Sunday, December 28, 2014

For NFV and SDN to succeed management of the virtualized infrastructure must be implemented Day 1

Every five years a technology trigger comes along that generates hype, confusion, and inflated expectations that fade as the realities of commercial success and operationalizing the new technology settle in. This was the case with IMS 5 years ago. Today NFV is cresting the wave of inflated expectations. Our industry is ripe with innovation but when it comes to deploying it with carrier grade scale and five nines of reliability, the reality is that the implementers need time to prove it out.

I often hear comparisons to OTT and internet companies and how Telcos need to become more like these companies. This sounds good on the surface but CSPs businesses are not the same businesses as OTT, social media, or internet companies. The networks are more complex, the regulatory rules more onerous, and some of the legacy services remain cash cows despite the cries that voice is free. Managing voice services brings with it a complex web of home grown tools, islands of databases, customized COTS solutions, and inter-connection systems. CSPs will not move at the speed of OTT and social media companies because of the legacy services anchored in their traditional business, the mandates enforced by regulators, and the culture instilled in companies over the past century.

Instead what I expect to see is a gradual shift in business models as technology enablers like NFV and the move towards cloud based services permeate in the CSP business. To succeed in this journey CSPs need to revise their thinking on the deployment of new technologies such as NFV and SDN. For virtualization and control plane separation to move from PoC to commercial deployment, management of the technology must be implemented day 1. This is a departure from the thinking of the past where new technology infrastructure deployments led management systems by six months or even a year in some cases.

A significant barrier to more wide scale deployment of NFV and SDN is rooted in managing the virtualized infrastructure. At the center of the debate is the work that has originated out of the ETSI NFV MANO working group which acknowledges the OSS and BSS components but documents them as place holders in the overall blueprint.

A key challenge still under development is how the NFV orchestration process will bind with the assurance processes – also known as “management” under the ETSI nomenclature. The ETSI MANO document provides a useful blueprint only. The work to date is concentrated on the orchestration layer which takes an order, de-composes the service components, and coordinates the activation process across both the virtual and physical components. Service chaining is an important process in this orchestration activity.  A chasm still exist in linking together the orchestration process with the network delivery and end to end service assurance processes. This chasm can be bridged in the service management process which comprises elements of infrastructure performance, component failure, and other functions which assure end to end availability and service impact.


More work is necessary in defining how the legacy OSS and BSS systems will interface with the NFV orchestration layer. The TMF has embarked on bringing clarity to this area with ZOOM. TM Forum’s Zero-time Orchestration, Operations and Management project defines an architecture that that is based on the seamless interaction between physical and virtual components. ZOOM provides a framework to help service providers manage the entire lifecycle of both virtual networks and services. I will be looking beyond the catalyst demos to identify where CSPs are gaining real business benefits in the NFV orchestration and management domain over the next year.



No comments:

Post a Comment