 | Issue: | Europe II 2016 |
| Article no.: | 8 | |
| Topic: | The network is sexy again | |
| Author: | Paul Gowans | |
| Title: | Wireless Strategy Director | |
| Organisation: | Viavi Solutions | |
| PDF size: | 188KB |
About author
Paul Gowans, Wireless Strategy Director, Viavi Solutions
Paul Gowans, Wireless Strategy Director at Viavi Solutions, has more than 20 years’ experience in the communications and wireless industries. He currently leads the team responsible for Viavi’s Mobile network enablement portfolio, spanning solutions for the RAN, the edge, backhaul and core. Paul’s expertise in IP, Mobile, LTE, VoIP, IMS and VoLTE comes from developing, supporting, marketing and consulting on major mobile communications projects. Prior to Viavi, Paul was Global Marketing Manager for Agilent Technologies’ Network Systems Division.
He is a seasoned spokesperson on many mobile issues for a variety of international conferences and forums and has written many articles. Paul is a graduate of Edinburgh University in Scotland with a Bachelor of Science in Computer Science and Management Science.
Article abstract
Global network traffic has been doubling every 18 months, and therefore increases by an order of magnitude every five years. The most significant driver of that increase in the past few years has been video. Mobile video traffic now accounts for 55% of total mobile traffic. However, another inflection point is on the horizon, as the Internet of Things (IoT) brings billions of new connected devices online. Whereas video typically entails large streams between a limited number of endpoints, with buffering offsetting latency issues, IoT will involve dozens of devices in a connected home or car that, while not individually requiring significant bandwidth, will need to be always connected with low tolerance for latency. Consider a medical application and how crucial one “bit” of data may be.
Full Article
Remember when all the power was in the network, and devices were generic terminals that merely served as entry points? It didn’t matter what type of network: landline telephones were chunky units that sported a keypad and a receiver, and mobile handsets one-upped them by adding a small screen for telephone numbers and short messages. Then smartphones changed the game, ushering in a period of time when the device became the centre of gravity of the user experience, even making exclusive relationships with network operators a thing of the past.
The smartphone era has resulted in tectonic shifts in information and communications technology, from the consumerization of IT to BYOD to acceptance of the cloud. Perhaps because the smartphone has become so integral to our daily lives, it has become commoditized, and a search for the next big thing—tablets, watches, virtual reality headsets—has yet to bear fruit. Simultaneously, one can almost hear the pendulum swinging back to the network, as virtually every type of network is now undergoing revolutionary changes to address consumers’ appetites for new user experiences. To understand these changes, we need to first appreciate what’s driving them, and then explore the way network operators and their ecosystems are responding.
Drivers
New entrants, changing roles
Up until a few years ago, broadcast and cable television were the primary means for consumption of video. Now, over-the-top services such as Netflix and Amazon Prime are taking over a significant share of that consumption. Traditional telcos are increasingly turning their focus on becoming content providers as well (AT&T DirecTV, Verizon Go90). And conversely information and content providers such as Facebook and Google are stepping in to directly offer network connectivity to increase reach of their services.
(R)evolution in the volume and nature of network traffic
Global network traffic has been doubling every 18 months, and therefore increases by an order of magnitude every five years. The most significant driver of that increase in the past few years has been video. Mobile video traffic now accounts for 55% of total mobile traffic. However, another inflection point is on the horizon, as the Internet of Things (IoT) brings billions of new connected devices online. Whereas video typically entails large streams between a limited number of endpoints, with buffering offsetting latency issues, IoT will involve dozens of devices in a connected home or car that, while not individually requiring significant bandwidth, will need to be always connected with low tolerance for latency. Consider a medical application and how crucial one “bit” of data may be.
Decline in the cost of maturing technology
In 2008, a 100G optical interface cost US$30,000. By 2018, that cost will drop to US$1,000, and will level out at US$500. This technology, which has until recently been confined to core and backbone networks, will become more economically viable for deployment at the metro and aggregation layers of the network.
Responses
Gigabit to the end user
Communications service providers in every sector are racing to deliver 1 Gbps connectivity to end users. This trend crosses wireless (5G, 802.11ac) and wireline (DOCSIS 3.1, FTTH, PON) domains. In their most recent Quarterly Access Report, Dell’Oro Group concluded that the broadband access market—comprising cable, DSL and PON technologies—generated record revenue of over US$12 billion in 2015, with half of that coming from PON. Furthermore according to the same report, new technologies such as DOCSIS 3.1 will add more fuel to the growth in this market in 2016.
Of all access technologies, none has more far-reaching potential than mobile. According to the mobile industry trade group GSMA, there are over 4.7 billion unique mobile subscribers worldwide, implying that nearly two thirds of the world’s population are connected through this technology. Mobile operators worldwide are at different stages of technology evolution but are increasingly converging on a path to the fifth generation (5G). In addition to gigabit speeds and sub-millisecond latency, 5G is bringing with it advancements in the network that inch us closer to always-on, always-fast and always personalized networks. As access speeds cross the gigabit threshold, capacity will need to increase correspondingly at all aggregation layers as well as in the backbone. 100G metro networks are coming into view, and service providers are working with their ecosystems to accelerate 400G in the core.
True convergence
Network operators are fulfilling the promise of voice, data and video traversing a unified network. On the mobile front, voice over LTE (VoLTE) is off the ground and rollouts will continue in 2016, as it becomes the de facto voice service over the legacy voice service. The GSA forecasts VoLTE revenues will reach US$2B by the end of 2016. But VoLTE cannot exist as an island. It needs to evolve to reflect the way people communicate today, which comprises not just voice but also data, messaging social media, video and other multimedia-rich services. This implies that assurance systems must empower more granular and flexible control over performance parameters and thresholds to meet the needs of these different applications, alongside the visibility to react in real-time to unpredictable user behaviors.
The interaction between VoLTE and VoWi-Fi will mature, characterized by soft and seamless handoffs between the access methods. Managing VoLTE end to end—meaning understanding service quality from handset to the radio access network to backhaul to core—will be a key operator goal as they ensure that their services deliver high customer quality of experience. This means deploying sophisticated assurance platforms to know in real time where VoLTE services are performing poorly and where there is a stress in the network.
Virtualization, the cloud and hybridization
In order to both scale the network and make it more flexible to meet more distributed, unpredictable demands, operators are embracing network function virtualization (NFV). NFV can significantly lower network operating costs and increase flexibility and service velocity. Today, industry guidelines are for the most part in place to allow introducing the virtualized functions themselves, but management and orchestration standards for the self-configuration required to truly enable NFV are still in their infancy.
While 2016 will see a significant increase in NFV deployments, these will primarily revolve around semi-automatic configuration – in other words, not the full-blown automation required to realize 100% of NFV’s benefit. The NFV industry is therefore likely to put a great deal of effort into developing guidelines for the management and orchestration side of NFV deployments.
The benefits of NFV will only be realized if network performance management tools can access these new, virtual network interfaces. One major challenge for operators adopting NFV and hybrid networks will be comprehensive, real-time visibility. Networks will ultimately be distributed across a hybrid design consisting of legacy physical, virtual and cloud environments, and it will be essential for operators to be able to have a comprehensive, integrated and real-time view of network performance, states and issue root causes. Operators will need to invest in solutions that ensure they can satisfy quality-of-service needs, including resiliency and latency in initial virtualization deployments. This next year should show a major ramp-up in the availability of test and assurance solutions able to provide truly actionable performance insights for virtualized network environments.
Machine-learning, self-organizing networks
Self-organizing networks are essentially the key to a connected future. By automating configuration, optimization and healing of the network, this frees up operational resources to focus on what’s truly important—better quality of experience and aligning revenue to network optimization. And, with the number of connected “things” positively exploding, managing and keeping up with the sheer number of devices requires an automated approach that also yields a new set of network-assurance challenges operators will have to deal with in 2016.
Today, many SON techniques simply baseline a network. In 2016, as the extreme non-uniformity in the network becomes more apparent, it will take a new, end-to-end approach to SON to keep these benefits coming. In particular, machine learning algorithms will enhance SON by predicting the effects of changes to help ensure stable operation, rather than the prevailing SON paradigm of trial and error.
The network will become more sporadic and this will manifest in several forms: time, subscriber, location and application. For example, take subscriber and location: a recent Viavi Solutions customer study found just 1% of users consume more than half of all data on a network. The study also found 50% of all data is consumed in less than 0.35% of the network area. To achieve significant performance gains via SON, operators can apply predictive approaches using analytics that reveal exactly which users are consuming how much bandwidth – and where they are located. This level of foresight is key to not only unlocking the full potential of SON in the RAN, but also to maximizing ROI for software-defined networking and NFV in the core.
When does all this come together?
Industry hype machines, ever in overdrive, would have one believing that these network revolutions are on the verge of market launch. In reality, industry leaders—the major network equipment manufacturers and Tier 1 operators—have announced their visions for the future, and their interconnected ecosystems are in the process of executing upon them. In several areas—such as 5G and NFV—industry bodies are still defining and agreeing on standards, with real-world tests being conducted for validation purposes. While elements of this future will be rolled out gradually, it will likely be 2020 before it is fully realized and broadly benefiting consumers. And by then a whole new network vision will be under discussion.
