TELCO NETWORK VIRTUALIZATION – WHAT NEXT?

2020 was certainly an active year in terms of telco virtualization of the core and RAN. There were dozens (if not hundreds) of announcements throughout the 3rd and 4th quarters of the year from telcos announcing trials of virtualization with one of the growing number of vendors and integrators that have emerged in this area. And as this change takes hold within the telco industry we are seeing the emergence of a new supply ecosystem. The traditional vendors dominant role is being threatened by a combination of hyperscalers with their telco cloud and edge offerings, a growing stable of disruptive vendors with their virtualized core and RAN network functions, and SI’s who have invested in growing their expertise and ability to support this changing of the guard.  This has also rippled through to the standards world, where the pre-eminent positions of 3GPP, TMForum, ETSI etc.. are being challenged for relevance by O-RAN, ONAP, Telecom Infrastructure Project (TIP) and others.  This brief article touches on what I view as some of the most notable changes that are underway, and on some of the challenges that still remain!

Lots of high profile announcements across the global telco industry

I tend to look at this from three perspectives: The moves the telcos are making; the shifts in the supply-side; and the geo-political moves.

• In the telco community, the poster child for virtualization is Japan's Rakuten, who launched a fully virtualized, cloud native 4G network in early 2020, and followed this with their 5G launch at the end of September 2020. This has certainly attracted huge attention across the telco world with everyone waiting to see how successful it proves to be.  Another high-profile virtualization prospect that has everyone interested is Dish Networks in USA. They expect to go live in Q3/2021 with their fully virtualized 5G O-RAN compliant, cloud native offering, positioning them as the trailblazer in the USA for this technology. And underneath these high-profile examples are the dozens (if not hundreds) of announcements by brand-name telcos everywhere in the world of their virtualization trials in partnership with hyperscalers and various disruptive and/or traditional vendors. For example, Telefonica Deutschland announced in Q4/2020 that they are planning to shift their core network into the cloud (AWS) in partnership with Ericsson. 
• On the supply side, the traditional vendors have been relatively quiet when it comes to virtualization (with a few notable exceptions), allowing all the running to be done by the disruptive vendors such as Mavenir, Altiostar, Parallels and others. The hyperscalers have also been very active with AWS, Google and Microsoft Azure being particularly visible during 2020 in developing new offerings targeted specifically at telco virtualization. This move is potentially significant, as the power, reach and financial muscle that the hyperscalers can bring to bear, alongside the synergies they can create with platform offerings in analytics, AI etc,, could permanently alter the market. Major players such as IBM/RedHat and HPE have also begun to invest heavily in this market opportunity. Rakuten have also made an interesting virtualization move by teaming up with NEC to offer the Rakuten Communications Platform. This draws on all the experience they have gained in trailblazing their 5G virtualized solution and could help other telcos to accelerate their own journeys to become a cloud-native virtualized 4G/5G operator. This positions Rakuten as not just as a disruptive telco, but more importantly as a disruptive new entrant into the vendor market. And while it’s still early days in Reliance Jio’s rollout of 5G, they have announced that they will be developing their own ‘indigenously developed’ 5G solution. Its not rocket science to think that once developed and proven in the enormous Indian market, that the hugely impressive Jio will take their offering on the road as a disruptive vendor into other markets.
• The geopolitical perspective: I mentioned earlier that the standards landscape has evolved with the growing importance of O-RAN, TIP, etc.. But another important industry move to consider is the formation of the OpenRAN Policy Coalition. This is a US-led, global initiative to advance the adoption of open and interoperable solutions in the RAN. The US has been a laggard over the past decade in telco infrastructure, unable to boast any major indigenous telco infrastructure supplier. With the emergence of virtualization it has the opportunity to press the reset button on this. If successful, this may well shift the balance of power in telco infrastructure supply, back into US-owned hyperscalers and virtualized component suppliers.

Still a Long Way to Go

But despite the hype and excitement surrounding virtualization, there is also a recognition that it has many hurdles to yet overcome before it is an economically viable replacement for traditional networks. In no particular order, these concerns encompass:

• Availability of low-cost white-label radio units. These will be essential if the rollout of virtualization is to make economic sense. The slightly arbitrary target price point that the industry has set for these is $1,000 per radio, and organizations such as TIP are driving hard towards this goal.
• Power consumption compared to traditional network equipment. We already know that 5G is power hungry compared with earlier mobile generations. As virtualization shifts the processing into the cloud there is a concern that this may make it even more power hungry. Exactly how much more power hungry is unclear. The increased dynamic utilization of resources should help reduce power consumption, but the more power-hungry servers compared with FPGA’s could more than cancel this out. Exactly how this balances out could become important both from an OpEx perspective and also from an environmental perspective.
• Engaging with a new ecosystem of suppliers and partners. The one stop shop (or dual vendor approach) that telcos currently have is much maligned but is something that telcos have become comfortable with. Virtualized components will likely come from multiple different players (particularly across different generations of technology) and will create a cultural and technology challenge that the telco will need to adapt to.  Engaging with a new ecosystem is as much about modifying procurement behaviours as it is about learning about how to operate a disparate set of technologies. Only time will tell whether the telco will find this new challenge difficult to overcome.
• Skills & Culture: Closely aligned with the above point, and as I have mentioned many times in my books on transformation, one of the biggest challenges the telco faces in any sort of transformation (equally applicable to this infrastructure transformation) is the corporate culture and the availability of skills to make this happen. Becoming a cloud native telco involves a lot more than changing technology. Unless the telco can develop the skills of the organization to support a radically new way of working, then the shift to vitalization could stall.
• Security: There has been growing chatter about the security challenges presented by virtualized RAN. There is some concern around expanded threat and attack surfaces in the network due to multiple new interfaces (and vendors). Again, it is not clear whether these are significant issues to progress or minor bumps in the road.
• Cost comparisons with traditional networks:  Finally, and ultimately most importantly, are the whole life cost comparisons between deploying, operating and maintaining a traditional 5G network compared with the virtualized equivalent. The accepted thinking is that virtualization will provide significant cost advantages (in both CapEx and OpEx) and additional ‘pay-as-you-grow’ financial flexibility over traditional asset heavy networks. I have yet to see any good independent cost modelling of this and I’m holding off judgement until I have this. In fact, I have a project underway at present to try and develop this model, so if anyone wants to get involved or to contribute some ideas to how best to do this then let me know!

If you have an opinion on these challenges, you might want to click on the attached survey. I’m running this to try and get a feel for what people see as the real blockers, versus what are likely to only be passing issues. If you share your email address on the survey I’ll make sure to share the results of it with you.