Australia will be an interesting test market for VDSL. With a new government and the broadband infrastructure company NBN Co basically in agreement, it is most likely that VDSL will be used to bring for example fast broadband to multi-dwelling units (MDUs).

It was mainly for political reasons that the previous government stopped NBN Co deploying the VDSL technology in MDUs for this purpose. Whether or not any more VDSL will be deployed beyond that will largely depend on the NBN Co review of its current plan. If there indeed is not a cost blowout, and if the timeframe can in fact be maintained, there is a good chance for the majority of that plan to survive.

Obviously another issue that will need to be addressed here is customer expectation. Will people in MDUs accept that they will receive a VDSL connection rather than the FttH connection they were promised under the previous government?

VDSL an interim solution

The principal consideration of VDSL and its vectoring option continues to be that it is an interim broadband technology and will eventually have to be replaced by fibre-to-the-premises, FttP, an assessment also supported by the Australian Minister for Communications, Malcolm Turnbull.

But there are other more immediate issues that need to be considered in relation to a fibre-to-the-node FttN rollout based on VDSL.

In any economic and technical sense VDSL vectoring can only be done in Australia by its national telco Telstra, and it will therefore be a monopolistic activity that needs to be regulated. In any case it will not give the government its much-wanted infrastructure competition. It is rather puzzling that this government is hanging on to infrastructure-based utility competition, a policy they had already trialed in the 1990s with HFC and which failed miserably—it only provided 25% penetration, was (for 90%) overbuild, and led to financial write-offs of billions of dollars.

However, I argue that Instead the government should concentrate its policies on maximizing competition on top of that infrastructure.

But apart from these policy issues there are also other considerations in relation to the cabinets that need to be installed for a VDSL rollout.

The aesthetics of the cabinets

Estimates vary greatly on the size of the FttN rollout, but up to 70,000 VDSL street cabinets will have to be deployed for the FttN rollout as it was outlined in the Coalition’s policy document of April 2013, when they were still in opposition.

The size of these cabinets has shrunk somewhat from that of a double-sized fridge, but it is still considerable—something the size of the operating boxes near traffic lights. Where these boxes have been deployed efforts have sometimes been made to make them more attractive, e.g., by planting flowers around them; but many more have proved to be ideal graffiti targets. As with everything, it also helps if the installation can be promoted as a positive development for the community, but in Australia, where the FttN is a backward step from the original promised FttP deployment, it could be difficult to put a positive spin on these cabinets.

Putting aesthetics aside, the more important question that may need to be asked is how reliable are the cabinets. My Dutch colleague, Hendrik Rood, brought the following to my attention…

Reliability and performance of the cabinets

In 2007 BTG, the Dutch association of MNE (enterprise users) questioned the company installing these cabinets, Dutch incumbent telco KPN, about availability/performance of the FttN cabinets within local loop architecture. KPN is rolling out FTTN infrastructure in some cities entirely by itself, while at the time of questioning it didn’t own a share in Reggefiber, which was still a competitor who rolled out FTTH. That changed end of 2008, when KPN acquired 41% in Reggefiber.

Finally in September 2009 KPN, now active in both FTTN and FTTH came up with a presentation based on data collected by TNO (the national R&D company).

The key points of this presentation were:

1. According to KPN the FTTN street cabinets contain two-hour battery backup. According to installing contractors it is, at best, one hour. This compares to eight hours of battery back-up for similar equipment in local exchange buildings and even diesel-back-up in large central offices.
2. Reggefiber/KPN’s FttH AreaPoPs (points of interconnect) batteries get four-hour back-up with a fast replacement service. In the Netherlands, where it is at most two hours (without traffic jams) to drive from central operations headquarters to every corner of the country except the islands; two hours is a fairly feasible window, but one hour is obviously not enough to avoid a shutdown during any serious power outage.
3. With 2-hour battery backup and based on data for power outages (short- and long-lasting) for Greater-Amsterdam, it was calculated that unavailability due to power outages would be contained to 1 per 9 years (instead of 1 per 3.5 years for minor outages of >15 minutes without battery backup).
4. However, for a business with a few hundred branches around the country this still averages out at about an outage per week. And it is also an issue if you run your alarm systems over it etc. All these issues aren’t currently happening with ADSL (because central offices locations have ample back-up batteries and power consumption is declining in those buildings with less and less power-hungry electronics/switches).

Bottomline: while power outages in the Netherlands are much lower than in, say, Australia or the USA, a mean-time-between-failures (MTBF) of 9 years after battery-backup measures is still a considerable burden for a firm with 450 branch offices spread over the country. That’s an average of one outage per week.

Of course you can argue that your PC wouldn’t work etc., but Hendrik argues that one has to be aware that battery backup in homes/offices for critical communications installations (fire and security alarms etc) is becoming increasingly common and it is in relation to these functions that the ‘lifeline’ functionality is essential, and FttH is superior to FttC/FttN as was the old fashioned PSTN/ADSL powered in local exchange buildings.

Issues in relation to savings coming from sharing infrastructure

In the TNO review it also turned out—KPN was obliged to confess—that it would never route its base stations backhaul through street cabinets (not even with G.shdsl.bis or FTTC+fibre extension etc) but would connect passively to the (regional) central offices.

Most of the many thousands GSM/3G base stations of KPN are currently backhauled over either microwave or multiple G.Shdsl (those on rooftops and nearer to central offices) that run passively through the company’s local loop to their central office location. Higher bandwidths for base stations (to support more mobile data traffic for 3G and 4G) will be done either by high bandwidth microwave or passive fibre loops, maybe fused through in the manhole near the VDSL2 street cabinets on the city or regional ring linking those cabinets.

The latter implies trenching a spur directly from the ring that also feeds to the VDSL2 street cabinets towards the mobile towers. This last-mile spur will be full cost—that is, it cannot be connected with fibres from the FttH local loop access networks. This inability to share cost is, of course, more expensive.

In the Netherlands, when an apartment building gets FttH there will be spare fibre loops deployed to enable the connection of a rooftop base station (that currently is typically supplied by G.Shdsl from central office over passive copper to the rooftop installation). The difference between a separate spur and a shared local loop with the FttH plant is between ca. €6000-€8000 per base station (separate spur) and €1000 (passive FttH loop).

So the added cost is, say, 10 thousand base stations x €7000 = €70million (10k is the estimated number of high-bandwidth base stations at rooftop positions needed for high-bandwidth 4G service).

Of course that won’t be the ‘deal breaker’ for lower CAPEX with a VDSL-based access network, but it is one of the hidden costs in choosing VDSL. The other one is raised levels of power consumption, as VDSL2+vectoring is far more power hungry per line (including CPE and DSLAM-cards) than an FttH-based solution.

The risk of short-circuiting

Explosion in VDSL street cabinet of AT&T in US – Big hole blown into the nearby wooden fence (Source: Light Reading)The TNO review also looked into this issue. While it is an exceptional event it cannot be excluded. The energy density of Li-ion batteries is much higher than conventional (lead-sulphur) batteries used in central offices. As a result a serious short-circuiting in equipment can lead to an internal chemical reaction that causes an explosion—such an explosion happened in a VDSL street cabinet of AT&T in the USA. Lightreading collected some pictures including the big hole blown into the nearby wooden fence.

If you want to have small street cabinets and long battery lifetime (e.g., a few hours) you have to deploy dense Li-ion and the risk of a seriously damaging explosion increases.

Of course, short-circuiting isn’t a frequent occurrence in a V-DSLAM, but it cannot be entirely ruled out for any equipment, and with a large installed base of street cabinets the chances grow of one or more short-circuits anywhere in the country over the technical lifespan of this new equipment deployed in street cabinets.

Until now telco operators haven’t had any serious scale experience with the impact of battery-backed-up street cabinets, and neither has CATV (as they didn’t use them because they hardly served SMEs and MNEs over their HFC-plant).

A number of mobile base stations also had relatively short-run UPSs (typically 30 minutes, which causes the mobile networks to shut down at large-scale power outages, despite people walking around with devices that last a day without charging). However this attitude is gradually changing at the operators (and they earn enough money with services to consider a serious battery back-up at the base station site that lasts many hours).

A hidden cost is that a typical UPS-supplier (the 110-230 VAC systems) that last 15-30 minutes (the time needed to power up the back-up diesel generator) charges an annual maintenance fee of ca. US$3000-5000 per device. UPSs are cheap in CAPEX but expensive in maintenance (due to the commercial model), which is why most telcos do the different design with AC-DC converters and batteries.

So, for the total VDSL picture it is worthwhile to consider not only the bandwidth issues, which are widely discussed, but also battery issues and the power supply to the FttC/FttN street cabinets and compare this with the FttH option (without the need for outside plant battery feeding). Another issue that needs to be taken into account is the short-circuit risks in those street cabinets.