Service assurance cases in carrier SDN

There is a lot of interest in carrier SDN and the reasons why telecom operators are migrating to it. The recent focus has been on technical and economic feasibility with some promising trials. Even though much of the business case for SDN/NFV relies on these missing pieces, the topic of service assurance is almost absent from the industry discussion.

[[185303]]

Software-defined networks have many powerful capabilities, the most important of which is their ability to make intelligent service deployment decisions when dynamically providing services to cloud applications. SDN networks can automatically make bandwidth available when needed to meet the needs of specific cloud applications. In unpredictable cloud environments, where network resources are consumed and released on demand and traffic patterns can change in an instant, this intelligent resource allocation is critical.

However, this dynamic and responsiveness to user needs is only half of the story. Dynamic allocation needs to be complemented by assurance capabilities, ensuring that the network and services are healthy at all times. Making automated, corrective changes to the network to ensure network performance is different from simply responding dynamically to service requests. In commercial deployments, operator SDN will have to leverage analytics, key performance indicators (KPIs), and other assurance metrics to ensure overall network and service reliability. While assurance in traditional networks involves time-intensive manual corrective actions, in new virtualized networks, assurance must be made to enable changes to occur quickly with automated decisions.

On-demand cloud services have changed many key characteristics and requirements of the network. This has nothing to do with T1/E1 leased lines. Services are automated so that they can be negotiated, used, torn down, and billed within hours or days. Each of these network service connections has specific requirements for bandwidth, latency, and other metrics. In this rapidly changing environment, the network must take corrective actions guided by policies and data triggers on its own, rather than waiting for operators to assess problems and take manual corrective actions. Just as SDN automation enables intelligent service instantiation, SDN guarantees the intelligent provision of service health and efficiency during service.

As operators move toward commercial deployment of SDN in the pilot phase, service assurance is a key component of vendor solutions. Having assurance capabilities will enable service providers to ensure that their SDN services are operational, that constraints on latency and performance are always met, and that the underlying network infrastructure is always operating at peak efficiency.

How does this work? Technically, the analytics, KPIs, and correlations that drive assurance can be integrated with the SDN controller so that they can drive dynamic changes to the network. For example, triggering actions at the network layer to avoid network congestion. This provides optimal flexibility to redirect traffic, establish new IP/optical paths, and dynamically adjust existing IP/optical paths, all driven by KPIs, analytics, and correlations from the IP and optical layers, as well as the physical and virtual domains. This is often referred to as closed-loop or dynamic assurance.

These dynamic assurance techniques have been successfully implemented using operators. One global operator is currently able to offer dynamic, on-demand IP VPN services that are automatically delivered over an IP/optical network. To ensure the service health of the business, the operator uses dynamic assurance capabilities to check link utilization and adjust the path as needed at the IP or optical layer, in this case adding Ethernet links to a link aggregation group (LAG). In commercial terms, this is their service description to the user.

Another Tier 1 operator in Asia Pacific leverages dynamic network assurance by checking congestion on multiple links/paths leading to and from the data center. When congestion occurs, it dynamically remaps the traffic to secondary links, which ensures link efficiency and user satisfaction. When the congestion is over, it redirects the traffic back to the primary link.

Integrating assurance into carrier SDN deployments is a critical step toward a new dynamic networking paradigm. Without it, carriers will not have the new capabilities of these technologies to provide guaranteed performance to their subscribers. Furthermore, as we move deeper into dynamically provisioning services as near real-time, on-demand transactions, assurance is the most missing link in carrier SDN. It will enable service providers to deliver truly reliable, dynamic, on-demand services for enterprise and cloud operations that grow on demand.