Telecom IQ
cancel
Showing results for 
Search instead for 
Did you mean: 

The Journey to NFV

Telco_Editor

NFV is a journey.  For network operators, it’s a journey from a network of boxes to a virtual box of networks.  In business model terms, it’s a journey from transport/connection services to services at a higher level.  For operations personnel it’s a journey from operations systems rooted in manual provisioning to systems optimized for event-driven service automation.  It’s a journey to new APIs, new ecosystems – influencing how and from whom  -  CSPs purchase products and services. 

 

Most of all, it’s a journey to the cloud.

 

The cloud is a pool of resources that can be molded by software to serve any service mission.  Combined with virtual networking and SDN where any connection or service can be defined by software, the cloud creates the first fully agile framework for provider infrastructure the industry has ever had.  Any connection set can build a service, any feature can build customer value and provider profit.  Everything can be tried quickly, adapted quickly, deployed at any scale.  Network Functions Virtualization, NFV, is an architecture created under the mandate of ETSI that embraces the full potential of both the cloud and SDN to free operators from silos based on vendor, technology, or protocols.

 

NFV can ultimately do all of that, but NFV is a journey and we’ve only just begun.  Network operators worldwide have a trillion-dollar infrastructure on which they, and their customers, depend.  They have evolved operations practices honed over thousands of man-years, and marketing practices and even brands that reflect the present more than the future.  All of this will limit just how revolutionary NFV can be, how quickly it can bring its benefits to realization.  It can’t be allowed to stall the NFV journey, and so we must bring all of this legacy infrastructure along with us.

 

HP’s OpenNFV support of this journey starts explicitly with all we have in networks today and evolves infrastructure, services, and operations in parallel to that future of the cloud.  In the real world, that’s the only way that change can come about.  Starting where we are, starting with what we have, we’re taking you to the cloud.

 

The Problem and the Opportunity

 

Operators have relied for over a century on transport/connection services for their revenues.  Today, consumer broadband and the Internet have created an all-you-can-eat world where additional network use doesn’t generate additional provider revenue.  A steady decline in revenue per bit has resulted, and even advances in network equipment and technology have been unable to push down costs at the same rate.  Figure 1 is a composite view of network operator revenue/cost-per-bit relationships globally.  By 2017/2018 operators expect the revenue/cost curve to cross over, making further infrastructure investment problematic.

 

 

Figure 1

CSP_cost_curve.png

You can already see the impact of this historic convergence of revenue and cost in the pressure on operators’ capital budgets.  For years now, financial analysts have been talking about “secular pressures on capex” and “deferred spending plans”.  Now, some are realizing that there’s more going on here than pushing back a few projects.  If revenue per bit continues to fall, the ROI on infrastructure will eventually slip below the CFOs’ guidelines.  Operators might stop investing in their home territories, focusing instead on emerging markets where returns are higher.  The Internet could be strangled by lack of capacity and the future of networking could be one of continued commoditization.

 

Nobody expects that to happen, of course; operators will take measures to correct the problem.  There’s no shortage of suggestions there too.  Software Defined Networking (SDN) promises to replace switches and routers with cheaper “white boxes” to reduce capex.  Network Functions Virtualization (NFV) proposes to substitute hosted software features for dedicated, expensive, network appliances.  Networks should be made more operationally efficient, some say.  Others think that they should be faster in introducing new services, more agile.  Are any of these approaches likely to succeed?

 

Capex savings are good, but not likely decisive in themselves.  Most operators think that all of the capex reduction strategies combined would likely produce between 15% and 25% savings, but because of the long installed life of network equipment it could take years to realize these benefits fully.  That leaves operators with the problem of driving change when compensating benefits would be limited in the near term.

 

Operational efficiency is the strongest of all cost-management suggestions, since most operators report that about half their network equipment TCO is operations costs and the other half capital cost.  Cutting opex by 50% could allow operators to increase capex by the same amount without changing their TCO.  If we could create provable operations efficiencies, we could certainly drive the cost curve much lower, creating headroom for continued investment.  The problem is that while technologies like SDN and NFV promise to reduce opex, it’s not yet clear even how these technologies are operationalized.  Assigning specific savings to one is even more difficult, yet that’s what a CFO would have to do to justify increasing the capital spending side of TCO.

 

Service agility seems to offer the most benefit of all.  It’s popular to point to the OTTs—to Google or Netflix or Facebook—and suggest that if operators were a bit lighter on their feet they could have lead in these markets.  There’s some truth in that; operators are efficient at making infrastructure decisions that preserve the value of long-lived equipment and so they’re less familiar with the pace of change that software-driven services can bring.  However, most OTT successes are funded by advertising, and the total global spending on advertising in 2015 is likely less than $600 billion.  In comparison, the estimated telecom industry spending in the same year is almost $5.5 trillion.  Even if all the ad spending was converted to online services without discounts or losses of any sort, the result would barely pay the capital budgets of network operators globally.

 

The good news for operators is that demand trends like mobile services and IoT have generated new opportunities and content delivery has begun to transition the Internet from a pure ad-driven model to a paid-service model.  In the future, users will move through a virtual web of knowledge and processing, created by a combination of their own location, their social frame of reference, and sensors around them. The “services” they’ll value will be a kind of “decision fabric” that can harness all this power, organize all of this information, to give them not “knowledge” but answers, actionable intelligence.

 

While “transforming operator business models” is a popular goal for vendors, the fact is that operators don’t need a new business model.  They are already selling directly to virtually every economically empowered buyer on the planet.  Operators know how to build services from expensive, long-lived, infrastructure and make a successful business from it.  What they need is new infrastructure—a path to take from building networks to move bits to building services hosted in the cloud.  The trick is making this new infrastructure capable and agile without making it impossible to operationalize, and then getting to the goal at a pace in terms of cost and risk that is acceptable to operators’ CFOs.

 

We have to build the network of the future by gracefully enhancing today’s network and coupling it tightly with the cloud.  We must limit the investment needed and the risks confronted so that they can be matched by early benefits.  We must also recognize that the future will be radically different from the present, and that the biggest area of change will be in the relationship between connection services and processing/information services.  If operators can harness this future trend without excessive risk along the way, the problem of network profitability can be solved, forever.  We believe they can do just that.

 

0 Kudos
About the Author

Telco_Editor

Labels
Events
See posts for
dates/locations
HPE at 2018 Technology Events
Learn about the technology events where Hewlett Packard Enterprise will have a presence in 2018.
Read more
October 9 - 10, 2018
Online
Expert Days - 2018
Visit this forum and get the schedules for online HPE Expert Days where you can talk to HPE product experts, R&D and support team members and get answ...
Read more
View all