In the future, the telecom industry may look back on 2016 as the year that software-defined networking (SDN) and network functions virtualization (NFV) really began to take hold. It could also be looked back on as the year we recognized the performance management and service assurance challenges that can arise from taking this next generational step in networking.
SDN and NFV and the open, bare metal networks on which these technologies run offer many benefits for operators and their customers. In tandem, SDN and NFV introduce greater simplicity, flexibility, and efficiency in how network components can be deployed, and how applications can be created and configured. Single-vendor networks full of expensive and monolithic hardware can give way to multi-vendor architectures using open, interoperable and modular appliances. Service creation can become more dynamic, more customized, and fully-formed solutions delivered to customers more quickly than in legacy environments.
There is every reason in this new environment that managing network performance, avoiding network downtime, and assuring service quality should become easier, too. After all, having a separate, centralized control function means that better, quicker decisions can be made about how to manage networks and applications in real time. Faults and other network problems arising from rapid increases in traffic, or event-related spikes should become easier to deal with in a more dynamic fashion. That should translate to better network performance and improved ability to avoid downtime.
However, the entire management story is not so simply put. For all the benefits to be realized from SDN and NFV in open network environments, this movement to a new generation of networking also can introduce some new challenges where network and service management are concerned.
A New Management View
In the past, managing networks to avoid downtime was mostly about having fairly comprehensive visibility of various network elements and other hardware devices, and keeping an eye on traffic flows moving through them. Network managers could see trends and start planning responses to them, and in the event of problems, they could find which devices were affected and where the source of the problem was. Of course, in some cases, even with that visibility, downtime could not be avoided because operators did not have the capability to dynamically respond to these problems, reroute traffic or repair network hardware.
Virtualization, and more specifically the dynamic nature of virtualized networks and services, can give operators more dynamic control over network and application management, but the problems in the network can be harder to pinpoint. In the age of SDN and NFV, many network elements are in effect becoming software or sets of control functions occurring at another layer.
We can’t simply ask the network management platforms we have been using for legacy infrastructure to now provide comprehensive visibility into this new environment, or to see what’s happening at the level of a virtual router function, for example.
Near-Term Hybrid Models
The challenge can become even more pronounced in the sort of hybrid virtualized networks that many operators are expected to have in place for at least the next few years. How do you manage network performance and service quality while having one eye trained on your legacy gear, and the other trained on your emerging virtual environment?
As operators migrate into these hybrid environments, and later, more fully realized virtual network environments, they will need management systems and service assurance solutions that can work in both worlds. What they really need is an assurance solution that can discover and provide visualization of the entire service chain along with both virtual and physical network functions to be able to determine the impact and root cause in real-time. That same system must be able to communicate back to an orchestration system to automatically optimize or remediate the issue.
Standards for an Open World
In addition to the practical challenges presented by these new network environments, a wide range of industry standards have emerged to guide operators and their suppliers into the more open, virtualized networking era. These include specifications from OpenDaylight, OpenStack, ONOS and others, aimed at guiding and defining the controller, orchestration system, and other operational support system (OSS) aspects and capabilities. In the coming years, there could be some shake-out among these guidelines, but right now they are default standards with which all management and service assurance solutions must comply.
There is so much excitement around SDN, NFV and open networks and the steps operators must take to adopt these technologies. The potential effects this migration may have on network performance and service quality has been a bit overshadowed by service orchestration, but it’s time to acknowledge these challenges and make sure operators are prepared for them.