Posted on Leave a comment

Have we finally reached the inexorable end unlimited data plans? Amdocs thinks we have.

During a business lunch in 2002 I had my first real conversation with my boss’s boss’s boss in France Telecom. Joining the table late, I found the discussion already heated. Jean-Jacques Damlamian, the longest standing board member the company has ever had was then the acting CTO (I say acting because there was no such title, JJD as he was known internally is now retired). He was adamant: « We have just made the biggest mistake in the telecoms industry’s history, since Graham Bell. » France Telecom had launched its first commercial ADSL packages for private subscribers just a few years previously. Damlamian was referring to the fact that these packages were limited only in speed, and unlimited in volume of data. Already it was becoming clear some users’ requirements were thousands of times higher than the average user. France Telecom was budgeting millions of Capital expenditure for just a handful of subscribers.

Last week, David Amzallag, Amdocs’ new CTO explained to me how he believes his company’s vision might contribute to getting the industry out of the fatal trap it set for itself over a decade ago. I met him in the run up to the 2011 Broadband Forum in Paris.

David just left from a 4-year stint as BT’s chief scientist. Those years spent on capacity planning, convinced him that even if networks get smarter and grow in capacity as fast as possible, current usage trends will outstrip our best efforts, leading to major bottlenecks and frustrated subscribers. David went as far as to say that T1s and T2 are in serious trouble as their bread and butter gets commoditized.

MPLS, one of those technologies supposed to make capacity management much more flexible, has delivered only part of the promises so far; capacity is running out anyway, however flexibly it’s managed. Amdocs see a metered future.

So based on the assumption that demand is going to exceed capacity, Amdocs commissioned a study on the future of data pricing from Heavy Reading. Amdocs’s core business is where Operations Support Systems (OSS) and Billing Systems (BSS) meet customer experience so they have a vested interest in the outcome. Reassuringly, the study concludes that operators believe users are willing to pay more for more and are willing to accept some kind of flexibility (Over 80% of interviewed operators said that their future plans include data plan shared between several devices e.g. tablet & smartphone. Also, over 65% said their future plans include data plans shared between several family members). Heavy Reading appropriately interviewed operators, because they are Amdocs’ customers. The research would carry more weight if it also included the opinions of real subscribers.

Amdocs don’t believe the problem will be solved with sponsored connectivity, where, for example, Facebook pay the ISP a few dollars to carry their traffic. David went on to say that the only way forward is for the network’s Operational Support Systems (OSS) to be better linked to the business issues.

He described several use cases with an overall data quota for the whole family across many devices. Parents might be prepared to pay a premium to be assured that during their single daily leisure hour, bandwidth was guaranteed. Children could swap their leftover bandwidth allowance amongst themselves. For the more tech savvy families, the hard-core gamer might even give up some bandwidth in exchange for better latency that the bandwidth hungry movie-buff sibling doesn’t need. Towards the end of the month, if the operator sends a warning message that data limits will be probably be exceeded, the family could decide either to extend existing plans for a premium or enforce lower usage until next month. Thus maybe watching a few older movies from the home NAS instead of streaming from the cloud.

For other customer segments like single adults, Amdocs sees people wanting to fulfil unique needs at specific times through different devices. Subscribers will be prepared to pay for this and data plans will need to be so flexible that David says the real name of the game will be personalization. He used the expression of “Quality of Service On Demand” and “dynamic customer profiling” to describe such cases.

These quota based premium packages could co-exist with unlimited ones, but Amzallag insists the latter would suffer much lower bandwidth. He said that net neutrality wouldn’t be completely gone as prioritization isn’t based on packet contents but on whether the customer is paying a premium or not. Of course the Net Neutrality activists would disagree saying that the corollary of prioritization is de-prioritization, which means blocking if congestion is too bad.

Other detractors can forcefully argue that there will be no turning back from “as much as I can eat” data plans. But the Amdocs vision addresses that pretty squarely saying unlimited data can co-exist with quota based plans. My remaining doubt is a central one. Do subscribers want this? The "Global Tribes" consumer research, conducted by Coleman Parkes, that Amdocs published earlier this year, addresses the question "are consumers prepared to pay more for more?" For most segments and markets it concludes reassuring that yes they are. However, having myself witnessed first hand how incredibly different markets around the world are; I avoid using patterns from one market to make deductions for another. UK and US customers are clearly being weaned away from unlimited plans. My gut feeling is that subscribers in the rest of Europe and places like Russia that benefit from fierce ISP competition might be harder to transition away from unlimited plans. Despite its recent problems, the Netflix model has proven it can fly; limited data packages could shoot the model straight out of the sky.

The context of the data plan debate will probably evolve rapidly as the boundary between the fixed-line and the mobile broadband markets gets fuzzier all the time. Quota based plans have been becoming the norm on mobile broadband. My daughter left home for a tiny 1-roomed flat in Paris last month and in looking for an ISP on a very tight budget, we concluded that using a mobile broadband subscription might be best – she only uses Facebook and email regularly and will keep away from streaming for now (which could incite illicit downloading when she’s back home, but that’s another story...) So I look forward to talking again with David maybe at next year’s Broadband World Forum, to see how things have panned out.

After writing this I found out on twitter that fierce competition on the iPhone 4S launch is pushing the big US operators back towards unlimited ... looks like the market may not be ready after all. Exciting yoyo times.

Posted on Leave a comment

New horizons for CAS and DRM companies beyond security.

It has always been a mystery to me why security vendors like Verimatrix always seem to punch above their weight in the pay-TV business. If you look at other value chains, beyond the content industry, stakeholders are credited with a relative importance directly proportional to their financial standing yet Verimatrix has even more mind-share than market-share.

During IBC this year, Steve Christian set me straight and gave me a glimpse into Verimatrix’ future, opening a world of possibilities for the Pay-TV security industry. It was a humbling “why on earth hadn’t I realized that before?” moment.

I was completing a small scouting project on cross-media content recommendation (expect a post soon on this) when we met last month, so I started by picking Steve’s brain on this topic.

A critical mass of users is required before an operator can map behaviour and usage patterns. You need lots of users to get a great service, but you need a great service to get any users. The end game is to understand the patterns of content consumption.

The problem is not only theoretical. Like all Telco’s, the one I worked for was very proud of owning precious customer data. Huge data mining projects on months-old data was the only use that data usually got. TiVo is an exception in collecting data used in almost real time.

Steve’s first point was that CAS and DRM vendors are already in the right place to transparently capture the critical real-time subscriber intelligence needed to deliver a recommendation service.

In their just-released white paper (Arming Digital TV Operators with Real-Time Subscriber Behavior and Usage Data), Verimatrix quaintly call this enabling Progress and Profit. C’mon Steve, these things always come in threes, so what’s it going to be: Profiling for Progress & Profit? Hmm not so quaint, because that’s where there’s still an unresolved issue: privacy. That’s the nice thing about white papers; you can skip the tricky bits.

If privacy does prove difficult, it could however be handled in a very transparent way. The whole recommendation concept requires transparency. People need to know why they are being recommended something to willingly make more personal information available.

The hard part to fix, in getting recommendation to work well, is collecting and understanding user data collected implicitly or explicitly.

Of course understanding the content metadata and classifying all the movie genres TV programs and whatever other content available, isn’t trivial either. This is no longer a real issue as there are now dozens of companies out there with variations on semantic analysis and other approaches with which to do this. Many of the algorithms are even available as open source. Suppliers range from TV specialists like TV Genius, bee TV or Orca to web-based solutions like, with people like Jinni somewhere in between. Of course, if you want to build a solution from scratch there are also some pure-players in the algorithm side like Think Analytics.

But all these solutions amount to nothing if you can’t get access to significant user transactions. That is why Verimatrix can solve one of the hardest parts of the problem in a more timely way than many of the above-mentioned vendors.

I didn’t discuss with Steve whether Verimatrix would be looking to develop such features, but the company’s track record suggests they are more likely to partner with whoever is best in class in this area.

Once you able to intervene in, or just under the TV, there are at least two other key areas you can intervene on: Quality management and social networking.

The burden of managing the quality of experience with an ever-increasing range of devices, and with a broadening scope of features and services, is becoming difficult to bear for operators. Some standards like TR-069 have at last emerged for basic requirements like firmware upgrades, but no standardized solutions are available for managing more advanced issues such as monitoring the service delivered.

After many years of caution, focussed on the risk of devices creating a storm of traffic if they all had the same issue to report at the same time, even the big safety-conscious Telcos are looking to deploy agents into both their STB and their home gateways.

Whether using a standard like TR-135, or as-yet proprietary products from the likes of Agama, Mariner, Witbe or Cisco, operators actively engaging in monitoring will develop a dynamic view of how services are being delivered into people’s homes. These operators will find themselves in a position to initially deliver services with a higher quality of experience and eventually deliver even better services altogether.

But the main issues operators have encountered, with embedded monitoring has been software integration. Vendors promise trivial two-week integration efforts, but this has often dragged on to yearlong projects. Here again the CAS vendors come up with a trump card: they are already integrated with all the end devices, that is: connected TVs, STBs, Tablets, smart phones and of course PCs.

But another angle Steve Christian developed was that CAS vendors are already doing monitoring. They maintain and monitor the security of pay-content. Extending from traditional pay-content to other types of content is not necessarily a huge step to take. As anyone who’s dared to get their hands dirty with something like wireshark (this is a free network protocol analysis tool) will know, there is almost no technology barrier to getting to grips with quality metrics like packet loss or network jitter. The most important metric of all is service availability that security vendors are in a prime position to report on.

Working in an ecosystem is part of a security vendor’s daily routine. To improve the quality of experience of the end user, security vendors could easily partner, say, with an EPG meta-data provider like Rovi to ensure that the right data is available at the right place at the right time. They could also use one of the EPG quality specialists like EPGenius to add value to the EPG data by analysing it, correcting discrepancies and adding things like missing series links.

The third point Steve Christian mentioned was Social networking. We didn’t really get into any details here. We were running out of time, but also the business case isn’t so compelling. Verimatrix can enable better Social-TV implementation, but I don’t yet see any clear path to market. I do believe the Social TV will be a reality soon (see my blog here). It is not yet clear who will prevail. The most compelling demos are multi-screen with the TV as a basic video output device and a laptop, smart-phone or tablet to interact with. Fitting into that ecosystem will require agility.

My talk with Verimatrix happened just about when Toy Story3 was released, which I still haven’t seen. Steve’s passion reminded me a bit of Buzz Lightyear’s mantra “From infinity to beyond” except that his could maybe be “from Security to beyond”.

Posted on Leave a comment

Back from my first Anga

My feelings from the Anga cable congress can be summed up by my reaction to Cologne’s main landmark.

A surrealist sight hits the visitor exiting the main Cologne train station. The gargantuan cathedral called the ‘Dome’ seems to rear up from the past; which is how I perceive the cable industry that Anga represents. The Dome is surrounded by modern, more ugly buildings, that seem to be slowly encroaching upon it’s lebensraum just like the Internet or DTT threaten Cable. The train station itself, with its underground lines, represents different hybrid ways of transporting things; I wonder if it’s undermining or on the contrary underpinning the great old cathedral’s foundations. From the outside it’s as if the majestic building, symbol of the cable industry, were dying. It’s blackened at places and has almost permanent scaffolding that seems to hold it up.

A different story emerges when you are inside. One’s jaw drops with the shear size. Wikipedia just told me that it was once the tallest building in the world. The vertical proportions of the arches stretch upwards as if some divine hand had pulled malleable stone upwards. Then you look closer and realise that no, this is the work of hundreds of humble stonecutters over centuries. All the carefully crafted slabs stack up in powerful columns, just like the innumerable insignificant single-valley cable plants, in nearby Switzerland, add up to a powerful force. The strength of this force will keep enemies at bay for the foreseeable future.

OK I'll leave the poetry there; now for some reporting:

A first surprise on my first visit to Anga is that it's marketed as a cable event when IPTV, FTTx, Satellite, hybrid and more abound. OK so there are a bunch of booths with nothing more than little bits of cable on display, but no-one ever stops on those anyway and one sees one or two even at IPTV World forum. The organisers must be doing something right though because at least one company I met, Zappware, were on the waiting list and didn’t get a booth.

Wondering around for two days wasn’t enough for my sense of direction to kick in so I kept on getting lost in the huge hall's two sections. There were an order of magnitude less booths than say at IBC, but almost all of them were mid sized. I used the Korea “Green IT” pavilion at one end as a landmark. It was a mistake because each time I passed it I felt a little worse about how clueless-marketers are trying to jump on the green bandwagon and have all but broken it before it has even left the station. The only green thing in that pavilion was the word.

On the positive side, the TivO demo on the Conax booth rocked. It was by far the most convincing illustration an industry oxymoron: walled garden OTT, where operators give access to all the content that is readily available out there, while reassuringly (?) never letting the subscriber out of their sight. The business models and content deals are not yet clear as earlier failures from the likes of Joost show, but the end-user proposition is now overwhelmingly compelling. It’s beautiful. I want it now in my home now!

The OpenTV 3D demo was the first, modest but effective, effort of using 3D in the interface itself, not just the video, which I’ve seen. 3D will not be revolutionizing mainstream interfaces soon, but I definitely got the feeling of peeking into the future.

I saw three interesting companies from my hobbyhorse Quality of Experience area. Ineoquest where present with a big booth clearly hoping to push their IP and head end leadership further into the cable market. Skyline’s Dataminer product was on prominent display. It’s an interesting way to commoditize the likes of Ineoquest by putting the effort on a central piece of software, which other quality systems then report to. Another outsider in the Quality area was the German supplier Axiros. From a background of managing zillions of legacy boxes, their approach is now built on the TR69 protocol. Axiros offer a new product that sits between the device management systems (ACS) and the devices themselves (STBs) so that more meaningful info can be taken from each device. Axiros performs some of the QoE monitoring functions themselves this way.

I got a private demo from German STB maker Winbox. They had a really simple “why-didn’t-I-think-of-that” ideas and an effective demo for simple « push » playlist. A VoD server pushes short content like trailers to device’s local storage (HD or Flash memory) so a personalised preview or barker channel is available. If not a killer app, this could at least be a killer VoD-ARPU generator...

Posted on Leave a comment

IPTV World Forum reports on Videonet

IPTV World Forum blogs are on videonet.

Overall there was a great attendance, with lots of people coming to Olympia to do business. The booths were decent although I didn't spot many exciting innovations this year and as usual the conference was of varying quality, from gems to blatant sales pitches.

The 2010 conference taught me that Canvas is a purely British thing for now, and a few companies companies stood out for me:

  • Netgem for their ability to do fancy stuff with run-of-the-mill chipsets,
  • Echostar for finally winning a deserved award for Sling-loaded
  • BeeSmart for their interesting freebee initiative,
  • Intel for getting Sodaville up and running, most impressive with Amino.

But overall it was a good show on my subjective scale.

Day 1 blog post is here

Day 2 blog post is here

Day 3 blog post is here.

Also some in depth analysis of some of the issues I became aware of at the show to follow.

Posted on Leave a comment

Why I’m going to IPTV world forum – March 23rd 2010

I’ll be staying for the three days at the event next week. I always spend long moments hesitating whether such a time commitment is reasonable. I thought I’d share my thinking in case it helps you make your mind up (and helps me decide whether to fork out a couple of k€ to go to NAB next month or not).

When I worked at Orange, heading international IPTV deployment, I gave one of the first talks from a major IPTV player at one of the first versions of the show in 2005 or thereabouts. It was about the technical challenges of IPTV deployment from a Telco perspective. IPTV World Forum holds a little sense of nostalgia for me.

In the early days Junction ran the show, and I remember it feeling like a special occasion. It’s probably the nostalgia speaking or maybe the fact that as it wasn’t yet mainstream we all felt a bit more leading edge. I suppose I can replace "feel good" factor from being a pioneer then to a "feel good" factor from knowing I was here first. To illustrate the newfound importance of the show, big decisions now get initiated at Olympia and the IPTV WF awards get fought over more and more.

We’re not supposed to decide which show to go to based on the quality of the tea and biscuits (no don’t pretend you never do). In this respect IPTV WF is pretty good on logistics, except maybe fort the A/V equipment that forces me to sit in the front rows if I want to both see and hear.

The main reason I’m going is networking. There’s only so much you can do with LinkedIn & Co and face-to-face meetings do make a difference. I know at least half the companies exhibiting and over a dozen speakers so it’ll be worthwhile just to catch-up. I also need to generate some new leads for my consulting business (;-€).

I'm looking forwards to awards ceremony at the end of day one. Only the English can make a pompous event fun as well (mind you the great food & drink helps). As I'm a judge, I can’t really talk about that till the results are out … but there were loads of good entries this year.

There will be demos of some really new things I want to see at the exhibition. This year I’m looking forward to seeing the BeeSmart free middleware that’ll be launched during the show. I’m also hoping ROVI will show their new promising looking EPG offerings. As I missed the NDS widget demo at IBC I was hoping to catch up on that but I can’t see them on the exhibitor list :o( - maybe I’ll have to go to the NAB show after all.

I wrote an blog entry here on the rosy future for the IPTV Monitoring market so I’ll be asking all the vendors like Mariner, Bridgetech, Ineoquest, Agama and the new kid in town from India called First Media what they think about that i.e. do they too see a blue ocean of opportunities?

I hope to do a post-show blog on the future of interfaces so I’ll also hop into booths from some middleware people like Dreampark and Nagravision.

Many of the usual suspects from the STB arena will be at the show so I’ll be checking out where they are in terms of chipsets & new deployments (although these tend to boringly all be confidential). But the ecosystem is constantly changing as the box makers move upwards or sideways in the ecosystem so I’ll be looking out for any exciting demos from booths like Netgem, PACE, SoftAtHome, Echostar, Awox and Amino.

I’m a bit disappointed in the content recommendation supplier line-up. Recommendation is still a stumbling block that we haven’t fixed. Hopefully Gravity R&D will have a better demo than they showed at the Prague show. I don’t know why the more mature suppliers like Jinni aren’t coming to the show. That’s food for thought for another in depth analysis.

I always drop into the Edgware booth not only because it’s invariably one of the nicest but mainly because they are a surprisingly interesting company to talk to; they have a real vision.

Oh and I’ll make a point of having a proper talk with the Canadians from Evertz because I kind of botched it last time in Prague and have heard they deliver a monitoring good job for Sasktel in conjunction with Mariner Partners who btw will also both be presenting at the show.

With over 100 exhibitors I expect it will take me at least a day and half to see everything I want to, and As I’m chairing during day one I’ll be there the whole time.

There’s some luck involved in choosing the best conference to listen to unless you know the speaker beforehand. Most speakers do go to the trouble of writing interesting fresh slides and are really worth listening to. However as with any mainstream conference, some vendors that pay a lot to get to say basically what they want amazingly get away with too much sales pitch. You should complain to the organizers if you see this. I certainly do. Telcos with big IPTV deployments who also get red-carpet treatment sometimes go around with the same slide deck from conference to conference; I’ve identified the speakers by now, but it's always worth listening to them if you haven't heard it before.

In the end I clearly do recommend going (twitter me @nebul2 to meet). If you decide not to come, several of us will be reporting from the show on Videonet.