The hybrid NFV service assurance challenge
This blog was written by Naor Mark, Product Marketing director at TEOCO and originally appeared on Vanilla+
The global operator community continues to identify new use cases and wider benefits associated with virtualized network architectures.
The move from physical network point solutions to agile, flexible and cost effective virtual systems is widely acknowledged as a strategic and commercial imperative. Overhauling network architectures will take time however, as operators are forced to overcome a few obstacles moving away from legacy systems – not least ensuring their customers do not suffer degradation in QoE.
Hybrid NFV – the short and medium term view
Industry analysts Heavy Reading, in its 2015 NFV Strategies Survey revealed that just over a third of operators expect to have more than half of network traffic running on NFV by 2020 – this is steady progress given the undoubted benefits of NFV migration. The same analyst, in a separate survey (Heavy Reading’s 2015 MANO and OSS survey) strongly suggests that the inability of service assurance systems to keep pace with NFV deployments is a big inhibitor of operator confidence, says Naor Mark, Product Marketing director, TEOCO.
The rate of transition between physical and virtualized networks will force operators to carefully manage ‘hybrid’ networks for the foreseeable future. That means that existing service assurance systems will need to have the flexibility to assess service and network performance in both physical and virtual environments.
The key issue here is that, in almost all cases, existing, statically defined service assurance systems are ill equipped to cope with the dynamic conditions of virtualization. This calls for a complete service assurance re-think to take advantage of new data analytics techniques and enable true real-time management of the network.
Service assurance for hybrid NFV environments: automation, analytics, integration
While different vendors across the industry will all have their own approach to solving the service assurance challenge for hybrid networks, there is a common acceptance that successful solutions will address three key areas: greater automation, enhanced analytics and better integration with other management systems. Overcoming challenges here, will deliver a service assurance solution that is truly fit for purpose.
When it comes to automation, in a perfect world, the service assurance system collects and analyses network data, but also pinpoints network service issues and makes recommendations of how they might be resolved. These recommendations are passed on to the NFV orchestrator, or if occurring in the legacy physical side of the network, to the service fulfilment system. This closed loop automation thereby reduces the time needed for network repair and service resolution and has the potential to create fully self-healing networks in the short-term.
In terms of improving analytics, service assurance solutions for hybrid networks must move beyond hand written rules for establishing root-cause analysis. Instead, algorithms or even artificial intelligence and machine learning must be introduced to parse a greater and broader data set.
By combining service and resource management into one unified system, operators can correlate across data sets and aid root cause analysis. By identifying and fixing problems at the lowest possible level in the network architecture, operators can save themselves both time and money, reduce time for necessary repairs and minimize network downtime.
Operators must also resist the tendency to treat NFV deployments as a separate silo to its existing network architecture. Service assurance must operate across domains and span both physical and virtual technologies. In physical parts of the network, service assurance must span the delivery of voice and data services. It should also cover all network and IT assets to provide holistic monitoring and cross-layer correlation.
Within the NFV environment, as well as collecting traditional data (such as throughput, drop rates and errors), CPU utilization of the underlying computing and switching hardware should be monitored. The management and orchestration (MANO) will be a key data source for service assurance, as well as its virtualized infrastructure manager and VNF manager sub components.
Within the OSS stack, service assurance should be integrated with policy management systems that define the SLAs that service assurance systems are required to meet. It should also be more closely integrated with service fulfilment to enable operators to rapidly provision and assure services.
Service assurance should not be an afterthought
By meeting the criteria related to automation, analytics and integration, operators should be able to identify and resolve issues that occur in a constantly changing virtualized environment. The resulting and overwhelming benefit to operators is improved network QoS and customer satisfaction.
Highly automated service assurance is key to enabling on-demand, self-service offerings. Service assurance, closely connected with service fulfilment enables operators to provision and assure services in minutes, not hours or days. After all, this heightened service agility and flexibility is the key driver behind NFV deployment.
Too often, operators have deployed service assurance as an afterthought. Other systems like order management and billing have taken precedent as they provide a more direct means of service monetization. Nowadays, there is as much, if not more focus on optimizing QoE for services. This is reflected in a Heavy Reading survey that revealed more than half (57%) of operators now feel a fully operational service assurance system should be in place prior to NFV deployment. Operators know they must re-engineer their service assurance systems to ready themselves for a hybrid NFV world.