Wednesday, June 27, 2007

FTTH is the endgame (but it will take a while)

Yesterday I attended the 'Next Generation Network Conference', hosted by Euromoney's Global Telecoms Business, in Amsterdam. Thanks for inviting me!
It was an interesting day, even if there wasn't so much really new. I enjoyed talking to telco and vendor officials, most notably Dirk 'Amsterdam' van der Woude.
Below I will summarize my take of the views on a number of trends (which I regularly write about on this blog):
  • FTTH, VDSL/LLU/regulation 2.0, Web 2.0, SaaS, WIMAX
  • Separation, co-op
  • Owning the customer, advertising as a business model
Companies and organisations represented included:
  • KPN, BT, Vodafone, Orange NL, Thus, 2 smaller Dutch MSOs (CAIW and Kabel Noord)
  • Xconnect, Alcatel-Lucent, BroadSoft, Sonus, AlwaysON
  • OPTA (the Dutch NRA)
  • Analysys, Fitch

1. Demand

The perennial question: will 10 Mbps be enough? or 100 Mbps? Some statements (not precise quotations) included:
  • There is no ceiling (KPN),
  • even if the new services are as yet unspecified (Analysys).
  • Why should the exponential increase stop now(BT)?
My comment:
  • KPN, through Nico Baken (senior strategist and professor at Delft University), proved to be among the most radical. By the way, when I asked Nico how he feels about KPN's current strategy, he responded somewhat in this manner: KPN is among the most respected telcos, and they allowed me to hire 12 PhD's to work on long-term strategies, in order to allow KPN to maintain its lead. Bravo Eelco Blok for gathering this team at the heart of KPN!

2. FTTH, NGA (access)

Statements included:
  • We see no business case, except for greenfield operations (BT).
  • FTTH is the endgame (OPTA, KPN).
  • Build-outs in the Netherlands (7.0m households) are projected to go from 115k at present to 580k by 2009.
  • Public/private partnerships (PPP) will emerge (KPN).
  • Wireless will be the way to connect over the 'last few meters' (KPN).
  • Within a few years, all munifiber in the Netherlands will be bought by either KPN or Cablecos (CAIW).
  • 2 Infrastructures (copper/telco and coax/cableco) are not enough to ensure real competition (OPTA).

My comments here:

  • KPN's Nico Baken was probably among the most impressive in his presentation. His visionary analysis underscores that KPN fundamentally believes in FTTH - as well as PPP!
  • Dirk pointed me to a new development: KPN plans to connect 11k homes in the eastern town of Enschede and eventually the entire city (155k inhabitants) will be covered. "We will try to convince any doubters that copper access is not sufficient in tomorrow's world." I suspect that OPTA's 580k number (see above) does not include Enschede, which would take the number up to 735k. By the way, Dirk added a new overview (as of June) of fiber developments to the Citynet site.
  • There was surprisingly little on VDSL. I feel that everybody present believes in FTTH, which makes VDSL a transitory if not outmoded technology before it is even launched.
  • OPTA's acknowledgement of FTTH as the endgame is positive (in fact, it was aired before, most recently last week), but saying that 'two is not enough' is puzzling (to put it mildly). I would say: all we need is one (FTTH), which needs to be regulated. I feel that OPTA regards LLU and even bitstream access as a separate infrastructure.
  • Now, if even KPN feels that PPP is the way to go, separation makes more and more sense to me. How about separating both the telco (KPN) and cableco ('Zesco') networks - which I feel could stimulate the two new network companies to build a nationwide FTTH network through some PPP/joint venture (with Reggefiber).

3. NGN (core)


  • The 21CN project started out as a cost savings measure, but grew into a complete business transformation (BT).
  • We offer NGN as part of our 'Business Transformation Partner' offering (Alca-Lu).
  • NGN implies cost savings, but at first a 'hump' will appear in capex and opex spend (Alca-Lu, BT). BT sees costs at a low in 2013, when the hump is coming to an end and normal growth is resumed (at a level less than half of what it is now).
  • BT established BT 21C Global Venture as a way to leverage its know-how that it is acquiring, doing the 21CN project (BT).
  • Apart from cost savings, NGN is all about new services (BT, Alca-Lu) for which SOA must be adopted (BT).
  • Telecom New Zealand wants to be a service provider and is less interested in being a network operator (Alca-Lu).
  • Altnets lack scale for NGN projects (Orange NL).

My comments:

  • SOA and SaaS will be recurring themes for telcos,
  • As well as separation. One could say that separation (and Saas) are ways of taking outsourcing to the extreme.
  • I wonder how the BT 21C Global venture fits into the IT services market. Do they have customers yet?
  • I will pound on one of my favorite subjects once more: why on earth do we see so few co-ops?

4. New services


  • Future services will include HDTV, social networking, software apps and Web 2.0. Many may not be really new but subsititutes. "New services are as yet unspecified", and business models are unclear (Analysys).
  • Many new services, such as triple play, aren't really new. Blending however (like on-screen caller notification) is what we will be seeing a lot of (Alca-Lu).
  • In offering IPTV, we focus on interactivity, not on exclusive content. We will offer "what is relevant for our customers" (KPN).
  • We aim at personalisation (Vodafone).
  • Data may actually make up for much of mobile growth decline, but IPTV will not do the same for fixed operators (Fitch).
  • "The customer experience needs vast improvement." (Fitch)

My comments:

  • Somebody mentioned that it is all about "owning the customer". I couldn't agree more. That is also why I question KPN's representation of WLR, which I believe distorts their net line loss numbers. Sure, WLR still adds to wholesale revenues, but the customer relationship is gone.
  • I wasn't terribly impressed with KPN's IPTV ('Mine') presentation. The service will be (re)launched after the summer, but not as a premium service anymore. The UI didn't look very fancy. The feedback they had so far (the low key launch was done in May 2006) must be a long shot at what they overambitiously describe as "what is relevant for our customers".
  • Blending sounds like mash-ups, in Web 2.0 terms.
  • Selling to Google or KPN is one business model, and otherwise it seems to be advertising. Sure, budgets move online and can be targeted a lot better, but in the end online advertising will prove to be a cyclical market. The Broadband Incentive Problem kind of raises the same issue: in the long run, things need to be paid for, preferably in a usage based (not flat-fee) model.

5. Other

  • I spoke to Orange NL and other people, who all seem to believe that T-Mobile will not dispose of the Wanadoo BB unit of Orange NL, once the acquisition is worked out. I always assumed that T-Mobile would be a mobile-only play (outside their home markets in Germany and Eastern Europe) in the US, the Netherlands, Austria, the Czech Republic. But who knows they will embrace the convergence story.
  • On the side, if T-Mobile do embrace a convergence model, selling T-Mobile USA must come into play again (remember the cablecos work with Sprint and the satellite companies teamed with Clearwire, so teaming with a fixed or WiMAX operator seems hard).
  • I am getting pretty fed up with people saying that the end user is not interested in technology - to the point that I start to feel that people are increasingly familiar with alphabet soup.
  • On the side, WiMAX was touted by someone in the audience as a technology capable of bypassing cellular networks in large cities. I do not wish to be overskeptical about new technologies, but I think we have to be realistic. It is an emerging technology, especially 16e (there are many 16d deployments underway, including Vodafone's Malta plans). Handset range will be a major issue. At first, the technology was supposed to deliver 70-120 Mbps over a distance of 50 km. Now, 16d seems to deliver perhaps 10 Mbps over 5 km (in a NLOS situation). Imagine what the performance will be for 16e, assuming the kind of usage we see in cellular networks today. And then I haven't mentioned building the network, from construction, backhaul and interconnect up to marketing ...
  • Not to end on a sour note: Xconnect is a very interesting story. Peering is a whole new way of saving costs (and enabling new services). In fact, it is like OTC trading. Actually, I included peering in my own overview of efficiency measures (including such seemingly unrelated things like DWDM, CDN, P2P, MPEG-4 and AJAX) at my Tiscali Wholesale presentation two weeks ago. Mail me for that presentation.

No comments: