 | Issue: | Latin America 2015 |
| Article no.: | 14 | |
| Topic: | Strategic service assurance: The missing piece in the SDN/NFV puzzle | |
| Author: | Anand Gonuguntla | |
| Title: | CEO | |
| Organisation: | Centina Systems | |
| PDF size: | 215KB |
About author
Anand Gonuguntla is the Co-Founder & CEO of Centina Systems. Anand has over 15 years’ experience in the telecom industry.
Prior to his current role, Anand was the Director of Systems and Software Engineering at Xtera. Anand also held management positions in software and program management at Fujitsu where he worked on FLM and FLASHWAVE product lines.
Anand holds a master’s degree in Electrical Engineering from the University of North Dakota and a bachelor’s degree in Electronics and Communications Engineering from Jawaharlal Nehru Technological University, India. He has published in Proceedings of ACM and holds patent in network management.
Article abstract
A comprehensive and adaptable service assurance solution is required to provide holistic service health and performance visualization across the hybridization of legacy and next-generation SDN and NFV networks. Furthermore the dynamic nature of SDN and NFV makes obsolete traditional assurance approaches and requires products taking dynamic new approaches.
Full Article
Software-defined networking (SDN) and network functions virtualization (NFV) are the hottest buzz words in the telecom space today. SDN offers the ability to program the behavior of the network by using well-defined interfaces, separation of control and forwarding functions, centralization of control plane and the ability to automate and simplify application deployment. NFV relocates network functions from dedicated, purpose-built hardware to generic, more interchangeable networking equipment. SDN and NFV work to complement each other by transforming the way networks are managed and used with their ability to maximize utilization of hardware resources.
It seems that not a day goes by without some new statistic, survey or trend being announced that points to the rapid growth in SDN/NFV deployments and the subsequent increase in market value. For instance, the following analyst firms continue to report data backing growing demand for network virtualization almost every day:
• IDC forecasts that the SDN market for enterprise and service providers will grow from $960 million in 2014 to more than US$8 billion in 2018. That number is up considerably from the forecasted US$3.7 billion in revenues the market is expected to generate by next year.
• IHS reports that the NFV market will grow more than five-fold through 2019. Michael Howard, Senior Research Director for carrier networks at IHS, said, “NFV represents operators’ shift from a hardware focus to software focus, and our forecasts show this. We believe NFV software will comprise over 80 percent of the US$11.6 billion total NFV revenue in 2019.”
• MarketsandMarkets predicts that the SDN and NFV market sectors will be worth US$45.13 billion in 2020, a CAGR of 86.1 percent.
• Infonetics Research, a part of IHS, conducted in-depth interviews with telecom carriers around the world to determine their plans for evaluating and deploying NFV and found that 35 percent are planning to deploy NFV in 2015.
Specifically in the Latin America region, IDC has predicted that software-defined everything (SDx) will become a mantra for infrastructure build-out. As Latin America sees increasing
competitiveness of cloud-based companies that are aggressively entering the market, this can also put a strain on traditional IT organizations which need to transform their hardware quickly in order to be able to leverage legacy apps and infrastructure investments with the growing need to become more agile and business service oriented. As a leading indicator of how fast the preference for SDx will impact infrastructure, take a look at investment in automation and orchestration. IDC predicts that investment in automation and orchestration software reaches US$785 million by 2018, growing at a 17.5% CAGR over the 2014-2018 period. System management software spending in 2015 is expected to reach US$737 million growing 11% over 2014.
Benefits of SDN and NFV
When you consider the overwhelming benefits of SDN and NFV, it is no surprise that global operators and service providers are planning to deploy solutions in the near future. One of the promises of SDN and NFV is to lower ongoing Opex and Capex costs through the commoditization of hardware. By envisioning networks that are simpler to configure and control, these architectures help to reduce spares management and in turn will reduce the time it takes to rollout new services. In addition, SDN and NFV networks will deliver improved network performance by centralizing and separating the control logic to make better traffic decisions. These new virtualized networks promise and increase in flexibility and agility for networks to make them more programmable, dynamic and more simplistic architecturally.
Even with all these significant benefits, there are equally significant risks including the challenge of service assurance at a time where it is more necessary than it ever was with traditional networks.
Assurance challenges of the SDN/NFV transition
The current transition to virtual services has been primarily focused on network architecture and orchestration. The next area of focus will need to be on operationalizing the virtual networks and that of assuring the performance and availability of these virtual services. There is more to the transition than deployment, fulfillment and orchestration. Service assurance and performance management is an area of great importance that must be taken into consideration as SDN/NFV services are being deployed. Another key issue to take under consideration is how service providers are going to monitor and assure their networks as legacy and virtual networks co-exist in this new hybrid environment that will be the norm for the foreseeable future.
There are compelling challenges to assuring virtual services that will make the process much more difficult to address than previously with legacy networks. Due to legacy networks being more well-defined and static compared to virtualized networks, these end-to-end services can be defined in assurance systems and monitored for alarms, performance and SLA conformance. This becomes a much bigger challenge for SDN/NFV networks primarily due to the increased complexity and dynamic nature of these new end-to-end services. Virtual service visualization and monitoring must go from Application to logical network to physical network and the composition of the virtual service can change at any time. How will monitoring systems keep up with the changes to the services and also provide consistency to the data as the network evolves, adapts and self-heals?
With the dynamic nature of virtual services, orchestrators and controllers will have the ability to modify, re-route, re-configure virtual services on demand, in real-time. This places a requirement on a Service Assurance system to be able to adapt in real-time to these service changes to be able to accurately reflect the current state and path of both virtual and hybrid services.
One of the biggest challenges to being able to support these network states will be the Service Assurance system’s ability to maintain the consistency of performance metrics for the services as the virtual and physical elements of the service change over time.
The critical role of service assurance
A comprehensive and adaptable service assurance solution is required to provide holistic service health and performance visualization across the hybridization of legacy and next-generation SDN and NFV networks. Furthermore the dynamic nature of SDN and NFV makes obsolete traditional assurance approaches and requires products taking dynamic new approaches. Given the challenges associated with the transition to virtualized networks, some of the important questions a Service Assurance solution will need to answer include:
• How is my network configured right now?
• How is it performing?
• Where are the bottlenecks?
• How to plan for future capacity needs?
• How to guarantee service level agreements?
• How to integrate this with the legacy networking I already have?
To answer these questions and have the dynamic capabilities described, Service Assurance solutions will need to be integrated with OpenStack, OpenDaylight, SDN Controller(s) and Orchestration System(s). The solutions must also have the ability to adapt the virtual service in real-time yet maintain all health and performance historical records as the service evolves. In addition, the Service Assurance solution for virtual services must have the ability to perform:
• Automated service optimization and remediation
• Closed loop feedback to assure service integrity
• Dynamic SLA management to be able to monitor SLAs in real-time even as the service evolves and changes
Also of critical importance will be the Service Assurance solutions analytics capabilities to provide performance roll-ups and trend analysis as well as point-in-time snapshots. Point-in-time snapshots will allow for service providers to be able to look back in time at any particular point to see what the network or service looked like and be able to see the performance of the service at that time. This will be critical for diagnosing, troubleshooting and triaging problems in live, dynamic virtual networks.
In conclusion, Service Assurance is an overlooked, yet critical part of the requirements for a successful virtual service deployment and service providers should start planning for what they will need as they plan for their deployment of SDN/NFV. What service providers need to implement is a solution that is uniquely positioned to address the complex service assurance challenges of SDN and NFV environments. From the initial hybrid nature of the infrastructure as virtualization works side-by-side with legacy networks to the real-time, dynamic imperatives of on-demand network adaptation and healing, a solution that can successfully support the current and evolving assurance needs of virtualized networks will ultimately lead to a successful assurance of hybrid networks.
