NFV is here to stay. Ignoring it won’t make it go away

1 of 8: Now is the time for telcos to read-in and skill-up on the transition to NFV

As many of you know, it’s been my honour and privilege to collaborate with the fine folks at Ericsson in various forms, blogs, articles, podcasts, videos, events etc.

You may also know, just maybe, that I am a HUGE fan of Fifth Generation mobile ( 5G ).

So much so I’d love to see a 5G icon on my phone tomorrow if Santa could make it my Christmas wish. Regardless, no matter how you look at it, 5G and NFV are coming, some might say they are already here.

So consider me biased in my efforts here – Ericsson have put in so much work to complete their own digital transformation, transitioning to Software Defined Infrastructure ( SDI ) and Software Defined Network ( SDN ) capabilities built on an OpenStack cloud platform, and mapping out what it looks like for the rest of us/industry, that we’d be fools not to learn from them (err, was that my out-loud voice?). I often say that Ericsson are their own first customer in this space.

A key foundational component underpinning SDN is that of Network Function Virtualisation ( NFV ), which in it’s own right is a heady topic to come to grips with. To address this, Ericsson have recently published a series of 8 ebooks on the topic of NFV transition.

This is the 1st in what will be an eight part series of blogs where I’ll review and share key highlights from those ebooks here, and I heartily encourage you to download your own copies and dive in yourself at ericsson.com/nfv.

I’m not going to belabour the point of “why should we”, or “do we have to.” The fact is “yes, we do”. To “do” 5G, telco operators and carriers *must* move from physical network functions ( PNF ) to virtual. The scope is daunting and the transition will not be easy ( Ericsson estimates it’ll take on average 2 to 3 years. *gulp*. ) The good news though, is that both the business and technology benefits have been validated and confirmed by early adopters, and the technology is performing well.

Ericsson quotes the telecom research analysts Heavy Reading with their list of the top 8 technical challenges, from security to power consumption, but Ericsson’s own list of challenges-to-anticipate looks different:

1. Lack of quantifiable business case
2. Multiple vendors and lack of accountability for breakdowns
3. Immature technology
4. Telco culture not ready for cloud
5. Components and source code not ready for plug-and-play interoperability
6. Lifecycle management requirements (tools, processes, skills) are intense and more multidisciplinary than existing telco workforce’s highly-specialised skill sets
7. Meeting & maintaining fault detection and uptime requirements in a cloud infrastructure
8. It’s complicated. Hardware-defined networks are very hard to mix w/ virtual network functions, making staged migration difficult and resource intensive

To paraphrase Sun Tzu, it’s best to know thy enemy ( and some might say to keep a packet of Nurofen or Advil handy ). But fear not, have faith, we can do this, especially with the help of the team from Ericsson and their partners, and we are already learning from Ericsson’s experiences thus far.

They themselves have made the shift as part of their own Digital Transformation, and have subsequently helped several other operators through their own pivot to virtual.

Like all journeys, it starts with a first step. I recommend everyone’s first step be to map out basic high level requirements – i.e what does “good” look like in your world:

• Integration, integration, integration. I will blog more on this. Just make sure that integration capabilities factor hugely in your planning. Pre-integration solutions are out there and may be worth considering.
• Automating and orchestrating all can seem complicated but it is essential. Luckily there’s been good progress made on standards-based management and orchestration architecture management and orchestration ( MANO ), which means you can expect and require interoperability between vendors.
• Don’t skimp on security. As if we needed to be told. Do your due diligence on the new security management tools, make sure they’re comprehensive across the entire stack.
• Be prepared to do this from scratch. Updating existing code to cloud has been tried and is an out-and-out fail. Skill-up for cloud-native development. New or refactored code may be required in some areas.
• Similarly, take the time to get your head around a software dominant infrastructure. The physical environment *goes away*. That’s massive. But, with the speed at which cloud moves, and the scale of new growth and consumer demand, automated software is critical.
• Make sure you plan to move your Business Support Systems ( BSS ) and Operational Support Systems ( OSS ) and tools over as well, so that you get the business rewards from all this.
• Performance optimisation strategies are different on cloud. Prep for technologies such as SR-IOV, vSwitch optimisations and cloud-native design principles.
• Open Source is at the heart of these new systems and tools, where possible leverage open source for all the obvious reasons, and be wary of proprietary closed systems, integration of open source and open API’s is likely to be far less challenging. Require clearly defined capabilities, APIs and a transparent roadmap.
• Put on your thicker skin, you are going have to make changes. Prepare to be Agile, get used to the Fail & Fail fast ethos, this is the world we live in now – Cloud & NFV move faster than anything you’ve seen before, and current or legacy telco service, business and development environments are all going to have to adapt.

Have I mentioned you need to do this, now? Waiting is likely to be deadly. Anyone arriving late to the NFV party is bound to simply be left behind. It’s very much a case of adapt or be eaten. There is no doubt though, that investment to transition is worth it, the benefits to NFV are massive

Benefits of NFV

• CapEx lowered by up to 30% and OpEx by 25%
• Mobile private network on-boarding time reduced from 21 days to 1 minute
• Global enterprise network on-boarding reduced from months to days
• Time to add NFV capacity reduced from 60 days to 2 hours
• With automation, troubleshooting time reduced from 60 to 3 minutes

I remember discussing many of these benefits with Mats Karlsson, Head of Solution area OSS at Ericsson last year when I was visiting the Ericsson Digital Studio in their head office campus in Kista, Sweden, as part of the first of my Transmissions from the Future themed podcast shows. Matts was actually my first Ericsson guest on the show. At the time he was Head of R&D and Portfolio, and he referred to the dramatic shift from eight months to stand up an environment to a mere eight minutes ( you can tune into that podcast show here ).

Ericsson points to their own ( and partner ) rollouts to validate those percentages. You know your hourly costs of operation. Simple math.

As I mention there’s a complete series of eight ebooks from Ericsson, which break the transition down step-by-step. I strongly recommend you grab all eight of them, and consume them at your own pace, in bite-sized chunks to avoid choking – this is a long-term strategic play, you’ll need your strength. Grab your copy of all eight free NFV eBooks from Ericsson here => ericsson.com/nfv

That’s it for part one in this eight part series, thanks for reading, and I’ll see you in part two.