Archive | Regulation RSS feed for this section

PSTN Switchoff Will be the Next Big Challenge

14 Oct



There’s a little nugget in the last paragraph of this article on Belgacom’s deployment of ALU vectoring (in French), one that I wish had been more developed. Let me offer my own translation here:

One last point, Standaert also reveals that these networks will continue to be converted to all-IP: “In Knokke le Zoute in particular, everything goes via IP. This year, we already migrated 400.000 lines. By 2018, Belgacom will no longer have any classic telephone line.” This means that the Siemens EWSD switches and the Alcatel-Lucent PSTN switched will finally all disappear. It should also free up a lot of physical space and allow Belgacom to sell even more of its buildings. By 2020, 30 buildings should be sold. Altogether, this should generate savings of 35 million euros.

I was astounded earlier this year to find out that the Australian NBN, despite the massive structural upheaval that it created had not been accompanied by a decision to phase out PSTN. Similarly the New Zealand policy makers, visionary though they might have been in establishing FTTH nationwide (or nearly so) and structurally separating their incumbent did not take this opportunity to dephase the legacy of PSTN.

Of course, this raises regulatory issues. In particular, what happens to voice regulation? An inclusive approach suggests that any VoIP service would become regulated, which may seem daunting and impractical. The alternative though would be that no voice services would be regulated, which seems somewhat dangerous.

I haven’t spent enough time gathering data and thinking about this, but I do think it’s going to be the major topic on the regulatory plate in the next five years. Stay tuned for more when I have more.


Photo Credit: Old Telephone (cc) by Macinate

The Maddening Schizophrenia of US Incumbents

11 Sep

(cc) Evan

(cc) Evan

I must confess to a particular fondness for those moments of synchronicity when different news items happen to collide in the same cycle, violently contradicting each other. This week is a great example.

On Tuesday, Techzone 360 published a really interesting editorial entitled Shifting Towards Symmetry: The New Broadband Landscape. I strongly encourage you to read it, but here’s the elevator pitch: AT&T and Verizon have finally realized that symmetry not only sells but differentiates them from cable. Suddenly, after years of arguing users didn’t need it, they flip around and extoll its virtues. Of course, to you and me, that’s not an incredibly forward-thinking stance, but think about where they are coming from!

The fact is that the market has moved beyond copper, whether they like it or not. That’s why Verizon can aggressively push symmetry (they’ve shed a lot of their copper, and have fiber everywhere else). AT&T, as is clearly highlighted in the editorial isn’t quite so at ease, having a large FTTC footprint that most definitely doesn’t support symmetry.

Still, no matter what the reasons, we should applaud them for being so modern, right?

Not so fast! That same day Ars Technica’s Jon Brodkin writes an article (AT&T and Verizon say 10Mbps is too fast for “broadband,” 4Mbps is enough) detailing how the same AT&T and Verizon are fighting the FCC tooth and nail to stop the regulator from increasing the requirements for a data service to be considered broadband. Today, 4Mbps down and 1 Mbps up is considered broadband, the FCC would like the download to go at least to 10Mbps and argues (quite convincingly) that you need that much for a perfectly normal evening at home. AT&T and Verizon are having none of it though, and their main argument is the same as ever (I quote AT&T): ” a 10Mbps service exceeds what many Americans need today”.

The real reason, again, is elsewhere: with a 10Mbps definition of broadband, suddenly the market doesn’t look so competitive anymore: 10% of users have no access to service and 30% only have one option for service. In other words, 40% of the market is either unserved or a monopoly. That doesn’t look so good…

So there you have it. PR schizophrenia at its most beautiful. I say the FCC should up the broadband requirements to a symmetrical 10Mbps. That should help AT&T and Verizon reconcile their paradoxical positions!


Photo: Insomnia, (cc) Evan

3G/4G as a landline data substitute

10 Sep

(cc) Philippe Henry

(cc) Philippe Henry

A guest post by Herman Wagter.

In the never-ending discussions about platform competition, the notion that mobile data is a substitute for landline data keeps coming back. I thought it might be interesting to recount a real life experience of trying to do just that.

One of my friends owns a couple of farms in rural France. They date back to the year 1100 and are located on the hills of the Rhone valley near Valence. He has restored them to apartments which he now rents to tourists.

Offering Internet access as part of the accommodation is a big problem. 10 miles of old copper lines doesn’t work for DSL, and there’s no line of sight to someone who might have better DSL to share.

The only option is to use 3G and (hopefully) later 4G as a substitute for landline data. Since the farms have a wonderful view over the Rhone Valley – which is populated with masts – this seems like a feasible idea. The signal strength though is quite low outside the buildings, and inside there is nothing as the walls are made of stone, and are of medieval thickness (80 cm).

In order to solve this, we installed:
- a 17 dBi planar directional antenna, with cable through the wall to
- a Teltonika RUT550 3G/4G router-to-wifi

and purchased some prepaid SIM cards (SFR as provider). I’ll spare you the administrative telecom-hell associated with these prepaid sims.

The planar antenna got -75 dB signal strength, which is fine. In all likelihood, the signal comes from a mast which is 7 km away, near line of sight. The Wifi signal inside the homes is fine as well.

The throughput fluctuates from 2 Mbps down to almost nothing, 100kbps up to nothing.

Looking at the throughput we see:
- sawtooth patterns, with a periodicity 5-10 seconds
- sometimes it takes tens of seconds to start the connection to a website

Still, it’s more or less usable to check mail, IM, and websites. For Youtube you need a fair bit of patience.

So we installed one per rented accommodation.

The big surprise turned out to be the data consumption: the prepaid SIMs had to be topped up constantly. 1Gbyte of data was usually gone within a day. We first suspected excessive Youtube viewing or the like, but every tenant swore they did not have any data-intensive kind of usage.

After analysis, testing and elimination of likely causes we found out that:
- the average number of connected devices per family member was more than one. Multiply that by the number of family members…
- phones and tablets are very frugal with data when they use a 3G or 4G connection. As soon as they see wifi (which they equate to a landline) the floodgates open.

So we now tell guests to:
- switch off icloud
- not update their OS
- switch off Dropbox, Gdrive and other cloud storage syncs
- check that muTorrent or other bittorrent clients do not start automatically in the background (which is the default on laptops)

That cuts back the consumption to bearable levels.

What this real life example of a mobile substitute to landline tells us is that you can do it if you’re willing to seriously cut back on most of what is considered “normal internet activities”, and if you’re willing to risk paying through the nose for overage charges.

In other words, it’s possible, but it’s not much of a substitute. We’ll see if 4G offers anything different, but one thing is for sure: the data consumption drivers aren’t going away…

Photo: (cc) Philippe Henry

TWDM and the regulatory framework for Fiber broadband

9 Sep


From the start, in markets where competition was valued over oligopoly or monopoly, regulation for next-generation broadband has been an issue. Trying to replicate the model for copper unbundling over a new fiber network was complicated for a number of reasons:

  • point to multipoint fiber, the cheaper way to deploy FTTH, does not allow for such passive unbundling
  • incumbent operators, the market players most likely to deploy some FTTH were happy with a technology that pushed hitherto passive wholesale customers into higher margin active products.

This led to point to multipoint technologies being deployed en masse by large scale players in most countries (exceptions: Sweden, Netherlands, Switzerland) which in turn led to essentially two regulatory models:

  1. either the regulator favored infrastructure competition, which in most markets is a very short term road to a duopoly in dense areas (cable / FTTP) and a monopoly everywhere else,
  2. or the regulator favored infrastructure sharing, with active wholesale products the norm and the incumbent operator retaining an upper hand on the market.

(France, typically, could not choose between these models so we have infrastructure competition with network sharing, and it’s a mess. But that’s a story for another day.)

The shortcomings of the latter model were partly addressed in some countries by forcing the incumbent to separate the network and services activities, either functionally as in the UK, or structurally, as in New Zealand.

Still, the wholesale price gap between passive unbundling over copper and active bitstream over fiber is a major issue as competing ISPs face either a significant margin drop as their wholesale costs double (roughly) or customers face a significant price hike, which may endanger the success of fiber products in general. In practice the market tends to price somewhere in between, but it slows down fiber adoption.

In the short term, it’s not a massive issue, simply because as with copper, no one would think of buying passive products (which require heavy investment to activate and operate) before they’ve acquired a significant market share in a given area. When that happens though, the classic “ladder of investment” concept that powered DSL competition in the last decade simply won’t work with PON networks as passive products cannot be delivered.

Or at least could not until now.

TWDM is a standardized technology that (on paper) delivers more capacity to PON architectures. In a (very) simplified way, what TWDM allows is to stack up to 4 virtual PON networks, operating on different wavelengths onto an existing PON architecture. These ‘virtual PON trees’ would each deliver 10G down and 2.5G up (although a symmetrical product has been specified, there’s no operator demand for it yet), which is about 4 times the capacity of current PON.

The beauty of it is that although you might use these stacked PONs to deliver way more capacity to end-users than what you currently can deliver with PON, you could also use them to lease these separate wavelengths to different ISPs, thus replicating a quasi-passive unbundling model over PON fiber.

This is probably not why some incumbents were so eager to choose PON for deployment, but it could eliminate a regulatory headache in the years to come as a more even form of competition is allowed to happen over existing architectures with limited reinvestment (and, most importantly, no outside plant work needed).

The technology is standardized and the first commercial products are expected from the big players (Alcatel Lucent, Huawei…) within a year. TWDM really is going to be the tech to watch for in the fiber space.

Behind the Great Firewall

3 Sep


I’ve been living in China for close to 4 weeks now. Admittedly not long enough to get a real sense for what it is to live here, especially since my living conditions are not exactly similar to those of your average Chinese citizen, but enough to have faced a number of challenges on the connectivity side that I thought were worth sharing, especially since they cast a different light on some of the debates that are still raging in the west around Net Neutrality and traffic discrimination.

To clarify a number of things up front, I have fiber in my home, or at least what’s advertised as fiber (I suspect it’s FTTB, which is indeed considered fiber in most of the world) with a nominal service of 100 Mbps down and 4 Mbps up. As a foreigner here, I have essentially two issues:

  • the general low quality of any traffic outside of China,
  • the large number of blocked services.

The former is a challenge because most if not all of what I access is not hosted in China. As a consequence, no matter what I’m trying to do, my speed is close to what I’d expect of a slow DSL or even worse. And there’s nothing I can do about that. It’s not censorship, at least not openly, but it’s traffic degradation for sure.

The latter is an addressable challenge, at least on paper. I subscribed to a VPN service before I came, and it’s been working fine, mostly, except that by definition the VPN tunnels to a non-Chinese destination (hence crappy bandwidth) and also that it’s regularly targeted by the Chinese authorities, so somewhat unreliable. Sometimes it kicks me out, sometimes it slows to a crawl.

What astonished me was the sheer scale of the stuff that’s forbidden. The most shocking discovery (and with a great impact for me) was to find out that all Google services have been blocked since the doodle on June 4th to commemorate Tienanmen Square. And when I say all, it really is everything: search engine, mail, maps, shared documents, everything. Dropbox is also blocked, for reasons unknown. Every large non-Chinese hosted blog platform (blogger, of course, but also wordpress, tumblr, etc.) is out.

And then there’s the sheer randomness of it all. Some services you can’t access for no apparent reason, others are so slow that you can’t figure out if they’re blocked or just snail-paced. And as I experience this, I wish some of our politicians and media people, those who see net neutrality as the enemy, I wish they’d come here and experience what a radical version of non-neutrality is. Again, I have a VPN service to overcome most of this (at the cost of speed) but most people don’t and/or can’t afford one.

Don’t get me wrong, I’m not suggesting that not enshrining net neutrality is the equivalent of doing what the Chinese (or Iranian, or Indian) government does. But I look at the UK’s blocking mechanisms supposed to protect children but really targeting just about any kind of site for arcane reasons that no one can figure out, and I think that what I have here is an extreme version of the same thing.

After years of experiencing a reliable internet, one where you don’t have to wonder before you access content or as you wait for it to load whether it’s actually accessible, being behind the Great Firewall is a humbling experience. You suddenly become aware of things you had forgotten, like the importance of your page loading within a reasonable amount of time.

Etsy last week came out with an impassioned plea for Net Neutrality for that reason alone : the smaller commercial sites will actually lose business (more business) to the large retailers under a net discrimination regime, simply because they won’t be able to afford the cost of bringing down their page load times. Meanwhile, a moronic French representative of rights holders comes out with a speech where he argues that Net Neutrality protects the big guys (meaning Americans) against the little guys (meaning French) and I’m thinking this guy should come here a bit and experience a radical version of what he’s wishing for.

They say you don’t know what you have until you lost it. In the case of an open internet, I’m certainly here to tell you that that’s true.

Photo: The Great Wall of China (CC) lutmans

How much money is there in Net Discrimination?

1 Jul

One of the striking realizations of my Analyst career was when I found out that very often companies in the broadband ecosystem defend, or even lobby for positions that they assume to be in their interest for ideological reasons, but without having worked out rationally if indeed they are. I have many an anecdote about crestfallen faces when real numbers are worked out and exposed.

And in fact, this has long informed my own approach to research: the idea is, based (ideally) on hard data or failing that on documented modeling, to assess whether a policy position actually makes sense or delivers what it’s supposed to deliver. This was the genesis of our short report Net Discrimination Won’t Buy You Next-Generation Access (still available, dirt cheap) in which we modeled a top-down revenue share between OSPs and ISPs to figure out the financial impact it would have. Long story short: not a lot, and certainly not enough to shift the lines in terms of network investment (as often argued by ISPs).

Fellow analyst and provocative thinker Dean Bubley has just gone one step further in what I consider to be a groundbreaking piece of analysis entitled Non Neutral Mobile Broadband Business Models. In this report, Dean doesn’t look at the classic arguments for or against net discrimination, he examines in-depth which business models net discrimination would enable and how much revenue they might generate.

You can get a feel for the material that’s in that report through the following presentation he’s made available on Slideshare:

The report is thorough, very well documented and enlightening. A highly recommended read.

Photo (cc) by Tax Credits

US Cable Reviled by its customers

9 Jun



It’s hard to believe that profitable businesses would be so detested by their customers, and yet survey after survey shows how US broadband users revile their cable operator. The latest is the subject of an article in the Washington post entitled A Soup of Misery, which shows (amongst other findings) that over half of US Cable customers would switch to another provider if they actually had an alternative.

The amusing thing (or ironic, or sad depending on how you want to look at it) about this is that cable still insists there is competition. If this market was a free market, with satisfaction ratings like that cable would be bankrupt instead of being amongst the most profitable industries in the US.

There’s an added bit of irony for me. A few weeks ago I got into a bit of tiff on twitter debating with Luigi Gambardella, the head of the European Telecom Network Operators’s Association (ETNO). ETNO has been lobbying fiercely for a regulatory model that’s more akin to that of the US, despite overwhelming evidence that that market is dysfunctional and anti-competitive. I naturally took exception to this view (as well as to the preposterous assertion that wherever fiber was being deployed, it was not regulated), and the back and forth went south very quickly (you can read the whole exchange here, assuming it doesn’t get deleted). The point here is that Gambardella’s final stroke was the following:

Needless to say that baffled me…

Anyway, all this to say that looking at the US for a functional model for Europe is not just ridiculous, it’s dangerous…

Net Neutrality Threatened…

24 Apr

Last night the US regulator FCC announced that they were carving out exceptions to Net Neutrality rulings for “fast lanes” that ISPs could charge to OSPs. Only in La La Land can this still be called Neutrality. I’ll write about this more at length when time allows, but in the meantime let me share this wonderful drawing on the topic by Susie Cagle:




TPG Telecom’s FTTB play a headache for Turnbull

11 Feb

See on Scoop.itConnected World

Although obviously very unhappy with the broadband monopoly created by the previous government, Communications Minister Malcolm Turnbull – as its shareholder minister – also has to protect the interests of NBN Co.

Benoit Felten‘s insight:

I must confess to a certain amount of pleasure watching Malcolm Turnbull struggle to keep together the NBN he has contributed so hard to undermine. It’s not good news for Australia, but you can’t help but feel he got it coming…

See on

Kansas Legislature Introduces Bill to Limit Internet Investment

31 Jan

See on Scoop.itConnected World

Benoit Felten‘s insight:

Looks like the US incumbents are at it again, trying to stifle competition through restrictive legislation, in Kansas this time. Commentary has been that it might be to hamper another Google Fiber, but Google Fiber isn’t a PPP. More likely they’re worried Google Fiber will give other Kansas communities ideas…

See on