While much of what lies ahead for service providers in the world of SDN and NFV is relatively unknown, one thing is for certain: if providers don’t automate processes, networks will become too complex to manage.
In part two of our three-part series on automation, we’ll examine the relationship between assurance and fulfillment systems and discuss how these components can support automation in a closed-loop system.
NFV is changing how we think about networks
NFV presents a fundamental shift in network operations as it allows for increased agility, higher service availability, and substantial cost savings. However, this technology demands highly-efficient processes that, if not managed properly, can drive up complexity.
With virtualization, service providers have the ability to spin up new virtual network functions in real time and move them throughout the network as conditions change. This results in dynamic topologies that humans would struggle to understand, and current OSS cannot support.
So, how can these next-gen networks be managed in a way that avoids increased complexity? It begins with how service providers approach the creation, design, and delivery of services and ends with how they are assured
One ring to rule them all
One of the principle concepts regarding closed-loop automation, as seen by observing other industries who deploy similar processes, is that there must be a single method to control a system.
As I covered in part one of the series, achieving automation begins at the lowest level of the network by applying a single control loop process to individual domains to achieve lower latency and greater scalability.
Closing the loop on assurance and fulfillment
Although assurance and fulfillment have traditionally been independent processes, this can no longer be if service providers want to achieve lights-out operations.
In order to converge these two systems and establish self-correction in the network, service providers must first have a real-time model of the underlying network and resources. Any changes to resources and topology are tracked, correlated, and turned into intelligence which is shared with other resources. This is an assurance function, and will be used as the foundation for all operations: complete orders or isolate and fix faults.
Once this real-time view of the network is formed, it enables the system to identify any issues in the network, take corrective measures, and figure out the best path to deliver a service. Through this control loop, the service fulfillment process can be completed.
A real life example of this would be to think about what happens when a failure occurs to a device in the network. This single closed-loop method would work to identify that box “X” is experiencing an issue, figure out what services are affected, and would re-route services to carry out fulfillment.
In the end, this process would see assurance become the engine that drives the direction of operations from fulfillment, to auto-healing and auto-scaling, and would reduce workflows and execution times for complex multi-domain services dramatically.
Stay tuned for part 3.
Download the Appledore Research Group’s white paper to learn how increased automation is the only way forward to support the demands of cloud services.