Posted on by — Leave a comment

Day 1 of Big Data In Focus: so where is it on the hype cycle?

I came to TMForum’s Big Data In Focus Summit in Amsterdam today expecting one of two things: either that hype would still be the main driver behind this market, or that we’d finally moved forward and fallen into the trough of disillusionment.

For Rob Rich, TMForum’s Managing Director of Research, it’s clearly the latter, which made him all the more proud that attendance was on a par with last year. But I think something else may be happening on the Big Data scene. Last year’s conference was still all about building the infrastructure for collecting and storing all that data. This year, Jessica Rausch, TMForum’s conference producer, set it up to be bullish with as many real world success stories as possible. As Rob pointed out to me nobody really wants to talk about failures, at least not their own. This positive feeling, during day one at least, was furthered by a move away from setting Big Data up, towards Big Data Analytics, which means really using it.

On the down side though, I was disappointed with the actual use cases described on the podium this year. Despite being interesting to listen about and the fact that Big Data made them more accessible, none were really new.

I mean, Big Data for Fraud detection, really? Credit card companies have been using algorithms since the 1980s.

I was also a bit startled by some of the revelations given about data collection with vendors easily spliping into the “I’m just doing what I’m told” roles while their operator clients don’t necessarily take the responsibility either. I couldn’t help thinking that it’s as if the Snowden affair never happened and Big Data is here to give the NSA a new lease of life.

I had a passing feeling of déjà vu with Big Data in the 2010s reminding me of Objet Orientation in the 1990s: it’s probably hugely important and underpins the future of much of IT, but may not ever become a market in its own right.

To counter this, a characteristic of Big Data that came up as much as last year was the strong linkage between process, organization and Big Data projects or as @yifatkafkafi tweeted: “People-driven transformation was key to Big Data success at Skype - @DvirYuval #bigdatainfocus” (you can find other tweets with the #bigdatainfocus overly long twitter hash tag).

Nobody used the term “Machine Data” last year, so Alice Crook of Splunk’s Marketing team put my out of my ignorance telling me that it’s “anything unstructured, created by machine”. I’ll come back to this and dig a bit deeper to get to a clearer conclusion in the next blog, where I’ll also write about the three companies I met: Cvidya, Splunk and Ontology and give TMForum’s view after my interview with Rob Rich.

Posted on by — 1 Comment

Has the French Telco Market derailed, again?

Let’s wind back the clock.

70’s – France’s young president elect Giscard d’Estaing comes to office and realizing that the French Telephone system is years behind those of other countries. Huge investment is ploughed into the state owned PTT. During the decade, France moved from an antiquated system to a state-of-the-art Signaling System 7 (SS7) Telephone network.

When I came to France as a boy in 1974 it took months for our family to get our first phone line installed – that was already a vast improvement over the previous decade’s standard 2-year waiting list. By our second move in 1977 it was just a few days, which seemed a bit like science fiction at the time and vastly impressed family and friends back home in the UK.

Before most people had heard of the Internet or IP, we were busy in France discovering online services with the Minitel system which saw it usage boom in the 80s and early 90s. The leading edge was taken off French Telecoms as there was then much hesitation of how to best react to the emerging Internet in the 90’s, embrace it or try to impose that national Mintel standard? By the time it became clear that the Web would replace the Minitel, other countries got to catch-up with the network savvy French and the European ISP business from the mid-90’s seemed relatively balanced although we were all trailing behind the US.

A combination of a sound regulatory approach and French business acumen the French Telco sector back on track in the second half of the 90’s.

Competition was introduced effectively without bringing the incumbent to its knees. Compared to the UK where BT was almost killed off by Thatcher or Germany where DT’s stranglehold on access networks is still to be fully broken, the French seemed to have got a pretty good balance.

So when Free, a new entrant in the ISP business built from a Mintel empire (ILIAD) aggressively launched their first triple-play offer in 2003, France Telecom was able to reciprocate and the other challengers like 9-Cegetel and Club Internet were also able to follow suite.

The French Telecoms subscribers saw the most incredible decade until about 2010 with always more services for a fixed price of about 30€. But Iliad/Free surreptitiously broke the 29,99€ rule it had forced on the industry, by removing a few standard features that became add-ons (like €1,99 for TV service). This improved ARPU and although detailed figures are hard to come by, it is no secret that ILIAD/Free make significant profit from their ADSL ISP-Business (a 40% margin is often cited).

This profit and a favourable regulatory approach (Free bought the 4th mobile 3G operator licence at a discount in 2009 paying just 240 M€ where the other 3 operators had paid 619M in 2000) meant that Free was able to launch a very aggressive mobile offering in January 2012. The success was immediate with about a million clients by the end of that month alone.

Despite being a free market economy, most French commentators agree that the 1000 layoffs competitor SFR announced in 2013 can be clearly attributed to Free’s price “dumping”. As an independent consultant based in Paris and focussed on new services, I’ve seen I’ve seen many operator projects canned this year due to innovation budgets being cut at Orange, Bouygues Telecom and SFR. French 4G is at least a 2-year behind what US operators have already deployed.

 

So what’s going wrong?

Too much of anything – even a good thing – can be bad. After proving that a balanced approach to (de)regulation really does work, France seems to have lost its unique touch.

The three incumbent Mobile operators were hoping that 4G would help fight off the low-cost battering from Free. But Free has entered the 4G market with a bang, announcing that 4G services would be provided at the same price as with 3G. All operators have had to follow suit ruining the basic tenet of Telco strategy, where superior services requiring significant Capex can bring in extra revenues. The only way forward seems to be lowering costs.

This situation of all-out war has led Bouygues Telecom to announce seriously low-cost ISP services to be detailed in January 2014, with as-much-as-you-can-eat triple-play offerings going for 15€ per month (compared to Free’s current 30-37€).

The price war has got out of hand to the extent that the French government has become involved in the industry-wide slanging match, although it’s not at all clear what it can actually do except maybe give the regulator some more power. If the situation keeps on escalating, subscriber glee will be short-lived in 2014 as one of the 4 operators is bound to go bust and maybe two other merge.

But if subscribers have so far only won out from this situation, shareholder prospects are less clear.

ILIAD-ORANGE-STOCK-2YRS-TO-EOY2013
In the 2 years to December 24th 2013, Iliad’s share price rose well over 50% (red) while Orange’s (blue) dropped more than 25%

The situation is almost reversed if you look just at the last 6 months

ILIAD-ORANGE-STOCK-6MONTHS-TO-EOY2013
In the 6 months to December 24th 2013, Iliad’s share price was stable, but recently dropped 12% (red) while Orange’s (blue) grew 20%

Posted on by — 2 Comments

LTE Broadcast part 3/5: Reasons to be cheerful

With three mobile operators, Verizon Wireless, Telstra of Australia and Smart Communications in the Philippines, now having conducted successful trials of LTE Broadcast, it is a good time to assess why at last multicast video transmission over cellular is coming of age.

There have been various failures in the past, such as DVB-H, mostly in Europe, and Qualcomm’s MediaFLO in the US, but they came too early and did not meet all the technological challenges. The underlying factor is the proliferation of tablets and ever larger smartphones, providing users with compelling mobile devices conducive not just for snacking but also consuming longer form content. This in turn has created a surge in video consumption over cellular that threatens to swamp both radio spectrum and the backhaul capacity of radio operators.  LTE Broadcast can save spectrum and fixed network capacity for popular live or linear video content that is consumed simultaneously by a sufficiently large number of people that would otherwise generate multiple unicast streams down to the cell level.

If there are only one or two people in each cell watching a given stream, then LTE Broadcast will not save much spectrum. But there is usually at any one time some relatively popular content being streamed over a given cellular network. Therefore, as various surveys from Ericsson and others have indicated, on average a mobile operator can cut traffic by at least 10% through use of LTE Broadcast. But that does not give the full picture, since the savings will be greater at peak times when more popular content tends to be consumed.

But the killer use case is for venue broadcasting, where hundreds or even thousands of people at say a large sporting event or concert may simultaneously want to tuck into action replays or content relevant to the occasion. Then LTE Broadcast becomes essential both to feed the cell where the venue is located and to distribute the content to devices over the RAN (Radio Access Network). Of course earlier cellular broadcast technologies could also serve venues in principle, but fell down over cost of infrastructure, lack of complaint devices and mobile screens being too small for compelling viewing. DVB-H required significant additional infrastructure investment to deploy, as well as specific upgrades to handsets.

LTE Broadcast will run over all LTE infrastructure and while it is not supported by current commercially available handsets, it will come out of the box with the next generation ready to hit the market in the second half of 2014. It also delivers video much more efficiently than its ancestors such as DVB-H, partly through being based on the 3GPP eMBMS (Evolved Multimedia Broadcast Multicast Service). eMBMS uses OFDM (Orthogonal Frequency Division Multiplexing), which has been incorporated a while in other wireless systems such as Wi-Fi, Wimax and DVB broadcast, but is new to cellular. The key point is that OFDM is a form of inverse multiplexing, splitting the broadcast signal over multiple low bit rate carriers that are therefore closely spaced in frequency, but do not interfere with each other because the waves are out of phase with each other.

This greatly increases robustness against fading at a given frequency, since the signal is split out across a range of frequencies. This robustness enables LTE Broadcast to deliver premium broadcast quality video over cellular for the first time. Apart from eMBMS, LTE Broadcast will in future be able to take advantage of associated developments in video transmission, in particular MPEG DASH for standardized streaming and HEVC (High Efficiency Video Coding)/H.265 for stronger video compression. Taken together these technologies will ensure that LTE Broadcast both reduces bandwidth consumed by mobile video and increases the quality. This is why despite having been bitten before the cellular industry is convinced that LTE

Broadcast will be a great success. Part one in the series by Philip Hunter is here, part two is here and part four is here.
Philip Hunter

Posted on by — 2 Comments

LTE Broadcast 2/5: operators get new revenue options

Now that LTE Broadcast is off the ground following the world’s first session on a commercial LTE network transmitted by Australian operator Telstra, attention has naturally turned to potential use cases and revenue generating opportunities. The initial attraction for cellular operators, once they have a sufficiently widespread LTE network in the target region, is the significant bandwidth saving they will get when streaming popular live video, with the consequent ability to optimise capacity and improve service quality for customers, especially in crowded areas. That is a business rather than a specific use case, but at least one major operator is planning to come out with all guns blazing with the latter from day one.

Verizon Wireless has already confirmed it intends to time its 2014 launch of LTE Broadcast in the US to coincide with one of the nation’s iconic sporting events to use the technology for streaming video in the stadium. This event is now widely believed to be SuperBowl, the annual championship of the US National Football League, to be held in New York on Sunday February 2nd 2014. This could be a big day for establishing LTE Broadcast’s most widely touted use case for broadcasting coverage, replays and associated video around sporting or music events where there are large numbers of people concentrated into a venue or arena. Without LTE Broadcast such content has to be unicast, which would cripple a cell and leave many users unable to access the service, while also causing congestion on the associated backhaul networks.

For operators this use case, live event streaming, offers potential both for cost savings by reducing or avoiding need for network expansion and also for generating revenue in various ways, for example subscriptions, pay per view, pay per event, a seasonal pass, or revenue sharing with content partners. Apart from live event streaming, other interesting use cases are on the horizon under five broad categories: real time video streaming across the whole network, news services, broadcast radio, off peak media delivery and cell based advertising. Real time streaming again offers revenue sharing opportunities, most likely with broadcasters or pay TV operators as the mobile part of their TV Everywhere strategies, perhaps combined with Wi-Fi hot spots. Already Verizon Wireless is in talks with US cable operators along these lines. Then for news services, LTE Broadcast has the potential to extend the scope of existing offerings by delivering news and sport as live updates or clips. There is the option of combining these with personalized lower bandwidth unicast services, such as specific stock updates.

These news services could be given away free as part of a premium offering to extract higher subscriptions, or could be advertising driven. The broadcast radio category is also creating some interest as a way of saving unicast capacity while again creating scope for upselling to premium packages and carrying advertising, as well as revenue sharing with providers of content such as music. Then the off peak delivery use case brings a range of opportunities, especially perhaps for tablets as they have greater storage capability, turning them into mobile DVRs (Digital Video Recorders). TV shows, movies, YouTube videos and newspapers could all be drip fed to devices this way, as could software updates. Apart from saving bandwidth this has pay per view opportunities as well as revenue sharing. Finally cell based advertising can be used to deliver ads targeted on the basis of location rather than personal preferences, with live events themselves being an obvious case, but also shopping malls and airports where particular retail outlets or restaurants might want to advertise to mobile users while in that vicinity.

This can be achieved by partitioning LTE Broadcast on a cell by cell basis, so that separate content would be sent to each cell. Most of these use cases are unlikely to be deployed in the immediate future as they will have to wait for widespread availability of devices compatible with LTE Broadcast, which will not be until well into 2015 at least. Meanwhile though some questions are arising, such as how LTE Broadcast will affect Wi-Fi expansion.

There has been interest in Wi-Fi broadcast as well for live events, with Cisco already offering this with its StadiumVision mobile app, which users can download at venues to access event specific broadcast streams over Wi-Fi. This could well compete with LTE Broadcast for live events streaming, but that is another story. Part one in this series is here and part three is here.

Posted on by — 3 Comments

LTE Broadcast 1/5: live debut heralds battles with Wi-Fi to come

The first live LTE Broadcast session delivered recently by Telstra in Australia raised the natural question of why things should be different this time.

The recent history of cellular communications is littered with the skeletons of mobile broadcast’s dismal past, with Telstra’s demonstration as it happens coinciding with the death of the last DVB-H service in Poland. Other notable mobile broadcast failures include Qualcomm’s MediaFLO in the US, which is significant in that the company is now strongly backing LTE Broadcast. Qualcomm’s chipsets are at the center of Ericsson’s LTE Broadcast platform used by Telstra for its live demonstration that served various devices with concurrent video feeds, including a sports match replay, general news and a large video file over the single LTE Broadcast channel.

The same Ericsson/Qualcomm platform will be used in a much more significant test of LTE Broadcast by Verizon at the 2014 SuperBowl in early February 2014. As the most popular US sporting event this will be the perfect springboard for LTE Broadcast, giving it the chance to demonstrate its ability to serve large numbers of users with concurrent streams within a single 4G/LTE cell. This would not of itself prove the case for LTE Broadcast, given that its predecessors could also deliver concurrent streams to multiple users. That after all is the whole point of mobile broadcast. Yet there are important differences this time that suggest LTE Broadcast will at least be a contender for delivering mobile video.

One big difference is that there are now eligible handsets for viewing video, notably tablets but also larger smartphones and a host of emerging hybrid devices that have got consumers hooked onto mobile video consumption when they weren’t before.

The other big difference now is the backing of key industry players and the fact that LTE Broadcast, or more precisely the eMBMS (Multimedia Broadcast Multicast Services) technology on which it is based, is an integral part of the LTE ecosystem underpinning current and emerging 4G cellular services.

All the big industry hitters, including Alcatel-Lucent as well Ericsson and Qualcomm, are full square behind it, along with a host of key second string infrastructure vendors like MobiTV. DVB-H was also an open standard but it required significant additional infrastructure investment to deploy, as well as specific upgrades to handsets. LTE Broadcast will run in principle over all LTE infrastructure and while it is not supported by current commercially available handsets, it will come out of the box with the next generation.

Qualcomm's Mazen Chmaytelli, senior director of business development at its Labs, is on record saying he expects the first LTE Broadcast capable handsets to come to market in the second half of 2014. For these reasons LTE Broadcast will be quite widely deployed, with AT&T as well as Verizon Wireless planning to do so in the US, while Korea Telecom is collaborating with Samsung towards a launch in South Korea. In Europe France Telecom’s Orange and EE have announced firm intentions to deploy LTE Broadcast.

Yet at one time there was equal momentum behind DVB-H and despite the fact LTE Broadcast is much better placed its success is still not a done deal. One reason for that is the advance of Wi-Fi, which may enable venues to cater for large scale events more cost effectively through temporary deployment of hot spots. At the same time versions of existing digital terrestrial standards, such as the DVB’s T2 Lite, could be better placed to meet the requirement for general mass delivery of video to mobile devices. There is a good reason for this.

Outside major events such as Super Bowl where large numbers of people will be consuming the same video streams such as sporting action replays, there will not often be more than a handful of people watching the same content in a single cell and often it will only be one. In the latter case LTE Broadcast collapses to unicast. Yet at any given time there may well be a number of people watching different streams in a given cell, so there is still a need for an efficient video delivery infrastructure, which DVB T2 Lite would be as it has a much larger coverage area than an LTE cell. Within the much larger digital terrestrial coverage area, even mid-tier content would often by consumed concurrently by several people, so that mobile broadcast would save a lot of spectrum. Mobile operators such as EE have stated that the initial “monetization motive” for deploying LTE Broadcast will come from more efficient delivery of video both over the backhaul networks and at the radio level inside cells.

My contention is that these efficiency savings will not materialize outside major events. Even within such events, Wi-Fi may be a more cost effective way of addressing the “Super Bowl” effect and could be offered as a service by mobile operators, which would benefit by offloading the traffic directly onto the more efficient fixed broadband infrastructure. On this count there are already a variety of products available, for example from Birdstep in Sweden, which enable automatic selection of traffic for offloading to Wi-Fi according to specified business rules.

With some operators already talking about temporary LTE Broadcast channels for venues as a future business model, it will be interesting to see how this approach will stack up against Wi-Fi and the answer may be not very well. At least with the upcoming Verizon demonstration at Super Bowl 2014, the battle lines are being drawn. Part 2 is here.

Posted on by — Leave a comment

IBC 2013 part 2: Brightcove & Envivio

Brightcove

I met up with Albert Lai who is CTO of Media and Broadcast Solutions. He told me that Brightcove has “traditionally been a big end-to-end black box”. But that is changing as more and more requests come in for just a part of the workflow.

Brightcove has seen the Software As a Service (SaaS) model move to a Platform As a Service PaaS (platform) one, but where modularity and Flexibility have become the key criteria.

All this modularity has come about because of the very different use cases Brightcove now caters for, ranging from the Cloud transcoding services for HEVC to the Viacom native apps including the Nikelodeon one that just recently won an Emmy.

Brightcove was founded in 2004 and IPOed in February 2012. As of December 31, 2012, Brightcove had 335 employees. It is headquartered in Boston with offices in San Francisco, Seattle, New York, London, Paris, Barcelona, Singapore, Seoul, Tokyo and Sydney.

The leitmotiv of my interviews this year was 4K and Albert told me that the Cable Show 2012 was when Brightcove first started getting requests about 4K.

Brightcove has conducted internal tests on 4K content management and has concluded that it’s a very promising approach. At present it is “less a technology issue than a general marketplace one, where availability is still an issue”.

I pointed out that Cinema content was already widely available in 4K but Albert responded, “Sure cinema content is there, but its just a small amount”.

Albert sees 4K representing a 100% to 400% increase in storage and transmission costs although he thought that HEVC will alleviate some of the pain by doubling the quality within the same bandwidth and providing for a better experience. The monetization question must however be addressed, so Brightcove is listening to the market and is ready for, but not pushing 4K right now.

The Zencoder purchase of last year is probably part of that readiness campaign. Zencoder is a pure-cloud software based encoding solution for live and on-demand content.

Envivio

This year I spoke to Jean-Pierre Henot the company’s CTO based in Rennes.

I went straight to the point in asking about HEVC so Jean-Pierre first explained the main demos at the front of the booth. Three screens were showing an HD demo of live content at 24 frames per second and all looked identical. One screen was showing MPEG2 at 8 Mbps, the centre screen displayed MPEG4 at 4 Mbps and the last one was showing HEVC decoded content, currently at 3 Mbps but expect to be reduced to 2 Mbps by EOY 2014. Note that the latter still has a CPU power requirement four times greater which really is an issue in the short term.

All the demos used software encoding.

Jean-Pierre noted that the hardware decoding part of HEVC is stable already for HD content, but that for 4K HEVC decoding is still only available in beta versions as the protocol is still a bit new. Fully compliant reference designs are expected for CES 2014.

This is inline with some minor issues still pending with the specifications for HEVC transport, which is otherwise ready for HD.

Like a decade ago with MPEG4, the situation regarding royalties is still being sorted out.

We discussed the cheaper 4K sets available today and Jean-Pierre scoffed at the 30HZ limit as Envivio sees 60fps as a requirement for sport. The hardware limitation of HDMI should be gone thanks to the new 2.0 specification. This should also be available at the beginning of 2014.

All in all, Henot confirmed the general industry view that royalties, devices, frame-rates and HDMI 2.0 have been the stumbling blocks so far for 4K.

As these are gradually removed, HEVC will take off but start initially below 4K.

HEVC has no noticeable impact on ABR, which is the key enabler for many OTT services.

There seems no doubt that HEVC will be next “best solution” available for video compression and with the shorter lifecycles all round, it will take less time than MPEG4 did to penetrate market.