic.

NFV makes this possible by normalizing network functions into software so that they can share and maximize the utilization of common hardware. On the other hand, Software Defined Networking (SDN) provides an abstraction of transport facilities to make the network easier to reconfigure in realtime. As a result, SDN improves the programmability of the network to make it more responsive to changes and reconfiguration.

When combined, NFV and SDN are expected to enable unprecedented operational dynamicity, enabling the network to adapt to changing business conditions. This means more frequent and complex capacity planning of the underlying physical resources e.g., data centers, transport equipment and facilities, and logical connectivity.

To improve the profitability of new business models and service ecosystems, resource capacity needs to be handled in a more comprehensive, automated fashion. In this context, capacity planning becomes more than just a process for adding resources. It also needs to become adept at factoring in the release of resources in a timely manner. That brings capacity planning closer to the realm of service management.

This type of service management approach should be structured to bring in what, traditionally, had been passive capacity planning and turn it into dynamic capacity management that pursues equilibrium between investments and savings. Achieving this equilibrium requires the enrichment of service management with analytics, dynamic service and network design, and fast, accurate reconfigurations.

Analytics are needed at the customer, service and resource levels to provide just-in-time intelligence that enables optimal deployments and real-time workload management on an on-going basis. This addresses the lack of operational visibility in dynamic, virtual environments by providing comprehensive analyses of service functions, computing, network, and storage resources across physical and virtual domains, and operational cost elements such as power consumption.

These analyses make the gained visibility actionable by feeding the decision process with highly detailed, dynamic data. Multi-dimensional policies drive these decisions to guarantee that workloads are managed to optimize both systems and financial performance.

Actionable results must then be implemented via a dynamic service and network design platform that automates inventory assignment and provisioning processes, including network rearrangements, while continuously keeping track of the placement and state of both static and dynamic components at the various service and network layers. This includes data centers and the traffic among all the nodes in the network.

This service and network design plat form should collect data from disparate inventory systems and other sources to maintain a dynamic view of the network topology at all times. It should also negotiate with service and NFV orchestrators to determine the best utilization of virtualized resources. Finally, this platform should include discovery and reconciliation functions to recover stranded and lost assets, reveal unused capacity and keep track of how capacity is configured.

Execution takes shape in the form of fast, accurate activation and reconfigurations of services and resources. The challenge is to handle the multiplicity and variability of emerging services without introducing delays from configuration inconsistencies and errors. Emerging services are bringing many different kinds of network resources, systems and entities into the picture. And these are constantly changing as new service features are added, unsuccessful ones modified or removed, and new technologies are introduced.

As a result, activation processes, and the subsequent reconfigurations, need to accommodate frequent changes quickly and with accuracy, thus requiring an automated platform that can consistently apply best practices across the enterprise. These processes should be flexible so that it becomes easier to change and extend them to accommodate evolving services, networks and technologies. In addition, they should also leverage and even improve the accuracy of inventory systems.

The goal is to adopt a comprehensive service management software environment that harmonizes service and cloud management with NFV/ SDN control, under a common end-to-end orchestration scheme that factors in dynamic capacity management throughout the entire lifecycle of resources from planning and design to the provisioning, activation, optimization and decommissioning of resources. This should allow service providers to dynamically manage capacity and therefore offer more stringent SLAs, while realizing savings from having a built-in resource utilization scheme that can optimally morph the network to take advantage of fluid emerging services behavior.