Content Barons, Smart Dust

March 31, 2017 • Blog Post • By Christopher Surdak, Chief Field Technologist and HPE Big Data Platform Business


  • By 2016, data traffic will be 200 times greater than voice traffic, presenting a serious problem for telco vendors
  • Mobile data is massive today, but it is a fraction of the volume we should expect in 2020 as the Internet of Things takes off
  • Consumers continue to recognize the importance of cyber security and are likely to pay carriers a premium to protect their sensitive data

Expert predictions for the future of Telecommunications

In preparing for this forecast for the next five years in Telecommunications, I reviewed literally hundreds of predictions made by other industry observers. Many of these predictions were deeply technical: indexes of refraction, quantum computing, network capacity, technical innovation, evolving standards, regulatory responses, etc. Indeed, one needed an engineering PhD and an acronym encyclopedia in order to follow the conversation. In an industry heavily reliant upon technology and carefully monitored by a smorgasbord of regulatory bodies, this emphasis upon internal factors and controls comes naturally.

In coming up with this list of predictions for the future of Telecommunications, I instead wanted to focus upon external factors that will drive the industry. Technical innovation will continue to be critical to the industry; however, there are many external forces that will significantly impact how this industry will evolve over the next five years.

1. Integration

As I pondered the coming half-decade for the Telco industry, perhaps the most obvious disruption that has appeared is that of Vertical Integration. Communication, or more precisely connectivity, has quickly become a utility. It is no less critical to modern living than water or electricity, and yet for the most part people who have access to connectivity take it completely for granted. And as with most things that we take for granted, the perceived value of connectivity continues to fall, as long as we don't lose it.

The commoditization conundrum

This leads to several interesting and predictable results. First and foremost is the commoditization of connectivity. Being connected continues to become cheaper and cheaper, adhering rather slavishly to Moore's Law of diminishing costs. Ever-increasing capacity and diminishing marginal utility when combined with just a little bit of competition, leads to what is sometimes called the Pricing Death Spiral. The cost of providing such a service keeps falling, and competition means that the price keeps getting smaller and smaller in a strong, negative feedback loop.

Users of communication services, particularly cell phones, are very familiar with this effect. Every few months, one of the cellular service providers makes a tiny, incremental reduction in the price of their service, in order to gain or retain customers. Shortly thereafter all of their competitors respond in-kind, and a new price floor is reached until someone again reduces their price to gain a miniscule advantage. These tiny cuts continue to accumulate over time, dramatically reducing the price paid for the service.

Each communications service provider has invested billions of dollars to build their network and each spends a similar amount each year to maintain that network. There is a tremendous fixed cost, hence vendors must maintain a high level of utilization just to break even. Also, capacity is a fleeting asset; every percent of network capacity that is not used second-by-second is wasted forever. As such, vendors work to maximize their customers' use of their network, all of the time.

The data disconnect

In December 2009, data traffic on mobile networks first surpassed voice traffic, and has since exploded. By 2016, data traffic will be 200 times greater than voice traffic. This has presented a serious problem for telco vendors. When voice traffic dominated the market, there was a clear connection between cost and value. When a customer spoke for five minutes, they paid for five minutes of connectivity. They paid proportionally more or less, depending upon how long they talked.

However, this is not how data connectivity works. Providers are still charging according to capacity used (mega- or giga-bytes per month, for instance) but customers' consumption of capacity is totally different. Data customers don't perceive value according to how many bytes they download. Rather, they perceive value in how many videos they download, how many pictures they upload or how many Skype calls they make in a day. To a customer, a 140-character text from a family member may be far more valuable than a two megabyte video they watched while waiting for a taxi, regardless of the dramatic difference in their associated costs. The perceived value of content is no longer directly tied to the cost of delivery, leading to a great deal of disenchantment by customers of these services.

Network operators don't necessarily get this effect, but other segments of the information value chain do. The providers of the products and services delivered through telecommunications networks absolutely understand their value to their clients. They also recognize that while connectivity is critical to their business, it is again a utility. Connectivity is capturing an ever-smaller proportion of the information value chain, while content, service, and product deliverers capture ever-more.

As a result, it's not terribly surprising that content companies are moving into the infrastructure game. Google, Amazon, Microsoft and other major content players have amassed enormous network capacity in their own right, and it makes sense for them to own the underpinnings of their business (the networks) in order to ensure that their customers are always online. Indeed, AT&T's market capitalization in 2014 was about the same as it was in 2006, while Google's value has more than doubled. Facebook didn't even exist in 2006, and by 2014 it was worth roughly 40 percent more than AT&T (approximately $222 billion).

By 2020, it is likely that one or more major telco companies will be acquired by a content company. And once this process begins, a feeding frenzy likely will ensue. Regulatory bodies might slow this process down somewhat, but as with so many industries throughout history, vertical integration in the Internet is almost certain to happen.

2. IoT: the traffic explosion

The next major trend that will impact Telecommunications is the explosion of connected devices. This Internet of Things, or Thingification, will add billions if not trillions of new connected data sources globally by 2020. Objects throughout our lives will become connected, aware and chatty, constantly transmitting information across our global networks.

The upswing of all of these devices will be an astronomical growth in data volumes; we will quickly push through exabyte volumes and enter the world of zettabytes per year. As sensors are added to everything, and everything starts sending out signals, the trunk lines of our networks will have to carry this crushing load of information. Interestingly, this problem will begin to look bleak, start to look better, and then turn a corner again to become downright daunting.

As Thingification begins, there are relatively few smart devices out there. And, they're not particularly smart. A connected light bulb may report on its status every thirty seconds, but let's face it, it doesn't have a lot to say; it's either "on" or "not on." Either way, that's not a lot of data. The challenge grows as millions and then billions of bulbs and toothbrushes and microwaves all start pushing more and more information, all of the time. The aggregated traffic from all of these devices will be enormous, and will stress our networks to the max.

From talkative to smart

Most first generation "smart" objects aren't really smart, they're chatty. They use sensors to describe their situation, but they're not really analyzing or thinking about that data; they just measure and transmit. As traffic explodes and as these devices evolve, we will embed more and more intelligence in these end-points. Then, they won't just measure and transmit, they'll process, evaluate, synthesize. As objects get smarter they can become more independent and transmit less. They won't just broadcast the conditions around them, they'll make decisions independently and simply phone home to notify someone of what they did.

Second and third generation smart devices will produce significantly less network traffic, but only briefly. By 2020, there will likely be sufficient numbers of these devices that they will actually start interacting with each other, rather than with human controllers "back home." This Machine to Machine (M2M in industry speak) category already is experiencing serious growth, and we haven't even scratched the surface.

As this transition occurs (and in some industries, such as the stock market, it already has), devices will start to have entire conversations with each other. Millions of devices will negotiate with each other, make deals, and act as agents for whoever is the lead device. These smart agents will constantly work on our behalf, performing vast quantities of transactions, calculations, and dealing in order to secure for us the latest widget we desire.

Quintillions of conversations

Hence, by 2020 the network traffic from smart devices will plateau somewhat, and perhaps even fall slightly. But, this is simply a deep breath before yet another meteoric climb. Self-aware, negotiating smart devices will create a volume of data that will look more like Moore's Law squared or even cubed. And, the more consumers recognize the benefits that these smart devices yield, the more they will be in demand.

3. Mobility

It shouldn't come as a surprise to anyone that mobility is a huge trend in Telecommunications. In 2014, there were an estimated 1.2 billion people with mobile devices worldwide. By 2020, this number is forecasted to exceed 5 billion, or about 80 percent of humanity. (Interestingly, there will be more than 9 billion subscriptions, reflecting a large number of people with multiple accounts.) Global growth of mobile connectivity is far outpacing hardline connectivity.

This makes sense, as most growth is occurring in the developing world and amongst poorer populations. Such consumers may not even own a home, let alone a FiOS connection. For these people, mobile is cheaper, more convenient and more useful, even when landline connectivity is an option. And further, many customers in the developed world are retiring their landline connections in favor of their broadband mobile connections.

This growth in mobile traffic, and in particular mobile data, has been mind blowing. Annual growth rates of 80 to 100 percent have been normal, and are likely to accelerate. A substantial proportion of this new growth will come from Thingification. As a trillion smart devices coat the planet with sensors, a large proportion of these devices will connect wirelessly. Some of the most interesting use cases for smart devices will have to connect through mobile networks because they will themselves be mobile in nature. Drones, robots, in-car Wi-Fi, and smart dust used to track wind patterns or pesticide use are all going to create massive amounts of data, and will do so wirelessly.

As mentioned under Thingification, as these devices get smart they will interact more and more with each other, and less and less with people. Machine to Machine traffic (M2M) will become Mobile Machine to Mobile Machine (MM2MM) traffic. These devices will hold trillions of conversations with one another on mobile networks and drive the growth of mobile traffic by up to an order of magnitude over the coming decade. As huge as mobile data appears to be today, it is a fraction of a percent of the volume we should expect in 2020 as Thingification takes off.

4. Saturation: the search for growth

It is widely noted in the Telecommunications industry that while traffic has grown dramatically, per-customer revenues have remained relatively flat. Customers are demanding more and more bandwidth, yet the price is dropping at about the same rate as data growth. This shouldn't come as an enormous surprise. Consumers view connectivity as a utility, with a given cost per month for use. Nature hates a vacuum, and customers hate unused minutes and megabytes. Hence, it's fairly normal for consumers to use the capacity that they've paid for, and rarely consume more.

Digital junkies continue to consume more and more bandwidth, but at roughly the same ongoing price. This is a direct effect of Moore's Law. I'll willingly allocate the same proportion of my household income to connectivity, but I'll expect that connectivity to be twice as good every 18 months or so.

Demographic challenges

In mature markets like the United States, market saturation is a growing problem. By 2015, broadband technologies (both landline and mobile) have been widely available and rapidly adopted by those who value these services. Network providers have reached the point of saturation, where those customers who want the service already have it. Conversely, those customers who don't want broadband are not likely to be swayed by changing price points, greater utility, or higher throughput.

This group of non-customers are Digital Luddites. These are people who, despite the digitized, socialfied, twittered world around them, still exist without being constantly plugged into The Matrix . For those of us who are junkies, this existence just seems weird. Indeed, my father owns a basic cell phone, which is only turned on when he wants to make an outgoing call; that's just bizarre.

In many ways carriers appear to have given up on these Luddites, and for good reason. These consumers have demonstrated that they not only can survive without perpetual connectivity, they actually prefer to connect rarely, and then only on their own terms. Not being connected is a lifestyle choice, and its one which many of them are proud to make. Such people are notoriously difficult to move, and most carriers have given up doing so.

I would offer that this perception of saturation is about to change, and for good reason. Many people are digital have-nots by choice. They have the funds, they simply don't feel that they have the need. Not surprisingly, many of these people are older, and many of them are part of the aging, baby boomer generation. While these consumers may choose for themselves not to be connected, their friends and family may no longer give them that option.

As they age, many Boomers will be cared for by their children or grandchildren who are digital addicts. Their entire existence is centered on connectivity, and they actually have some difficulty in interacting with the physical world without digital enhancement. They will care for their elders just as much as prior generations, they will just do it from hundreds or thousands of miles away, from a smartphone app, using sensors, robots and drones.

As they retire, Boomers will enter retirement communities and assisted living facilities which are fully-digitized in order to be as efficient as possible. Indeed, care in the United States is being digitized by law, and this trend will likely accelerate as more and more objects in our world become connected, self-aware and smart. Digital Luddites will be forced into using these technologies by the world around them and will likely consume vastly more bandwidth than they, or their carriers, ever imagined.

One of the keys of such adoption is that service provided by smart devices doesn't seem invasive. Retirees' kids may force them into adopting digital living, but they'll want to be nice about it. As this occurs, the last remaining percentages of market penetration will be achieved, and the market will be thoroughly saturated. At this point, carriers will be hard-pressed to find revenue growth through customer acquisition, while also fighting to maintain revenues in the face of both Moore's Law and the pricing death spiral.

5. Security: the network is the threat

No discussion of the present or future of the Internet can ignore the issue of security. In 2013, one in five Americans had their identity stolen, and that number is increasing rapidly. In February 2015, Intuit, maker of Turbo Tax software, actually suspended sending tax returns to state tax authorities, as there were apparently millions of fraudulent tax returns being filed in the first week of tax season. Security used to involve setting up a firewall and then walking away. Things have changed, and dramatically so. Security is no longer about prevention and protection; the walls of the castle have all been breached. Security is now an issue of detection, intervention, and interception of threats that are ever-present.

The challenge that this poses to Telecom providers cannot be understated. As custodians of the networks, carriers play a pivotal role in fighting the new threats that are emerging. Persistent penetrations, distributed denial of service (DDOS) attacks, botnet wars and other sophisticated, massed attacks are now commonplace, and can bring down entire segments of the global network. Customers will begin to expect, then demand, more proactive protection from the entire Internet value chain, and carriers will be expected to support these expectations with a range of technical and operational innovations.

Encryption will become much more widely utilized, however with the expectation that network performance won't suffer as a result. Virtual Private Network technology also will advance, making networks both more flexible and more secure. As more and more data moves into the cloud that data will be more accessible, valuable, and vulnerable. Security will move from being passive to active and from defensive to offensive. As many in the security space lament, "the defense has to be right all of the time; the offense only needs to be right once."

The desire for greater security may be a boon for carriers, if they embrace the need. While there is severe pricing pressure for their commoditized offering, providing added security to their services represents an excellent avenue for upselling to customers. Not all people will pay more for all of their traffic to be secure; not all data is created equal. But, customers are likely to pay a premium to protect data and transactions that are particularly sensitive, and the price they're willing to pay may be rather large. Network providers who respond to this need will show improved earnings, better customer retention, and slower erosion of their value proposition.

From talkative to smart6. Ascension: Skynet finally gets real

With this last prediction I'm going to stick my neck out a bit. In the 1990s several companies attempted to bring broadband technology into orbit. Teledesic, Globalstar and Iridium all planned on deploying fleets of tens, hundreds, even thousands of satellites which would provide broadband networking to every place in the world. The vision behind these networks was bold, courageous, forward-thinking, and fatally flawed. While companies poured billions of dollars into building and deploying these networks, most of them never left the ground, and those that did have gone through repeated cycles of bankruptcy and restructuring.

It was troubling to me that these networks in the sky, Skynets, failed because I actually worked on some of their designs back when I started my career as a satellite engineer. Their designs were very advanced, sophisticated, even elegant. Some of their capabilities are pedestrian today, but at the time they were breakthroughs. While we applied the best engineering available in designing these networks, their failure was due to a fundamental failure in understanding how consumers think.

There are obvious reasons why these satellite constellations wouldn't work. Launch costs were enormous. Teledesic would have required hundreds of launches at $100 million a pop. The networks weren't really useable until the entire network was in place, which would require years of launches to complete. The satellites themselves had useful lives of only a handful of years, so once the entire networks was up and running the first satellites you launched were already waiting to be replaced, before they fell from the sky for lack of fuel. And, the painful truth was that in 1995, there just weren't that many people looking to place a call to Paris or New York from the top of Mount Everest. Those that did, weren't likely to make up the tens of billions of dollars invested in these systems by paying $10 million per minute of talk time.

Nonetheless, I'm predicting that Skynet 2.0 is about to reappear. These space-, balloon-, or drone-based systems will provide high-quality broadband access to anywhere and everywhere in the world, they'll do it affordably, and they'll likely start arriving around 2020. And this time, they'll be wildly successful.

What will be different this time around? A number of things. First, launch costs are lower, and if innovators like SpaceX, Blue Origin and Virgin Galactic continue to innovate as they have recently, they're about to drop dramatically. This is a huge factor. Second, the core technologies are better. Continued miniaturization of electronics means that the satellites in these constellations can be much smaller than twenty years ago, while at the same time being much more capable. The technical hurdles are smaller as are the relative costs.

Despite these positives, technical innovation won't be the reason Skynet 2.0 will succeed. Rather, this time around there's an entirely different business case. The Skynets of the 1990s were supported by the mentality of "If we build it they will come." That is, once such a global network was in place people would naturally want to use it. They'd come up with novel uses which would make remote connectivity valuable, and the network would soon have lots of users taking advantage of these innovative new use cases. In this model, capacity leads demand; indeed capacity creates demand.

The problem with this model is that it just doesn't work. Infrastructure projects are inherently expensive; it takes a lot of capital to lay the foundations of society. For such projects to pay for themselves there must either be a governmental mandate and lots of public funds (think universal access to electricity or the telephone), or produced by industrialization. The U.S. superhighway network was spurred on by both growth in car ownership, post-war suburbanization, and the need to move military assets quickly during the height of the Cold War. And the Internet would have languished for years or decades as ARPANET, a novel tool of university professors, if the personal computer had not taken off the way that it did.

The lesson here is that infrastructure must not precede demand; the economics are too onerous. The space networks of the 1990s tried to deploy infrastructure that would then lead to demand; the recipe for an epic fail. The Skynets of 2020 will be completely different, because the demand will already be there. The other five forces listed here all create demand for ubiquitous, constant connectivity. Thingification will mean that trillions of devices will want to connect from anywhere, everywhere, all of the time. Indeed, collecting data in remote locations will be particularly valuable because they are remote, and hence largely unknown. There are billions of customers using pervasive communications all over the world, and they want that same experience everywhere they go. Our cell towers are Conestoga wagons heading west, and now we're ready for trains, planes, and automobiles to take their place.

As a result, I think that the efforts of Amazon, Google, Microsoft, Facebook and others to develop off-world networks will ultimately succeed. This ascendency of the network will be rapid, disruptive, and compelling, and it will likely start apace around 2020. You may not like some of the implications that it'll have on our society and our daily lives, but Skynet is coming to a mountain-top log cabin near you! there must be a huge pent-up demand for the capacity as its deployed. Most successful capacity projects in history were in response to demand, rather than the cause of it.

The American Transcontinental Railroad would have been an epic failure if there had not been a California Gold Rush. Conestoga wagons were a poor means of getting across the continent, but they were sufficient in light of the demand for such capability of 1848. Similarly, there would be no reason to dig the Suez or Panama Canals without a significant and growing shipping trade in both raw materials and manufactured goods.


KU Leuven and Hewlett Packard Enterprise Advance AI Capabilities through New Supercomputer

Press Release

Technology for a Better World: A Q

Blog Post

HPE enhances partner program to deliver higher payouts and more growth opportunities

Press Release