Mobile World Congress 16 hype curve

Mobile World Congress 2016 was an interesting show in many aspects. Here are some of my views on most and least hyped subjects, including mobile video, NFV, SDN, IoT, M2M, augmented and virtual reality, TCP optimization, VoLTE and others

First, let start with mobile video, my pet subject, as some of you might know. 2016 sees half of Facebook users to be exclusively mobile, generating over 3/4 of the company's revenue while half of YouTube views are on mobile devices and nearly half of Netflix under 34 members watch from a mobile device. There is mobile and mobile, though and a good 2/3 of these views occur on wifi. Still, internet video service providers see themselves becoming mobile companies faster than they thought. The result is increased pressure on mobile networks to provide fast, reliable video services, as 2k, 4K, 360 degrees video, augmented and virtual reality are next on the list of services to appear. This continues to create distortions to the value chain as encryption, ad blocking, privacy, security, net neutrality, traffic pacing and prioritization are being used as weapons of slow attrition by traditional and new content and service providers. On the network operators' side, many have deserted the video monetization battlefield. T-Mobile's Binge On seems to give MNOs pause for reflection on alternative models for video services cooperation. TCP optimization has been running hot as a technology for the last 18 months and has seen Teclo Networks acquired by Sandvine on the heels of this year's congress.

Certainly, I have felt that we have seen a change of pace and tone in many announcements, with NFV hyperbolic claims subsiding somewhat compared to last year. Specifically, we have seen several vendors live deployments, but mostly revolving around launches of VoLTE, virtualized EPC for MVNOs, enterprise or verticals and ubiquitous virtualized CPE but still little in term of multi-vendor generic traffic NFV deployments at scale. Talking about VoLTE, I now have several anecdotal evidence from Europe, Asia and North America that the services commercially launched are well below expectation in term of quality an performance against circuit switched voice.
The lack of maturity of standards for Orchestration is certainly the chief culprit here, hindering progress for open multi vendor service automation. 
Proof can be found in the flurry of vendors "ecosystems". If everyone works so hard to be in one and each have their own, it underlines the market fragmentation rather than reduces it. 
An interesting announcement showed Telefonica, BT, Korea Telecom, Telekom Austria, SK, Sprint,  and several vendors taking a sheet from OPNFV's playbook and creating probably one of the first open-source project within ETSI, aimed at delivering a MANO collaborative project,.
I have been advocating for such a project for more than 18 months, so I certainly welcome the initiative, even if ETSI might not feel like the most natural place for an open source project. 

Overall, NFV feels more mature, but still very much disconnected from reality. A solution looking for problems to solve, with little in term of new services creation. If all the hoopla leads to cloud-based VPNs, VoLTE and cheaper packet core infrastructure, the business case remains fragile.

The SDN announcements were somewhat muted, but showing good progress in SD-WAN, and SD data center architecture with the recognition, at last, that specialized switches will likely still be necessary in the short to medium term if we want high performance software defined fabric - even if it impacts agility. The compromises are sign of market maturing, not a failure to deliver on the vendors part in my opinion.

IoT, M2M were still ubiquitous and vague, depicted alternatively as next big thing or already here. The market fragmentation in term of standards, technology, use cases and understanding leads to baseless fantasist claims from many vendors (and operators) on the future of wearable, autonomous transports, connected objects... with little in term of evidence of a coherent ecosystem formation. It is likely that a dominant player will emerge and provide a top-down approach, but the business case seems to hinge on killer-apps that hint a next generation networks to be fulfilled.

5G was on many vendors' lips as well, even if it seems to consistently mean different things to different people, including MIMO, beam forming, virtualized RAN... What was clear, from my perspective was that operators were ready at last to address latency (as opposed or in complement of bandwidth) as a key resource and attribute to discriminate services and associated network slices.

Big Data slid right down the hype curve this year, with very little in term of  announcement or even reference in vendors product launches or deployments. It now seems granted that any piece of network equipment, physical or virtual must generate rivulets that stream to rivers and data lakes, to be avidly aggregated, correlated by machine learning algorithms to provide actionable insights in the form of analytics and alerts. Vendors show progress in reporting, but true multi vendors holistic analytics remains extremely difficult, due to the fragmentation of vendors data attributes and the necessity to have both data scientists and subject matter experts working together to discriminate actionable insights from false positives.

On the services side, augmented and virtual reality were revving up to the next hype phase with a multitude of attendees walking blindly with googles and smartphones stuck to their face... not the smartest look and unlikely to pass novelty stage until integrated in less obtrusive displays. On the AR front, convincing use cases start to emerge, such as furniture shopping (whereas you can see and position furniture in your home by superimposing them from a catalogue app), that are pragmatic and useful without being too cumbersome. Anyone who had to shop for furniture and send it back because it did not fit or the color wasn't really the same as the room will understand. 
Ad blocking certainly became a subject of increased interest, as operators and service providers are still struggling for dominance. As encrypted data traffic increases, operators start to explore ways to provide services that users see as valuable and if they hurt some of the OTTs business models, it is certainly an additional bargaining chip. The melding and reforming of the mobile value chain continues and accelerates with increased competition, collaboration and coopetition as MNOs and OTTs are finding a settling position. I have recently ranted about what's wrong with the mobile value chain, so I will spare you here.

At last, my personal interest project this year revolves around Mobile Edge Computing. I have started production on a report on the subject. I think the technology has potential unlock many new services in mobile networks and I can't wait to tell you more about it. Stay tuned for more!

2015 review and 2016 predictions

As is now customary, I try to grade what I was predicting for 2015 and see what panned out and what didn't. I'll share as well what I see for 2016.

Content providers, creators, aggregators:

"They will need to simultaneously maximize monetization options by segmenting their user base into new price plans and find a way to unlock value in the mobile market.While many OTT, particularly social networks and radio/ audio streaming have collaborated and signed deals with mobile network operators, we are seeing also a tendency to increasingly encrypt and obfuscate online services to avoid network operators meddling in content delivery." 

On that front, I think that both predictions held true. I was envisioning encryption to jump from 10 to 30% of overall data traffic and I got that wrong, at least in many mature markets, where Netflix is big in mobile, we see upwards of 50% of traffic being encrypted. I still claim some prediction here, with one of my first post indicating the encryption trend 2 years before it started in earnest.

The prediction about segmentation from pricing as OTT services mature has been also largely fulfilled, with YouTube's 4th attempt, by my count, to launch a paid service. Additionally, the trend about content aggregators investing in original content rights acquisition is accelerating with Amazon gearing up for movie theaters and Netflix outspending traditional providers such as BBC with a combined investment by both company estimated in the 9$Bn range. Soon, we are talking real money.

In 2016, we will see an acceleration of traditional digital services that were originally launched for fixed line internet transitioning to predominantly mobile or mobile only plays. Right now, 47% of Facebook users are exclusively through  mobile and account for 78% of the company's revenue. More than 50% of YouTube views are on mobile devices and the corresponding revenue growth is over 100% year on year. 49% of Netflix' 18 to 34 years old demographics watches the service on mobile devices. We have seen signs with Twitter's vine,  and Periscope as well as Spotify , MTV and Facebook that the battlefield will be on video services.

Network operators: Wholesaler or value providers?
The operators in 2016 are still as confused, as a community as in 2015. They perceive threats from each other, which causes many acquisitions, from OTTs, which causes in equal measure many partnership and ill-advised service launches and from regulatory bodies, which causes lawyers to fatten up at the net neutrality / privacy buffet.

"we will see both more cooperation and more competition, with integrated offering (OTT could go full MVNO soon) and encrypted, obfuscated traffic on the rise". 

We spoke about encryption, the OTT going full MVNO was somewhat fulfilled by Google's disappointing project Fi launch. On the cooperation front, we have seen a flurry of announcements, mostly centered around sponsored data or zero rated subscription services from Verizon, AT&T.

"We will probably also see the first lawsuits from OTT to carriers with respect to traffic mediation, optimization and management. " 

I got that half right. No lawsuit from content providers but heavy fines from regulators on operators who throttle, cap or prioritize content (Sprint, AT&T, ...).

As for digital service providers, network operators are gearing themselves to compete on video services with services such as mobile TV /LTE broadcast (AT&T, EE, Telekom Slovenje, Vodafone), events streaming (China Telecom, ), sponsored data / zero rated subscription services (Verizon, T-mobile Binge On, Sprint, AT&T, Telefonica, ...).

"Some operators will seek to actively manage and mediate the traffic transiting through their networks and will implement HTTPS / SPDY proxy to decrypt and optimize encrypted traffic, wherever legislation is more supple."

I got that dead wrong. Despite interest and trials, operators are not ready to go into open battle with OTT just yet. Decrypting encrypted traffic is certainly illegal in many countries
or at the very least hostile and seems to be only expected from government agencies...

Mobile Networks Technology

"CAPEX will be on the rise overall with heterogeneous networks and LTE roll-out taking the lion share of investments. LTE networks will show signs of weakness in term of peak traffic handling mainly due to video and audio streaming and some networks will accelerate LTE-A investments or aggressively curb traffic through data caps, throttles and onerous pricing strategies."

Check and check.

"SDN will continue its progress as a back-office and lab technology in mobile networks but its incapacity to provide reliable, secure, scalable and manageable network capability will prevent it to make a strong commercial debut in wireless networks. 2018 is the likeliest time frame."

I maintain the view that SDN is still too immature for mass deployment in mobile networks, although we have seen encouraging trials moving from lab to commercial, we are still a long way from a business case and technology maturity standpoint before we see a mobile network core or RAN running exclusively or mostly on SDN.

"NFV will show strong progress and first commercial deployments in wireless networks, but in vertical, proprietary fashion, with legacy functions (DPI, EPC, IMS...) translated in a virtualized environment in a mono vendor approach. "

We have seen many examples of that this year with various levels of industry and standard support from Connectem, Affirmed Networks, Ericsson, Cisco and Huawei.

"Orchestration and integration with SDN will be the key investments in the standardization community. The timeframe for mass market interoperable multi vendor commercial deployment is likely 2020."

Orchestration, MANO has certainly driven many initiatives (Telefonica OpenMANO) and acquisitions (Ciena acquired Cyan, for example) and remains the key challenge in 2016 and beyond. SDN NFV will not take off unless there is a programmatic framework to link customer facing services to internal services, to functions, to virtual resources to hardware resources in a multi-vendor fashion. I still maintain 2020 as the probable target for this.

In 2016, the new bit of technology I will investigate is Mobile Edge Computing, the capacity to deploy COTS in the radio network, unlocking virtualized services to be positioned at the network's edge, enabling IoT, automotive, Augmented Reality or Virtual Reality services that require minimal latency to access content even faster.

In conclusion, 2016 shows more than ever signs that the house of cards is about to collapse. Data traffic is increasing fast, video is now dominating every networks and it is just starting. With 4K and then 8k around the corner, without talking about virtual or augmented reality, many of the players in the value chain understand that video is going the next few years' battlefield in mobile, OTT and cloud services. This is why we are seeing so much concentration and pivot strategies in the field. 
What is new is the fact that if mobile was an ongoing concern or barely on the radar for many so-called OTT, it has now emerged as the predominant if not exclusive market segment in revenue. 
This means that more pressure will rain on network operators to offer bandwidth and speed. My reports and workshops show that mobile advertising is not growing fast enough in comparison to the subscribers eyeball moving to mobile screens. This is mostly due to the fact that video services in mobile networks are a pretty low quality service, which will get worse as more subscribers transition to LTE. The key to unlock the value chain will be collaboration between operators and OTT and that will only happen if/when a profitable business model and apportioning of costs is worked out.

At last, my prediction about selfie kills seem to unfortunately have been fulfilled with selfies now killing more people than shark attacks. Inevitably, we have to conclude that in 2016, commercial drones and hoverboards will kill more people than selfies...

That's all folks, see you at MWC next month.

LTE World Summit 2014

This year's 10th edition of the conference, seems to have found a new level of maturity. While VoLTE, RCS, IMS are still subjects of interest, we seem to be past the hype at last (see last year), with a more pragmatic outlook towards implementation and monetization. 

I was happy to see that most operators are now recognizing the importance of managing video experience for monetization. Du UAE's VP of Marketing, Vikram Chadha seems to get it:

"We are transitioning our pricing strategy from bundles and metering to services. We are introducing email, social media, enterprise packages and are looking at separating video from data as a LTE monetization strategy."

As a result, the keynotes were more prosaic than in the past editions, focusing on cost of spectrum acquisitions and regulatory pressure in the European Union preventing operators to mount any defensible position against the OTT assault on their networks. Much of the agenda of the show focused on pragmatic subjects such as roaming, pricing, policy management, heterogeneous networks and wifi/cellular handover. Nothing obviously earth shattering on these subjects, but steady progress, as the technologies transition from lab to commercial trials and deployment. 

As an example, there was a great presentation by Bouygues Telecom's EVP of Strategy Frederic Ruciak highlighting the company's strategy for the launch of LTE in France, A very competitive market, and how the company was able to achieve the number one spot in LTE market share, despite being the "challenger" number 3 in 2 and 3G.

The next buzzword on the hype cycle to point its head is NFV with many operator CTOs publicly hailing the new technology as the magic bullet that will allow them to "launch services in days or weeks rather than years". I am getting quite tired of hearing that rationalization as an excuse for the multimillion investments made in this space, especially when no one seems to know what these new services will be. Right now, the only arguable benefit is on capex cost containment and I have seen little evidence that it will pass this stage in the mid term. Like the teenage sex joke, no one seems to know what it is, but everybody claims to be doing it. 

There is still much to be resolved on this matter and that discussion will continue for some time. The interesting new positioning I heard at the show is appliance vendors referring to their offering as PNF (as in physical) in contrast and as enablers for VNF. Although it sounds like a marketing trick, it makes a lot of sense for vendors to illustrate how NFV inserts itself in a legacy network, leading inevitably to a hybrid network architecture. 

The consensus here seems to be that there are two prevailing strategies for introduction of virtualized network functions. 

1. The first one, "cap and grow" sees existing infrastructure equipments being capped beyond a certain capacity and little by little complemented by virtualized functions, allowing incremental traffic to find its way on the virtualized infrastructure. A variant might be "cap and burst" where a function subject to bursts traffic is dimensioned on physical assets to the mean peak traffic and all exceeding traffic is diverted to a virtualized function. 
2. The second seems to favour the creation of vertical virtualized networks for market or traffic segments that are greenfield. M2M and VoLTE being the most cited examples. 

Both strategies have advantages and flaws that I am exploring in my upcoming report on "NFV & virtualization in mobile networks 2014". Contact me for more information.

LTE Asia: transition from technology to value... or die

I am just back from LTE Asia in Singapore, where I chaired the track on Network Optimization. The show was well attended with over 900 people by Informa's estimate. 

Once again, I am a bit surprised and disappointed by the gap between operators and vendors' discourse.

By and large, operators who came (SK, KDDI, KT, Chungwha, HKCSL, Telkomsel, Indosat to name but a few) had excellent presentations on their past successes and current challenges, highlighting the need for new revenue models, a new content (particularly video) value chain and better customer engagement.

Vendors of all stripes seem to consistently miss the message and try to push technology when their customer need value. I appreciate that the transition is difficult and as I was reflecting with a vendor's executive at the show, selling technology feels somewhat safer and easier than value.
But, as many operators are finding out in their home turf, their consumers do not care much about technology any more. It is about brand, service, image and value that OTT service providers are winning consumers mind share. Here lies the risk and opportunity. Operators need help to evolve and re invent the mobile value chain. 

The value proposition of vendors must evolve towards solutions such as intelligent roaming, 2-way business models with content providers, service type prioritization (messaging, social, video, entertainment, sports...), bundling and charging...

At the heart of this necessary revolution is something that makes many uneasy. DPI and traffic classification, relying on ports and protocols is the basis of today's traffic management and is becoming rapidly obsolete. A new generation of traffic management engines is needed. The ability to recognize content and service types at a granular level is key. How can the mobile industry can evolve in the OTT world if operators are not able to recognize a content that is user-generated vs. Hollywood? How can operators monetize video if they cannot detect, recognize, prioritize, assure advertising content?

Operators have some key assets, though. Last mile delivery, accurate customer demographics, billing relationship and location must be leveraged. YouTube knows whether you are on iPad or laptop but not necessarily whether your cellular interface is 3G, HSPA, LTE... they certainly can't see whether a user's poor connection is the result of network congestion, spectrum interference, distance from the cell tower or throttling because the user exceeds its data allowance... There is value there, if operators are ready to transform themselves and their organization to harvest and sell value, not access...

Opportunities are many. Vendors who continue to sell SIP, IMS, VoLTE, Diameter and their next generation hip equivalent LTE Adavanced, 5G, cloud, NFV... will miss the point. None of these are of interest for the consumer. Even if the operator insist on buying or talking about technology, services and value will be key to success... unless you are planning to be an M2M operator, but that is a story for another time.

The war of machine 2 machine: Internet of nothing?

A recent Tweet conversation got me thinking about all the hoopla about machine-to-machine / internet of everything.

Many telecom equipment manufacturer hail the trend as the next big thing for wireless networks, both a bounty to be harvested and a great opportunity for new revenue streams.

There is certainly a lot to think about when more and more devices that were not designed for real time connectivity are suddenly able to exchange, report, alarm... All these devices that could have well suited rudimentary logging software or technology, most of the time for manual retrieval (think your home gaz, water  or electricity meters being read by a technician) could in the future be eligible for over the air data transfer.

A similar discussion I had at LTE world Summit where I was chairing the data explosion stream comes to mind. A utility company in Italy, I think, had rolled out these "smart" meters. The implementation in labs was flawless, the utility was going to save millions, with only a handful of employees monitoring the data center instead of hundreds scouring the countryside reading manually meters. What was unexpected was that all meters had the same behavior, sending keep-alive and reporting logs at the same time. This brought the wireless network down, in a signalling and payload storm that was self-inflicted.

When I look at all the companies that have created apps with no knowledge of how a phone or a mobile network behaves, I can't help but think about the consequences of meters, cars, irrigation sensors, gaz turbines, fridges and traffic light trying to send snippets of data and signalling through a wireless network with no understanding of how these signals and streams will affect the infrastructure.

This immediately bring to mind horrific headlines: "Sheep herds monitoring device bring down network in New Zealand!". "Water and electricity meters fighting over bandwidth..."

More seriously, it means all these device manufacturers will need to get some serious programmers who understand wireless not only to put the transmitters on the devices but also to code efficiently so that signalling and payload are optimized. Network operators will also need to publish best practices for M2M traffic in term of frequency, amount, etc... with stringent SLAs since most of this traffic will be discrete (subscription paid with service or device, no usage payment).