NBN 101: The case for wireless broadband

1 2 Page 2
Page 2 of 2

Wireless as it stands

Wireless broadband already plays a significant part of the telecommunications market in Australia. According to the latest figures from the Australian Bureau of Statistics (ABS), wireless broadband subscribers make up 2.8 million of the total 8.2 million broadband subscribers. It's important to note here that the ABS only takes dongles into account at this stage - not smartphones - and doesn't distinguish those users who have multiple connections. However, considering that half of the total 25 million Australian mobile subscriptions in June 2009 were made up of 3G-capable devices, it's likely that the number of connections to the Internet on mobile phones will surpass that of fixed broadband, if it hasn't already.

With Telstra and competing providers pouring an increasing amount of money into developing faster 3G and HSPA networks, it doesn't look like wireless is going anywhere anytime soon.

When it comes to building a national, open access broadband network with minimum committed speeds, though, wireless broadband becomes fraught with issues. The more users you have on a given wireless broadband network, the more base stations and back-end infrastructure a service provider has to implement to keep providing the same quality of service. According to Quigley, Australia would have to expand its current base of 16,000 mobile cell sites to around 80,000 in order to deliver a near-equivalent broadband experience to the proposed NBN.

That doesn't even take into account speed. Telstra's fastest mobile broadband network currently works at a theoretical peak of 42Mbps in lab tests, with real world speeds of up to 16Mbps with the right hardware. In premium conditions, Telstra's dual carrier HSPA network could technically be a candidate for the wireless portion of the existing NBN plan. But given that aspect of the rollout is made up of rural Australian communities it is unlikely a 12Mbps minimum speed can be guaranteed.

Next generation, "4G" technologies such as WiMAX and LTE provide better hope of comparable speeds and coverage, but can they really be fit out for an NBN?

*Note: We will be addressing the uptake and demand argument around whether people will choose wireless over fixed line services in a future article. This article focuses on the technology and their ability to deliver NBN goals.

The case for WiMAX

WiMAX has, so far, been a bit of a non-starter in Australia. Though several Internet service providers (ISP) are expanding their respective WiMAX networks, the technology has suffered from a fractured upgrade path that has essentially split it into three different standards: 802.16d, 802.16e and 802.16m. Like 802.11 WiFi, these different industry certifications provide different capabilities and speeds, but by the same token require different hardware and often significant upgrades to existing infrastructure by both operator and end-user.

The earliest version, 802.16d, is currently available in Australia through Unwired in Sydney and Melbourne, and Internode in South Australia, with varied speeds of between 256Kbps and a theoretical 9Mbps. The service was marketed heavily by Unwired - now owned by Seven Group - several years ago as a fixed wireless broadband access service dedicated for people on the move or those without access to ADSL services. However, slow speeds and patchy coverage led to slow uptake. The Seven Group has since put out a different WiMAX-based offering (802.16e) under the vividwireless brand earlier this year.

vividwireless has generated the most attention for 802.16e WiMAX so far when it launched in March with 150 WiMAX base stations from Huawei scattered throughout Perth. The network launched claiming peak speeds of 20Mbps but promptly dropped the figure and does not currently claim or promise any specific average or peak speeds to customers. However, users have reported an average speed of 9.53Mbps through speedtest.net, which the service provider's chief executive officer, Martin Mercer, recently championed at a 4G conference in Sydney as a success that provided a "superior experience to ADSL2+".

South Australian ISP, Adam Internet, uses 802.16e WiMAX to fill in gaps on the outer fringes of Adelaide where residents and small businesses are unable to receive ADSL2+ services, as part of the Federal Government's Australian Broadband Guarantee. At time of writing, 18 of the total 62 communities targeted by Adam Internet for WiMAX base stations are able to sign up to the service, with the remainder of the areas slated for connection before the end of the year. The service provider artificially caps the maximum WiMAX speed at 12Mbps downstream, but has reported that users experience an average of 11Mbps.

Another notable WiMAX experiment is currently under construction by NSW energy utility, EnergyAustralia. The retailer is in the process of rolling out 140 802.16e WiMAX base stations across NSW, with 20 expected to go into full operation soon as part of the Federal Government's $100 million three-year lt;Igt;Smart Grid, Smart Citieslt;/igt; trial. The wireless network works off 15MHz worth of spectrum leased from Seven Group's Wireless Broadband Australia (WBA) - which also owns vividwireless' spectrum in Perth - and is designed to connect to WiMAX-capable smart meters in homes as well as up to 3000 mobile field computers. While ambitious, the retailer has ruled out interests in signing up customers to the network, even if legislation allowing enterprises to become retail service providers (RSP) were to eventuate.

Put simply, WiMAX as it currently stands in Australia, is untenable as a nation-wide broadband network, and certainly isn't capable of delivering the committed 100Mbps speeds that the Federal Government proposes to deliver for at least 90 per cent of Australians.

Unlike optic fibre-based network technologies, the WiMAX technology’s greatest asset is also yet to make a strong appearance in commercial reality. 802.16m WiMAX, otherwise known as "WiMAX 2", purports to deliver peak speeds of 300Mbps and lower latency than previous generations to make applications like Voice over IP (VoIP) easier to deliver over the network. However, the specification is yet to be finalised and, while reports earlier this year pointed to 2011 as the beginning of the standard, the timeline has since been pushed back to 2012 according to Intel.

The WiMAX Forum says operators are able to co-locate WiMAX base station equipment within existing 3G towers, which means that telcos at least won't have to erect completely new towers if they choose to implement the technology. However, given that the technology is yet to be field-tested indicates that the upgrade path to WiMAX at a national level could potentially take even more time than currently slated for the NBN and, even then, 100Mbps on a completely even playing field for all users (the concept of ubiquity that is part of the NBN) are unlikely.

Next: LTE

Page Break

The case for LTE

The other major 4G technology, Long Term Evolution (LTE), has been marked as the direct successor to current HSPA services by industry body 3GPP, with greater capability for VoIP and telephony services under an all-IP network. While the technology as a whole is comparatively immature compared to WiMAX, full trials of the technology have already begun in Australia, and may come to fruition before the competing 802.16m WiMAX makes a local appearance.

LTE sees a shift in the way most broadband and telecommunications players operate. LTE’s all-IP technological foundations essentially require operators to switch off existing 2G and ultimately 3G networks. While this is inevitable at some point, it also means a potentially massive shift with similar repercussions to the shutoff of the Australian CDMA network in 2008.

Telstra has led the way in the trials, with the telco's six-month tests of the networks beginning in May. According to Telstra, the trial should see the network deliver theoretical speeds of 172Mbps, though one test by Nokia Siemens Networks has seen real-world speeds of up to 70Mbps. Even if trials are wildly successful, Telstra has publicly stated it won't jump into LTE head first, and the first LTE services aren't expected until halfway through 2012 at the earliest.

As part of a potential deal between Telstra and NBN Co, the telco will also be able to bid on wireless spectrum made available by the Federal Government in the future, potentially increasing opportunities for a more expansive and faster LTE network than is currently available over its flagship Next G service.

Optus' parent company, SingTel, has also announced trials to take place in Australia, Indonesia, Philippines and Singapore though further details remain scant.

Other major Australian telcos have expressed interest in delivering similar services by early 2013 but it's clear the technology has some time yet before it becomes a real world application. According to one report on LTE, mass adoption of the technology will only begin in 2015.

Thanks to better handling of radio spectrums, each LTE base station can handle up to ten times more users than 3G base stations are currently capable of; some report 200 to 400 customers each. As such, it's almost a necessity in order to provide better voice and data services to mobile customers without building more stations. However, the fact that a choke point continues to exist means that, as more mobile broadband customers join a network and use more and more data - as they inevitably will over time - the guaranteed speeds will fall unless operators upgrade infrastructure or build more stations. With no easy way to increase capacity at the operator end, the ubiquitous 100Mbps guaranteed speeds will easily drop.

The case for satellite

Australia is currently serviced by satellites operated by either Telstra or Optus, many of which fall under the Ku-band category, which means they operate on the 12 to 18GHz bandwidth of the radio spectrum. While these are largely devoted to supporting backhaul for operators' mobile networks, they are also used to deliver Internet services to rural Australian communities.

Those eligible under the government's Australian Broadband Guarantee program currently have access to Internet services ranging in speed from 256Kbps to 1Mbps at prices of between $19.95 and $165 if you're on a service on the Optus satellite network, such as those from BorderNET and HarbourSat. Telstra's fastest plan, on the other hand, currently costs users $499.95 per month for 800Kbps downstream Internet speeds and 4GB of monthly data quota.

Put simply, satellite currently exists as a last measure for those who can't receive fixed or wireless broadband; the capital expenditure costs are also too prohibitive for mass deployment. Under the NBN, satellite access would serve a similar purpose and is slated to serve about the three per cent of Australia for whom it is too expensive to service through one of the other two methods. Where the government's proposal differs, however, is through the use of two new Ka-band satellites, which operate on the 26.5 to 40GHz bandwidth. These are of course required to meet the minimum 12Mbps speeds required under the proposal.

The first such capable satellites were launched in May and with 38 satellites due to be launched with at least some Ka-band transponders by 2014, it seems most global operators are at least keen to replace existing infrastructure with the heavier bandwidth technology. However, whether that makes it a suitable replacement for fibre-type services or fixed broadband of any type is still unclear; the use of Ka-band technology simply refers to the microwave bandwidth used rather than the type of Internet access.

Notably, a Japanese program called Kizuna is working towards providing a maximum speed of 155Mbps downstream, 6Mbps upstream for “households with 45-centimetre aperture antennas (the same size as existing Communications Satellite antennas), and ultra-high speed 1.2 Gbps communication for offices with five-meter antennas”. However, the program is still in testing and is expected to also be cost prohibitive for mass deployment.

Next: The case for fibre

Page Break

The case for fibre

Fibre is hardly a new technology: Everyone from the government to corporations to service providers have been laying out fibre at every opportunity for years.

For some organisations, it's a present business opportunity; the NSW Department of Education and Training alone runs fibre to nearly all of its 2240 schools. For others, however, it's simply a matter of future-proofing when and where possible, and it's no surprise that NBN Co has been in negotiations with Telstra to secure its existing dark fibre network. AAPT has even complained of having too much fibre.

With plenty of fibre already in place, it's far ahead of existing or future wireless broadband in terms of saturation. For many businesses, it's already a present opportunity. For residents, the backhaul largely exists; it's a matter of running that last mile connection and providing active services.

Perhaps the biggest advantage of a fixed line broadband network, and one that is certainly important to the viability of the NBN over time, is merely the fact that fibre is capable of much more than the 100Mbps it has been initially tasked to achieve. The technology certainly isn't as susceptible to congestion issues as wireless broadband and the ability to scale up capacity from the operator endpoint doesn't necessarily require new infrastructure or firmware upgrades.

Distance isn't so much of an issue either. In any wireless broadband network, the user receives a differing experience regardless of distance from the base station. Similarly to ADSL2+, connection quality and inherent downstream/upstream speeds decrease the further a user is from each cell. While overlapping base stations and wireless coverage, as well as leaps in technology, certainly help to remedy this potential issue, the problem never quite goes away. This especially becomes an issue for those living on the edge of a cell, or in areas where there are only two or three base stations to a given community. Live on the edge of the coverage area, and chances are your Internet speeds are vastly inferior to those at the centre.

For fibre, this is no issue: data travelling down fibre is as fast at point A as it is at point B, with line degradation (or failure) and interference from devices along the network being the largest obstacles in continued, committed speeds.

More importantly, the 100Mbps speed proposed under the NBN is just the starting point for a FTTH network. Look to service provider AARNet and you'll see speeds of up to 10Gbps already in operation, and vision to move to 8Tbps and beyond in the future. A single pair of fibre is capable of 1Gbps and greater, and if the NBN is constructed properly, the use of multiple fibres makes scaling up capacity even easier.

If Mike Quigley is right, and Australians demand speeds of up to 1Gbps by the year 2020 and as we showed in a previous article isn’t an unreasonable expectation, fibre can meet those needs; current and foreseeable wireless broadband technologies can't.


Copyright © 2010 IDG Communications, Inc.

1 2 Page 2
Page 2 of 2
Shop Tech Products at Amazon