Monthly Archives: July 2012

Grahame Lynch, Paul Brooks and Simon Hackett debate NBN policy & OECD rankings


Yesterday’s column chastising the political class for focusing on Australia’s lowly fixed penetration rate and ignoring the country’s outstanding top ten position in wireless broadband seems to have hit a few raw nerves! Along with a batch of “not-for-citation” emails, at least one correspondent, the well-known technical guru Paul Brooks, had a go at my logic.

Posting to the Business Spectator site where my column had been cited, Brooks wrote: “Mobile broadband and fixed broadband are not exclusive. Grahame Lynch conveniently ignores that most of the mobile devices making up the wireless stats—standard mobile handsets and tablets—will be set to connect to fixed-line WiFi at home and office, only using the cellular broadband in transit. It’s not either/or the Australian public uses mobile broadband for expensive but intermittent convenience, and fixed-line for performance and lower cost.”

I might point out that I have written about mobile broadband hand-off to fixed networks before, notably in a February 28 editorial where I wrote: “While the NBN is excessively household-centric, it will still create tremendous utility for mobile operators, as a spill-over network for heavy data loads and in practice, as the secondary part of a bundle. …[The trend to mobile operator-supplied public WiFi] reduces loads on its mobile networks in areas of high congestion such as shopping malls, transport hubs and significantly, apartment blocks. I wonder if in ten years time we will see the NBN as primarily a vast publicly subsidised backhaul and data download relief network for the two major mobile operators, whose commercial scale and market relevance will dwarf everyone else.”

But I’m not entirely sure that Brooks’ characterisation of mobile broadband devices only making intermittent use of mobile networks and largely defaulting to fixed connected Wi-Fi nets most of the time accurately reflects the more complex reality of usage and customer value perception.

Mobile network operators have long observed that much use of their networks takes place in both homes and offices, and the simple fact that there are 15m active mobile broadband subscriptions compared to 5m fixed broadband subscriptions suggests that the platform is not necessarily seen as overly expensive and to be placed in cotton wool most of the day.

If people truly resented mobile data prices and switched to WiFi at the first opportunity then why has fixed line penetration largely flattened in recent years even as the network has increased its reach through rural investment and the Top Hat installations? And likewise, why would consumers so readily adopt the 3G versions of devices such as iPads that are also available in WiFi only modes?

Brooks isn’t entirely wrong but neither is he entirely right. For a start, if mobile broadband was really such a laggard in price and performance compared to fixed broadband, then what accounts for the massive differences in rankings? 21st in the world for fixed, 8th for mobile. Surely there would be more symbiotic convergence on those stats. Perhaps this is because the reality is a little more nuanced.

The late Exetel CEO John Linton observed a couple of years ago that lower end HSPA pricing was such that it had become a superior competitor on price to entry level DSL plans. Likewise other people enjoy incidental access to fixed broadband in their workplaces and, thus, forgo it in favour of a wireless subscription for their personal needs.

More prosaically, users will happily download using a 3G network for things they may value as important or timely—podcasts, work email attachments, MMSs from friends– and save things of lesser personal value such as a movie download or generic software update for when they are in the fixed zone.

The latter shows up as a bigger data download in terms of bytes but that size differential doesn’t necessarily equate to utility or importance—hence my quibble over the characterisation of wireless as inferior. There is, what I believe to be, a false idea in the industry that every bit and byte is equal: they aren’t! As long ago as 1994, comms bureaucrat Chris Cheah observed that an SMS is worth thousands of times more per byte than a movie download is.

But despite all that, perhaps there is an argument to be made that none of this condemns a $40.8b investment in the NBN, especially if the NBN has a twin purpose of supplying household desktop PC and television needs as well as becoming the default data transfer network for 3G devices when indoors?  Well, perhaps so, but keep in mind the findings of UK analyst Robert Kenny—who has recently performed some work for the DBCDE—and thus, cannot be dismissed as a partisan stooge as one parliamentarian recently attempted to do in a committee hearing.

Kenny used ABS stats to find that the monthly data usage of Australia’s fixed networks in late 2010 was 67 petabits. Assuming an average peak speed of 12Mbps, the monthly capacity of each line was 4TB. Multiply by 5.5m lines and the total potential fixed network capacity was 21,236PB. With 67PB utilisation, the implied overall network utilisation rate was 0.32%.

This, my friends, is why some economist and commentator types get quite upset when governments pay operators to turn off 100Mbps capable broadband infrastructure and issue press releases that in representing the level of water in the broadband penetration glass, misrepresent its flavour and temperature.


Re Grahame Lynch’s column in Communications Day last Friday. Thankyou Grahame for highlighting the discussion which has already wandered far from my original point, which was that the NBN won’t improve the OECD statistics for fixed connections greatly, or our ranking compared to other countries. Perversely, the continued deployment of the NBN is likely to see our ranking in wireless increase as the fixed-wireless and satellite portions bring broadband within the reach of regional and rural people that don’t have anything today. Our fixed-line ranking will barely shift, as DSL and cable connections are substituted by fibre, leaving the total fixed connections unchanged. Indeed, it is possible the fixed-line statistics might decrease, if people with poor DSL today switch to a better performing fixed-wireless connection.

Where we seem to disagree is the significance of the higher wireless ranking. I wonder if you aren’t trying too hard to see nuances and complex reality of usage and customer value perception that just aren’t there. None of this, or anything else in the OECD report, says anything about the relative value we put on fixed or wireless connections, or the value of the traffic exchanged over them.

The reason why fixed-line penetration has flattened could be many—perhaps everyone who wants and can get a connection now has one, and the current penetration is the best we’ll get in the footprint available – no point in forcing a broadband connection on someone who doesn\’t see a use for it. Perhaps the fixed-line footprint hasn’t expanded at all—these stats only go to Dec 2011, so much of the fixed-line expansion (and I doubt it is significant in national percentage terms) in regional new DSLAMs enabled by the Backhaul Blackspots program and the top-hat program won’t be in these statistics anyway – we\’ll have to wait for the next set to see those.

All it really says is that we are a nation obsessed with having lots of mobile devices. The OECD wireless stats include not just dedicated mobile data devices ( USB dongles, iPad SIMs and the like) which make up 46% of the numbers, but also standard data-capable phone plans (53%), including prepaid SIMs whether or not they are ever used to access the Internet.

I wonder if we aren’t somewhat overthinking the nuances buried in the OECD report. Just because an iPad user looks at email headers, or a handset accesses Facebook or maps from the train once in the past 3 months, doesn’t make it the equivalent of a single fixed-line connection that drives many devices inside a home or business. The value put on looking at email text while on the move, and downloading the attachment when connected to a fixed-line, is different because the uses are different—and equally valid. Neither is a replacement for the other.

Like reading tea-leaves, different people will interpret different patterns from something that may have no significance at all. In any case, and back to my original point—if we want to increase our relative fixed-line ranking compared to other countries and I’m not convinced that’s a meaningful thing to do, it won’t be by simply building the NBN.


Having read Grahame Lynch’s editorial in Communications Day (Friday 20th July 2012), I noted that he posed a number of questions. Here are answers to just a few of them..

Q: Why has fixed line penetration largely flattened in recent years even as the network has increased its reach through rural investment and the Top Hat installations?

A: Two parts to the answer:

1) Rural fixed-link broadband investment has been relatively limited because it has been impacted by the chequered history of such schemes with which we’re all familiar. The reality is that these are low population density areas and hence the absolute number of customers served is correspondingly limited. The Agile/Internode deployment of licensed WiMax in regional SA services around 1200 customers over a circa 20,000 sq km region, is something I\’m very proud of. While these services matter a lot to those who have been able to obtain improved broadband because of them, these numbers are clearly are dwarfed by metropolitan customer base size.

2) Top Hat installations have very little impact on penetration because the Top Hat installations are an upgrade of an existing DSL coverage area only. By and large they’re being done to address technical limitations in existing RIM cabinets (insufficient backhaul and no support for ADSL2+), and hence they are simply Telstra catching up (remarkably late) with delivering ADSL2+ to all of their existing coverage areas. They are not increasing the geographic coverage of broadband in the process. The very name (Top Hat) being the point: Its an overlay on existing cabinets only.

The point about fixed line broadband is that penetration has plateaued because the market is saturated. The issue is now one of what the market is saturated with; Australia has never had a purpose built broadband network, with existing broadband being an opportunistic re-use of networks built for other purposes (copper for voice and HFC for pay TV). Best-effort DSL often fails to meet customer needs for reliability or speed, and the achievable speed is inequitably delivered as a function of copper line length, creating a society of haves and have-nots, in terms of whether the broadband service available at a given location is sufficient for current or future needs.

One of the great benefits of the NBN will be consistency of access speed for all customers. So it’s not the NBN to solve issues of such import as the digital divide.

So it\’s not just about higher speeds, but also very much about the benefits of consistent and reliable service speeds for all customers. Application developers can’t deliver applications to Australians that rely on HD-video grade data speeds, if only a fraction of that existing saturated broadband coverage area are technically capable of accessing the speeds required at the reliability level needed. So the current market is saturated with the best it can currently get. The NBN is needed (in part) because that current best isn’t even nearly good enough for the long term. No sensible proponent of 3G/LTE networks is proposing to turn off the existing DSL network, because there simply are not enough megahertz of spectrum available to deliver that many megabits today—let alone tomorrow!

Q: Why would consumers so readily adopt the 3G versions of devices such as iPads that are also available in WiFi only modes?

A: Clearly because having the fallback to 3G is useful for roaming operations, but it’s clear that the heavy lifting tasks (substantial downloads of any sort) are both more effective and far more economic where WiFi is available. There is no doubt that people will pay more for the convenience of being mobile (whether for voice or data), but there is equally no doubt that WiFi delivers a huge benefit in network offload for 3G operators on a routine and increasing basis. The benefits of that offload (via both WiFi and via Femtocells for those operators who don’t have a religious objection to them) are enormous, and that offload is delivered pretty much automatically. Its a win-win when that offload occurs.

There isn’t a battle here between fixed and wireless! Rather, there is an emerging symbiosis. The clear answer to whether wireless or fixed will win is clear – they both help each other win, with the logical end state being a very large number of very short range, very high speed radio cells (WiFi, Femtocell and similar), using ubiquitous high quality reliable backhaul (via the NBN). This deep penetration fixed network that seamlessly supports a very large number of endpoints genuinely presents the best of both worlds, and delivers an outcome impossible with only one or the other.

Q: With 67PB utilisation, the implied overall DSL network utilisation rate was 0.32% (implicit question: why upgrade anything if what we have is barely utilised)

A: This approach to arguing I presume against the need for the NBN is driven by a misuse of statistics. Average theoretical total utilisation of every tail circuit all at once is a meaningless figure in this context. I expect the average utilisation of the entire Australian existing road network, including every side road and every driveway, is at a similarly low level and is a similarly meaningless number. None of these numbers do anything for someone stuck in standstill traffic on the way to work, or for someone who can’t stream an HD video because their DSL service can barely manage 3 megabits per second and stops working entirely when it rains.

The bottom line is that Australia is on the cusp of building a genuinely world class national broadband network that will have far reaching benefits for all consumers for many decades to come (providing it does *get* built of course). Meantime, arguing that the existing network is good enough just doesn’t cut the mustard. It isn’t. Sitting on our hands, as a country, and using inappropriate statistics to say (in effect) that don\’t need to fix what we have is as silly as relying on such statistics to claim that we don’t need more road capacity because, on average, we have more than enough already. Averages don\’t mean much here at all. What matters, again, is network behaviour at peak demand and appreciating that the peak demands rise non-linearly over time. I’d hope you aren’t implying (via averages) that everyone with broadband today has enough speed for an arbitrary future period. That’s right up there with saying that 640K of RAM is more than enough for a personal computer.


Let me start by saying I agree with many of the points made by Paul and Simon, especially regarding the likelihood that the NBN won’t improve genuine fixed penetration levels, that fixed and wireless networks are symbiotic and that there is a need for investment in improved access speeds on the fixed network. Clearly with the bulk of Australia’s fixed and mobile subscriptions supplied by the two same companies the two markets cannot be artificially separated.

My original contention was that Australian public policy exists in a parallel universe where everything seems to be viewed through the prism of the fixed network and the apparently essential requirement for the NBN to solve issues of such import as the digital divide.

Paul makes several references to wireless as “expensive”, used “intermittently”, switched on “once every three months” and so on which certainly hints at a certain attitude towards its usage even as he later describes both light and heavy usage of the net as “equally valid”.

I actually suspect that wireless is soaking up some of the low end demand for broadband access—for example, Vividwireless’ targeting of low income and itinerant university students, people with access to broadband at their work who don’t want to then make two further personal subscriptions to a DSL and HSPDA service and thus, opt for the latter and so on. Similar observations have been made in the US where it has been identified that lower socio economic groups are more comfortable with mobile than desktop devices and, consequently, universal access policy priority should be afforded to wireless.

Hence the great priority in the US, reaffirmed as recently as Friday with an affirmation to open up a sweet 1000MHz in shared spectrum, on wireless as the means to bridge the digital divide. And perhaps in promoting government 2.0, more attention should be paid to Android apps and less to NBN apps. The NBN seems custom built for power users, but not so much for others.

In recent years there has been a succession of Australian policy decisions designed to promote the quality and reach of fixed broadband such as forbearance on Telstra’s rural ADSL2+ upgrades which expanded the product to millions of people as well as a continued emphasis on unbundling and price regulation to stimulate demand.

This continues on with market moves such as Telstra’s Top Hat upgrades, the massive reduction in international transit costs, the continuing creep of competitive DSLAMs into more ESAs as their costs decline and, of course, the availability of higher speeds via the HFC DOCSIS upgrades and so forth. All these make fixed broadband a more attractive product, usually at a lower price, but, demonstrably, have had little effect on the overall penetration rate, even as household sizes slightly decline and national income surges.

By contrast, there has been very little public policy focus on wireless broadband other than some sporadic grants for such things as Adam Internet’s and Internode’s smallish WIMAX networks. There have been no significant spectrum allocations since the great wireless broadband explosion even as the medium is regularly dissed or ignored by central actors in policy circles as unfit for purpose.

But again market trends with the increasing proliferation of smartphones and tablets, combined with the intense infrastructure competition that has pushed access speeds up to tens of megabits and prices down (Telstra’s wireless broadband ARPUs dropped by over 25% to just above the $30 a month mark last year) have seen Australia emerge as a wireless over-achiever.

Let us not forget that one of the greatest constraints on wireless performance is limited spectrum and that this spectrum is artificially constrained by government so as to maximise its revenue-raising potential. The fact that one’s wireless broadband connection shares a score or so of expensive megahertz with hundreds of other users per sector as opposed to scores of users on hundreds of free or cheap megahertz is a function of spectrum planning influenced by tax collectors (and town planning), not a function of physics (just as the NBN’s planned virtual connectivity circuit is designed as a usage pricing mechanism and isn’t the result of optical science).

Ten years of public policy focus and now multi-billion dollar subsidy of fixed broadband and the result? Apparently still an underachiever in world rankings and, if Paul is correct, likely to remain that way on raw fixed penetration levels. The simple reality is that millions of Australian premises are electing not to connect to fixed broadband.

I return to my original point that Australia’s telecom policy making elite seemingly dwells in a parallel universe. To wit, simply read the ACCC’s justifications for approving the NBN-Optus deal which sets the rather bizarre precedent that an infrastructure monopoly, especially one that pays $2,000 per customer to shut down a rival, is a desirable model of efficiency and consumer welfare. If David Thodey offered Paul O’Sullivan the same he would face up to 10 years in prison.

To borrow from Robert Kenny’s methodology we are paying Optus to close a network with an aggregate 700 gigabits per second of end user access capacity—equivalent to Australia’s total international bandwidth usage. I find this utterly surreal.


COMMENT: Australia\’s parallel telecom universe problem

Late yesterday, the Organisation for Economic Co-operation and Development issued fixed and wireless broadband statistics for its 34 member countries—essentially the nations of Western Europe and North America with some Asia Pacific representation in the form of Australia, New Zealand, Japan and South Korea.

The number one takeout is that there are substantially more wireless broadband services than fixed broadband services across the 34 developed nations: 667m wireless broadband subscriptions versus 314m fixed broadband subscriptions or about twice as much.

The number two takeout, if one is Australian, is that this country does spectacularly well in the larger wireless market: it ranks 8th in the world for wireless. Australian wireless broadband subscriptions at 16.6m outnumber fixed broadband subscriptions at 5.49m. This is about three times as much.

What might a communications minister do in these circumstances?

Well, he might issue a press release congratulating the industry on its fine effort in bringing wireless broadband to the equivalent of 75% of the population. He would likely cite the amazing and extensive personal productivity benefits that accompany this fantastic adoption rate of technology. And he would almost certainly be entitled to give himself a pat on the back for presiding over a supportive regulatory regime that has allowed all this growth to prosper.

So what does the communications minister, instead, actually do?

He issues a press release that completely ignores Australia’s top ten wireless status and instead focuses on the drifting fixed broadband network penetration rate (21st, down from its mainstay level of 16th through 18th over recent years) as vindication of the need for a national broadband network, which will actually have a smaller fixed footprint than the current reach of the DSL network. (Of course, there being no voice-only services on the NBN, all its subscribers will be counted as broadband users whether they use broadband or not, which should at least help that OECD number in a few years).

Last night the Coalition did what Conroy regularly did when he was in Opposition: mock the government, in this case comparing Conroy to a medieval doctor applying leeches to a wound!

My initial reaction to all this is: who cares? When we are going to get over this facile idea that wireless broadband isn’t proper broadband and that fixed is the real deal?

Average occupants per household in Australia is 2.6. There are over 3 wireless broadband services for every fixed broadband service in Australia. The market is talking.

What’s more, it is likely that the typical Australian probably spends more time with their mobile broadband device—at work, in transit, at home, even in bed—than the fixed broadband service at their home or work. Certainly ARPUs and overall spend suggest that people place a greater value on mobile broadband than fixed broadband; the former represents a much more substantial economic sector, with actual growth. Yes, fixed users download nine times more data than mobile users do: but does this data have the same value per bit? In terms of revealed preference of actual spend per bit, wireless wins.

How much of that fixed data download is economic junk? Stolen movies and music? Bloated hundred megabyte updates to patch badly written PC operating systems? Unwanted pop-up ads, video plays and malware? Fixed BB communications appears to be the only field of human endeavour where obesity is a virtue. Compare this to the stripped down utility of the wireless medium: lean 20MB 3D games & 2MB apps, compressed 100k photos, the elegant economy of 160 character tweets and SMSs, the invaluable utility of a medium that is largely by the recipient’s side for every minute of their waking day.

And unlike with fixed, there appears to be no great universal service or capital investment problem with mobiles.

The wireless operators definitely put more into the Treasury coffers than they take out, very much so with spectrum auctions around the corner: the same cannot be said for fixed operators at the moment who are getting the best part of $40 billion of state-administered funds spent on them at minimal risk. The NBN may well revolutionise society as we know it but the evidence where fibre and advanced wireless exist side-by-side in pervasive supply would suggest one shouldn’t bet on it: wireless remains the more exciting story in Japan and Korea.

Now, none of this is to diminish the exciting future for NBN services.

Yet, we should appreciate the reality: Australia is not only a wireless nation but a world leader and it’s about time we celebrated this instead of ignoring it. If the minister can’t congratulate the industry on such a great outcome then this publication will.

Grahame Lynch 

COMMENT: Bad timing for proposed US undersea cable tax

The Federal Communications Commission plan to expand the scope of universal service fund assessments to cover foreign submarine cable operators including Southern Cross Cable and Pipe Networks could not come at a worse time for the US.

Right now the US needs every bit of international prestige it can muster to resist the push at the International Telecommunications Union for changes to a global treaty which could profoundly alter the basis on which today’s Internet is run.

The FCC is seemingly seeking to subsidise domestic universal service spend with proceeds collected from the pockets of foreign carriers and their end customers in the likes of Fiji and Peru. Just a little over a decade ago it was the same FCC which unilaterally dismantled the practice of inflated termination charges used by developing countries to boost the coffers of their domestic telcos—and by definition, fund capital expenditure and universal service programmes. In 1999, the FCC—heeding the arguments of US telcos paying out billions to foreign telcos to terminate their calls—effectively mandated the price they would pay at between 15c and 22c a minute.

Foreign telcos, some of whom charged multiples of this amount, were outraged at the presumption that the FCC could name the price that they charged for access to their networks. But ultimately the FCC prevailed—US carriers re-negotiated their rates downwards, the ITU reacted with a face-saving compromise that capitulated to the US position, and foreign telcos simply had to live with sharply reduced income from what was still their biggest source of international traffic.

One shouldn’t underestimate the angst this caused.

Soon after, Philippines telecom executives visiting a Pacific Telecommunications Council conference in Hawaii were literally subpoenaed from their Wakiki Beach hotel rooms to answer Department of Justice charges that they were ripping off US telcos. Bad memories linger! The hypocrisy of the latest FCC move will truly stick in the craw of many of those around 13 years ago, especially now they are likely to hold more senior positions of authority in their home countries.

As I say, this comes at a time leading up to a crucial test for the future of the Internet: the December World Conference on International Telecommunications which seeks to ratify a new treaty for international telecoms. Current drafts could have profoundly negative effects for the net as we know it. For example, there is the European telcos’ proposal to impose charging on Internet traffic, effectively adding imposts to the likes of Google and Facebook.

ILLIBERAL: Various illiberal proposals from the Middle East, China and Russia would provide legal cover for crackdowns on freedom of expression over the ‘net and the inspection of packet content as well as the transference of Internet governance from currently effective private bodies to government bodies. There is even a proposal to make such changes binding on members. There are currently only a few lone holdouts against these proposed changes: the US, Australia and Portugal. It will be quite difficult for the US to argue against a European internet content tax if it concurrently proposes a tax on foreign Internet bandwidth sales. It will be even harder for the US to maintain its centrality as the world’s bandwidth hub if all these things come to pass.

As for the submarine cable operators themselves, it is unclear what actual effect the USF levy may have on their finances. It is true that many of them have razor thin margins and an implied 15% tax could wipe them out if they were forced to absorb it.

But it would seem many contracts currently in force insure submarine cable operators against new taxes—they can be passed on to telco customers. Given international bandwidth accounts for less than 10% and probably as low as 7% of ANZ retail broadband costs, the additional USF levy would have a minimal effect on end user prices.

Trans-Pacific bandwidth and transit prices have dropped dramatically in recent years: Australia-US megabit per second prices have fallen from $135 to below $25. One contact tells me that a large capacity purchase completed in recent weeks effectively prices US connectivity at $8. Clearly with such disruptions in unit pricing and large volume growth, the impact of a revenue levy might be less dramatic than it seems. Add in the depreciating US dollar and the overall cost profile of delivering a broadband service in Australia is still heading downwards, levy or not. Yet things might not so be comfortable for the Pacific Island, Asian and Latin American nations that would also be affected.

The levy would likely be more problematic for new builds. Operators such as Southern Cross Cable, Pipe and Global Crossing are well established and have either paid off or nearly paid off their capital build or acquisition costs. A new venture such as Pacific Fibre faces the prospect of paying a levy for every dollar it ever earns, putting it at a competitive disadvantage.

In reality it is hard to see the FCC sticking with this. As law firm Wilshire & Grannis point out, “These proposals, if adopted, could also encourage Internet content providers including online video providers to shift content creation and storage outside the United States particularly as they are typically treated as end-users for USF assessment purposes, perhaps abetted by the spectacular growth of content delivery networks.” Throw in the complete spanner this throws into the US’ international negotiating works regarding the ITU treaty and it would be amazing if the FCC proceeds with this idea.

Grahame Lynch 

ANZ-US undersea cable operators may be hit with 15.7% universal service levy

A coalition of undersea cable operators including Southern Cross Cable and Pipe Networks is lobbying the US Federal Communications Commission against proposals which could see an American universal service levy currently set at 15.7% of revenues applied to their cable operations.

They have warned that the levy would slash their thin operating margins or force them to pass on the costs to existing customers.

The coalition, which also includes Level 3-owned Global Crossing and French Polynesia’s OPT, says the FCC is currently seeking to reform outdated universal service rules which exempt international services, mainly to counter avoidance of the levy by American calling card providers.

But the proposed reform would capture international undersea cable operators in the FCC’s ambit, effectively leading to Australian, New Zealand and other international internet users contributing to American universal service funds. Likeliest to be hardest hit would be new cables such as the proposed Pacific Fibre link: older cables such as Southern Cross have recovered their build costs.

The coalition of cable operators has retained respected Washington DC law firm Wilshire & Grannis and its leading international telecom partner Kent Bressie to represent its views to the FCC. In comments filed with the FCC and released late last week, the firm warns that the FCC’s actions would contravene previous legal findings on the issue as well as place the United States in violation of its World Trade Organisation commitments.

“Elimination of the international-only exemption … would eliminate operating margins for international undersea cable operators, which have, at best, a limited ability to recover their costs by passing Universal Service Fund assessments through to customers, particularly for those located outside the United States. Operating margins on many routes are already razor thin, and unsuccessful recovery of USF contribution costs from even a minority of customers could turn certain international undersea cables into loss-making enterprises,” the filing warns.

“Customers outside the United States consistently object that “domestic assessments” such as regulatory fees and USF contributions cannot and should not be passed through to such customers, but instead absorbed into the operator’s administrative overhead. This is particularly true when such charges come due long after the economic basis of the long-term capacity purchase is negotiated. Attempts to pass through such charges create tensions in the operator-customer relationship, and operators must expend significant personnel resources to persuade customers to accept and pay such charges, typically with little success,” the filing continues.

Wilshire & Grannis says that most IRUs and other bandwidth lease agreements do not provide the ability for operators to automatically pass on government-imposed fees, particularly given that over the past 15 years it has been repeatedly affirmed by the FCC and the Supreme Court that international-only provider are exempt from such levies. “Renegotiation of these agreements would be a major undertaking, as the operator would be seeking to alter fundamentally the economic terms of long-term arrangements that were intended to secure the supply of capacity at a known price,” the filing says, contrasting this with the intended target of the reform—calling card companies—who can simply raise their retail charges to absorb the cost.

The firm also warns that other countries may react with reciprocal actions. “If other countries were to make similar assessments on the same revenue streams for international services, such assessments could quickly equal the total revenues for an undersea cable system. For example, Level 3’s South American Crossing system lands in five countries (Brazil, Argentina, Chile, Panama, and Peru) in addition to the United States. Southern Cross lands in three countries (Australia, Fiji, and New Zealand) in addition to the United States. As capacity is often sold on an end-to-end or ring-configuration basis, such services would arguably fall within the regulatory jurisdictions of each of these countries. There is simply no way for international undersea cable operators to support multiple universal service assessments for services provided far beyond any one country’s territory.”

The expanded levy could also potentially damage America’s standing as the world’s central Internet hub. “[The] proposals could deter new cable landings in the United States and even encourage operators to land in other countries, particularly Canada. These proposals, if adopted, could also encourage Internet content providers including online video providers to shift content creation and storage outside the United States particularly as they are typically treated as end-users for USF assessment purposes, perhaps abetted by the spectacular growth of content delivery networks. Such outcomes would harm the telecommunications, Internet, and entertainment sectors of the U.S. economy, not to mention national security.” The filing notes that Canada’s universal service levy, at 0.66% of revenues, is considerably less onerous than the 15.7% currently charged in the US.

Sources close to the process suggest that the FCC had not anticipated the impact on undersea cable operators when it drafted the proposed USF revisions.

Combined with the highly negative effect the proposal would have on the US’ position as the world’s Internet hub, CommsDay understands that the cable operators are confident they can persuade the FCC to retain the exemption on international services. But conversely, there is an “insatiable” demand for universal service funds and the FCC may elect to press on with widening the base of contributing firms. Submarine cables landing in the US have previously been subject to international bearer charges so the idea of taxing foreign infrastructure that terminates in the US is not unprecedented.

Grahame Lynch



Aust gov’t softens stance on carrier security measures, but still targets some foreign vendors

The Australian government has softened its proposals for a package of telecoms infrastructure security reforms after industry consultation, shifting to an outcomes-based framework that would be less prescriptive on administrative and technical requirements.

But the latest version of the planned measures – aimed at “mitigating the risks posed to Australia’s communications networks by certain foreign technology and service suppliers” – could still see carriers obliged to demonstrate “competent supervision and effective controls” over their network. Failure to do so would invite penalties and directions from the Attorney-General’s Department. And those directions, while they might need signoff of the ASIO director general of security and the secretary of the Department of Broadband, Communications and the Digital Economy, could in turn see carriers compelled to change even existing infrastructure – at their own cost.

An extensive discussion paper prepared by the AGD and released by the parliamentary joint committee on intelligence and security – which has now commenced its inquiry into the planned reforms – publicly lays bare the current iteration of the plan in unprecedented detail. It confirms that the AGD originally consulted with industry on the proposal earlier in 2012, as exclusively reported in CommsDay on 30 March.  However, it also indicates that the proposal has subsequently been toned down somewhat.

“During the consultation about a possible regulatory framework that originally included a notification obligation in place of the requirement to provide information to government on request, industry expressed a preference for an approach that avoids the need for government approval of network architecture at a technical or engineering level and instead focuses on the security outcome, leaving industry to choose the most effective way to achieve it,” said the paper. “As a consequence an alternative regulatory framework designed with less focus on administrative processes and technical requirements, but greater emphasis on outcomes, has been developed for consideration.”

PROPOSAL FOR ACCREDITATION? The AGD added that government was now considering the means by which it might be assured that industry was taking reasonable steps to address risk – potentially including “accreditation of industry for self‐assessment purposes or a role for third parties in providing audit and assurance services.”

The main thrust of the proposals, though – to harden the security of Australia’s telecom infrastructure with a risk-based regulatory framework, established via amendments to the Telecommunications Act and/or other legislation – remains unchanged. “There is a lack of awareness of national security risks in business decisions by many C/CSPs, which means engagement often occurs late in the decision making process,” said the paper. “Government is concerned that the telecommunications industry is not fully informed about national security risks and is therefore not equipped to respond adequately to these risks. As both businesses and consumers are also exposed to the consequences of potential security risks, there is a compelling case to act now. Australia is at a critical stage of telecommunications infrastructure development driven by the NBN’s construction.”

The framework mooted in the paper would include an industry-wide obligation on all carriers and carriage service providers to “protect their infrastructure and the information held on it or passing across it from unauthorised interference to support the confidentiality, integrity and availability of Australia’s national telecommunications infrastructure.” This would require them to demonstrate a level of oversight of their network operations and location of data, as well as direct authority to protect their networks from unauthorised access – for example, by repatriating information and systems in response to such access. It would also contain a requirement for C/CSPs to provide government, when requested, with information to assist in the assessment of national security risks to telecommunications infrastructure; and powers of direction plus a penalty regime to encourage compliance.

The current suggestion is that the power of direction would be used after first attempting direct engagement with carriers who did not toe the line, and could require the concurrence of the AGD and DBCDE as well as the director general of security. Once signed off, however, “directions could involve targeted mitigation or remediation of security risks, including modifications to infrastructure, audit, and ongoing monitoring, with costs to be borne by the relevant C/CSP.”

“Should any legislative changes be agreed, this would require all C/CSPs to comply with the security obligations. In some instances this will require the application of mitigation measures to existing infrastructure. The security obligations would apply to existing and new infrastructure,” said the paper. “Government recognises that it would need to work closely with industry to ensure that there is a reasonable transition period.”

INTERCEPTION AND DATA RETENTION: The scope of the proposed reforms is not limited to tightening security for telco infrastructure. The other elements of the proposal package are aimed at “modernising lawful access to communications and associated communications data” and “enhancing the operational capacity of Australian intelligence community agencies.”

On the first point, the government is seeking the views of the Committee on expanding the basis of interception activities, establishing an offense for failure to assist in the decryption of communications, and instituting industry response timelines. It has also raised the spectre of data retention, asking the Committee to consider the possible application of “tailored data retention periods for up to 2 years for parts of a data set, with specific timeframes taking into account agency priorities and privacy and cost impacts.”

Submissions on the paper and terms of reference therein are due by 6 August.

Petroc Wilton 

Telstra reveals media strategy, plans changes at Sensis

Telstra Media GMD Rick Ellis is pushing a fundamental change through the firm’s content strategy – intent on transforming its digital media and IPTV arm into a profit-and-loss focused business in its own right.

And Ellis is confident that the fading revenues of the Sensis division, which also falls under his broad remit, will be on the road to recovery within two years.

Previously the CEO of Television New Zealand, Ellis was announced late last year as the head of the newly formed Telstra Digital Media, managing Sensis, BigPond, TradingPost, IPTV, Foxtel and other content arrangements; after beginning his term this January, he renamed the division ‘Telstra Media’ to reflect its mix of physical and digital assets. He’s also been serving as acting CEO of the Sensis business since the departure of long time incumbent Bruce Akhurst.

At a media briefing in Sydney, Ellis outlined the transformation he’s realising in the IPTV and digital content section of his portfolio. “We’re changing the model from a telco model, which is largely one where… investment in content is justified on the basis of the pull-through benefits of market share shift and churn reduction, to actually positioning that business as a P&L business in its own right,” he explained. “We’re changing the way that we think about users, market segments, communities, and… what content we’re going to package in which way to meet those market segments. We’re going from a product-centric approach, which is a telco approach, to a media company approach where we’re focused on viewers, users, audience, market segments and maybe what their expectations are.”

That shift will inform Telstra’s approach to content, whether licensed – Ellis noted that the telco’s AFL rights had resulted in 901,000 mobile video streams between March and May, up 800% year-on-year – or generated internally, such as via content production subsidiary Chief Entertainment. It’s the reason why Ellis has recently brought on board Adam Good as the new head of digital content, and will underpin its search for a new head of IPTV following the departure of Ben Kinealy.

The change could also see Telstra further decouple its digital content from its network assets. “The paradigm [within which] we’ve approached this business in the past, which is a telco paradigm, is up for serious review,” he said. “Taking our content assets, where we’re able, beyond a Telstra network connection is certainly something that we’re considering.”

But while Telstra is considering multiple delivery models and already offers Foxtel, for example, through its own T-Box platform, Ellis said that pure over-the-top IPTV players were unlikely to become a serious threat within the broader media industry. “There are a number of commentators who say the over-the-top players are going to kill local pay TV and free-to-air businesses. And I just don’t believe it, and I just don’t see it,” he said. “There were some interesting stats out two weeks ago that showed the OTT video on demand revenues in Europe were less than 1% of the gross revenues from pay and FTA TV… and then I look at a Netflix business, and I apply the same metrics of Netflix to an Australian market with eight million households, and it grosses up to about a A$90 million business. Who cares?”

“OTT is internet, and internet doesn’t deliver you ten or a dozen HD channels just at the flick of a switch when you turn your telly on. And even in an NBN world, the estimated cost of delivering HD linear channels to a television set is going to be like five times [that of] cable, satellite or digital terrestrial. I think OTT players, in the main, add an extension to core video consumption, which tends to be watched during primetime… I don’t think [they’re] going to change that at all.”

FOXTEL LINKS: A similar logic explains why Ellis believes that Telstra’s new media strategy won’t push it too far into competition with Foxtel, of which it is a 50% shareholder.

“I see Foxtel addressing quite a different market to the one that we’re addressing with our IPTV service,” he said. “If you think Foxtel is just a lot of channels delivered at a relatively low household cost over cable or satellite, there are constraints, obviously, in the IP world, whether it’s ADSL2+ or even cable, for that matter.”

“What we’re addressing is a solution for a household that only wants to spend A$20-40 a month as opposed to A$120 a month, to get more stuff than they would over digital free-to-air but not the sheer length and breadth that you get with a Foxtel package. And so, even if you think about Foxtel going to 50% of households, then [in] the other 50%, there’s a big chunk of those looking for something more, but not as much as [Foxtel]. And that’s where we’re targeting… [although] where we can leverage Foxtel content into that space, we will.”

RESUSCITATING SENSIS: Meanwhile, Ellis also expanded on his plans to turn around the fortunes of the troubled Sensis division and “transition from a directories business to Australia’s leading directories and digital marketing service.”

“The company\’s strategy around offering a value package which was a combination of print and digital has actually been remarkably successful, and at the same time the company’s continued to invest in its digital-only sales and fulfilment capability. And the uptake in that space has also been very pleasing,” he said. “We go into the next 24-month period [with] a clear line of sight to that transition, a clear line of site in terms of the types of products and services that we need to provide to SMEs in both metro and non-metro territories… and I’m confident that we’re going to see this transition through successfully.”

Ellis expects the ongoing increase in digital revenues to offset the decline in print by 2014, with a return to growth thereafter. He is currently searching for a new head of operations for Sensis, and for a head of digital partnerships and innovation – a role central to his vision for the business. “This whole place that we’re going to now is not a proprietary place; we need to be wherever users are going to look for information about a product that we want to secure,” he said. “To get to those places – just think Google and Facebook for obvious ones – we need to make sure that we have strong relationships with those sorts of organisations, and that the interface between our environment… and their environment as, effectively, the publisher is as seamless as it can possibly be.”

Ellis is also still in the process of recruiting a Sensis MD to replace Akhurst; he expects to make an announcement late this month or early in August.

Petroc Wilton

Telstra smart controls could extend to Big Pond, SMS

The controversial filtering system that Telstra has been trialling on its Next G mobile service is the first stage of wider plans the carrier has had to offer parental controls as a new fee-based service. CommsDay has learned that a follow-up in the Smart Control project looked to go beyond URL/web filtering and allow greater mobile device controls, such as parental controls on SMS traffic, while a similar URL filtering system was also being considered for its BigPond internet service.

Inside sources have also suggested that the trial was always intended to be conducted on an opt-in basis, but that problems with a separate policy management system in Telstra’s mobile network meant that it was unable to conduct the opt-in trial.

Telstra would not confirm whether this was the source of its problems.

Telstra had been trialling the URL filtering system from Canadian security firm NetSweeper but has attracted a barrage of criticism over privacy issues. The order to temporarily put a halt to the trial came from CEO David Thodey directly, according to sources.

Telstra has since told customers that it now needs to do more consultation before launching the service. A Telstra spokesperson said that it “should have consulted more with our customers prior to introducing the smart controls product.”

According to one person close to the project, Telstra had initially looked at an application-based internet filter similar to what Vodafone has already launched. However, the management team decided that a network-based service would be more flexible.

While the trial was initially supposed to be opt-in, the policy management problems meant it was unable to configure its network software to allow for this. Telstra had also wanted to build up a cache of known links that the system uses to filter out various web sites that it hasn\’t encountered before, given that the Netsweeper system has not been used in Australia extensively.

Geoff Long 

Facebook becomes a telco: invests in Asia subsea cable

Facebook has made its first investment in a subsea cable system: the Asia Pacific Gateway. According to a Malaysian stock exchange filing by Malaysia’s Time dotCom, the latest member to join the APG cable consortium, Facebook is already among the list of consortium members behind the project.

Facebook’s investment, which other industry sources have confirmed to CommsDay, would be its first direct involvement in a subsea cable project, and would follow a similar move by Google with its Unity cable investment across the Pacific.

Few details are available on Facebook’s actual investment in APG, but initial accounts of the system’s capabilities should give the company some terabits’ worth of pan-Asian capacity. For example, Time dotCom’s investment of US$45 million, which includes the cost of building a cable landing station, already gives the operator 3.4Tbps of capacity between Malaysia and Japan.

While Facebook keeps most of the information on its global network infrastructure behind closed doors, it has in recent times disclosed at latest some of its network purchases. Earlier this year, the company announced that it has tapped Teliasonera International for a terabit-level European network to support its Swedish data centre.

CommsDay understands Facebook currently has leased capacity on multiple pan-Asian subsea systems. Whether or not these contracts will be replaced by APG remains to be seen, as diversity will be needed to support APG’s linear design.

On the other hand, the investment could signal further expansion into Asia by Facebook. While the firm has previously denied news reports that it was planning a new data centre site in Taiwan, with 194 million of its 850 million subscribers in Asia as of March 2012, the prospect of Facebook investing into its own facilities in the region would not take a great stretch of the imagination.

In comparison, Google, which in addition to Unity is also an investor in the SJC pan-Asian cable, is building three data centre sites in Asia.

Other members of the APG consortium include China Mobile, China Telecom, China Unicom, Chunghwa Telecom, KT, LG Uplus, NTT Com, StarHub, Viettel and VPNT. NEC has been awarded the construction contract for the APG cable, which will link Korea, Japan, China, Hong Kong, Taiwan, Vietnam, Singapore, and Malaysia and offer a design capacity of 54Tbps using 40G waves upgradeable to 100G.

Tony Chan