Four Bold Telecom Predictions for 2019 and Beyond
Traditionally once a year, I sit down and write one of these ‘prediction blogs’. However, I would not be doing my job if I only thought about the future once a year. Rather, it is an ongoing process, one that the ECI teams and I do on a consistent basis.
A tradition it is, so it’s time again to share with you some of my thoughts about telecoms in 2019. While some may not pan out, I am hoping that they will still prompt a healthy debate and discussion, one that hopefully, will advance the industry in a meaningful way.
Prediction No. 1 – The ‘vendor agnostic’ movement needs to evolve to ‘carrier agnostic’
For years, carriers have been pushing the industry to take an open, ‘vendor agnostic’ approach. Prompted by the SDN movement, this approach was supposed to release them from vendor lock-in and enable Service Providers (SPs) to make business decisions based solely on the technology best suited to the application or service. While this makes sound sense, it was not easy getting proprietary hardware and software systems to seamlessly work together (and if truth be told, we still have more work to do). Funnily enough, while the industry has made great strides on moving this forward, carriers still work in their own, ‘closed garden’ world.
This may have worked in the past, however today’s OTTs have shown consumers a ‘better world’, one in which they can offer global services -- no bars -- across multiple providers regardless of infrastructure, cost structure or politics. Moreover, in this age of technical and business disruption, the web-scale giants like Amazon, Apple, Facebook and Google have gained a strong foothold in providing services customers want and use, leaving the carriers sitting on the wrong end of a losing battle.
Carriers need to move to the other side, and quickly. Inter-carrier cooperation is just as complicated and time consuming today as it was 20 years ago. It is tough, even internally for larger carriers, often made up of an amalgamation of smaller carriers, networks and equipment from years of M&A. This makes setting up services spanning multiple carriers a very…. slow…. moving…. process. The walled garden is as limiting for SPs as it is for their customers. SPs in search of new revenue streams don’t need to look too far, they only need to open up their gardens. Imagine, for example, extending VPNs beyond single SP boundaries, or extending other services via APIs.
Beyond today’s needs, end-to-end network slices of 5G will not accept the ‘boundaried’ world of today’s telecommunication landscape. Autonomous cars cannot be allowed to crash because they’ve moved to someone else’s edge node. Nor will remote surgery be possible without services that traverse multiple SP networks. So what are SPs doing about it? In spite of the long list of PoCs demonstrated at a recent MEF event, I would argue – not enough!
Prediction No. 2 – Best effort is not ‘good enough’ to ensure ROI on 5G investments. Assured, dynamic services are required
Today, we’re accustomed to meeting SLAs with a ‘best-attempt’ approach, coupled with over-engineering the network with more bandwidth than necessary, to ensure that there is capacity where needed. However, backing up 5G services with a 20th-century approach to service assurance will just not work. Why? Because in the long run, consumers will not pay more for just ‘more and faster’ capacity.
As always, consumers will be willing to pay more for services that provide them a unique value. 5G promises to enable just that. However, almost all 5G use cases will require proactive assurance and dynamicity. Otherwise, ‘real-time’ might hit a time warp, and we will likely experience jitter in our AR goggles. For these services, like many others, today’s static, best-effort assurance will not be enough.
As 5G rollouts progress, expect to see service requests change on several levels. Expect these to become increasingly sophisticated — as they are initiated by a large variety of edge devices, users and applications. The requirements can (and will) change each time a service is used. This is where dynamic, assured connectivity services come into play. A network transition of this magnitude won’t be easy, but without it, 5G will fail to deliver ROI.
Prediction No. 3 – Self-organizing AI-based networks will not materialize with existing infrastructures
The dialogue about the power of AI in telecommunications continues to gain steam as the industry has latched on to yet another hype. However, the cold, stark reality is that we couldn’t be further from this, as current networks suffer from the lack of data capabilities needed for Artificial Intelligence (AI) and Machine Learning (ML) to work. Our static and antiquated equipment and frameworks are not built for the dynamic reconfiguration needed to take advantage of AI’s agility. Likewise, we lack the capabilities on the business layer to unlock the potential of AI. Unfortunately, there is no hope that we will get there by simply ‘sprinkling some magic pixie dust’ or ‘waving a magic wand’.
A network that is truly assured and dynamic, needs to be agile and self-organizing – made up of both a programmable transport network and agile control layer. To achieve this, we’ll need 5 key components: AI, a multi-gear programmable network, big data, ML and human guidance. The result will be a network of the future, ready for 5G services.
Agile, self-organizing networks will exist in the future, but SPs must be vigilant in their digital transformation efforts today to build the foundation for these networks. The near-term focus should be on adding multi-gear programmable networks and putting in place the Big Data systems to report on network performance and events.
Prediction No. 4 – SDN is Dead!
I’ve traditionally been a protagonist for open, programmable networks. I realized that before SDN could excel, many carriers would require a network-wide environmental upgrade. Yet, the benefits of true SDN networks would make the journey worthwhile. However, in my humble opinion, SDN, as it is implemented today is far from delivering on its promise of open, programmable networks. Current implementations use SDN as ‘pumped-up’ network management systems, rather than for true programmability.
So, when the ONF announced the Stratum project, I eagerly read up and engaged in the conversation. Nevertheless, I find myself asking, “Does this new initiative (P4) really stand a better chance of success than its predecessor (OpenFlow)?”
My conclusions led me to believe that the answer is an unequivocal “no”! P4 still doesn’t properly deal with ALL three elements of packet operations: 1) Header processing, 2) Stateful (frame level) processing (e.g. load balancing, OAM), and 3) Separation of control and data plane. Moreover, Stratum, which is heavily leaning on implementation, still does not strike the right balance between standardization and implementation. And last but not least, Stratum lacks the industry consensus and backing required for its success. So alas, perhaps it’s time to say goodbye.
I leave you here, as I continue with our plans for a grand 2019. Rest easy, though, I will be back with another set of bold predictions next year.