Design-on-Demand is the Key to Success

Virtual Network Functions (VNFs), the virtual components that enable NFV, have to be managed like any other network element, including being included in asset and resource catalogs used to configure and provision services.

The ability to instantiate a VNF in minutes instead of weeks enables flexibility with time-to-market for new services as well as dynamic delivery of capacity on demand. The downside of accelerating service delivery in this way is that a virtualized infrastructure can rapidly grow to a point where it is difficult to contain or manage it.

Service providers are proceeding cautiously as they learn which virtual elements to instantiate where, when and for how long. When creating VNFs, service providers should establish basic VNF service elements at strategic locations. By using design templates that are pre-defined to optimize the use of VNFs, new services and customer instances can be rapidly and seamlessly configured across a hybrid network. Properly designed, the templates and core VNFs will meet most demands and can be pre-integrated with BSS/OSS platforms, which will then become the foundation for an automated fulfillment process.

Automating “Design and Assign”

In the physical world of service fulfillment, a “design and assign” workflow takes a generic infrastructure design and assigns assets to create a customer-specific service instantiation. In order to turn up a new service or customer, service providers first need a service design and then the unique physical elements required to connect customers with services and services with customers can be assigned to close the loop. When providers had to wait days and even weeks for the physical connections to be put in place and/or configured, automation was unnecessary.

But VNFs are changing all that.

As with physical assets, service providers benefit from the optimal placement of virtual assets. Before assets can be assigned, virtual functions that deliver security, network optimization or residential CPE must be intelligently deployed at the proper locations depending on service requirements for security, performance and latency. Programmable CPE or security functions for business customers benefit from being located in the enterprise data center while virtual CPEs for residential customers are best located at a service provider point of presence (POP).

Automated service orchestration and a unified SDN controller capable of managing the entire hybrid infrastructure ensure that connections from the data center to the customer are accurately assigned and delivered.

Virtualization dramatically alters the design and assign workflow, and the variety of virtual and physical elements available to instantiate each customer’s service and the speed at which the assignment occurs makes automation a necessity. From a customer request to add a service to an engineering request to optimize performance, service orchestration must be intelligent and flexible enough to understand where VNFs are, how they are being used and how the services and network are performing end to end.

Service Management Matters

In the context of the complete network infrastructure, VNFs represent only a fraction of the assets assigned to any given service or customer configuration. Yet, those virtual assets sometimes get lost in the service and resource catalogs after new services and customers are provisioned. VNFs pile up and, without proper monitoring and management, service providers can quickly lose track of which assets (both physical and virtual) are deployed, which are available to be retasked, and which assets should be retired.

As service providers potentially add thousands of new virtual elements, unification of network and service orchestration will be required to maintain service quality and ensure a positive customer experience. In concert with network orchestration, service orchestration must be inclusive of the virtual infrastructure. With so many instances and configurations undergoing continuous changes, service providers need to maintain a sense of how the entire service is configured and how it’s performing over time.

Image by Wagdi with Creative Commons license