Tarek Robbiati is the opinionated and outspoken chief executive of leading Hong Kong operator CSL, and newly appointed head of Telstra International. He joined CSL three years ago, prior to which he served at the operator’s parent company, Telstra, in Australia.

In October 2006, Telstra marked a world first with the launch of its Next G-branded HSDPA network, offering customers throughput speeds of 14.4Mbps. Appointed chief of the Hong Kong operation, Robbiati was given the task of replicating Telstra’s 3G business model in a new, and very different, market.

“When I came to Hong Kong I intended to replicate the Telstra model and success, using the same design and engineering principles,” he tells telecoms.com during an interview at the carrier’s Hong Kong headquarters. “I tried to do that for a few months before realizing the challenge in Hong Kong was a little different and we needed to replace our infrastructure in the country.”

It’s a realization that many mobile operators may well have reached but the difference with Robbiati is that he acted upon his own discovery without hesitation. CSL discontinued its long standing relationship with Nokia Siemens Networks (NSN) and ripped out all of the NSN supplied infrastructure, replacing it with an all IP 3G platform provided by Chinese vendor ZTE.

It was a move that would be deemed to risky by some but Robbiati is dismissive of what he calls the “sentimentality” of engineers and managers to their legacy network infrastructure. “You have your legacy infrastructure and you have engineers who are too enamoured with that legacy infrastructure, so they don’t want to take the bold steps of removing that legacy infrastructure. It’s what I call the opportunity cost of legacy,” he explains. “So you end up sticking with the old technology and just making incremental improvements, but if you want to do something radical then just change the way you do it.”

What CSL ended up with, according to Robbiati, is, “A data network optimised for voice, not a voice network sub-optimised for data. It’s an important distinction to make because, prior to this, 3G networks were effectively voice networks operating in 2100MHz – they over-promised and under-delivered.”

CSL launched its 3G network, under the same Next G branding as Telstra, in March 2009. The company progressively replaced its old 2G network  with HSPA+ capable equipment offering headline speeds of 21Mbps on its all IP network. Since then, CSL has launched UMTS900, by refarming its 2G spectrum, giving it a very powerful network, with much better in-building coverage, outdoor coverage and speed.

Spectrum is key to CSL’s business model, both current and future. Robbiati openly admits that CSL was “lucky” with the timing of the 4G licence auction in Hong Kong in January 2009, winning 2.6GHz of 4G spectrum and ending up with more than twice the amount of spectrum (127MHz) than its closest competitors. As Robbiati explains, “If you have all this spectrum and SDR (software defined radio) you can do a lot of things with that.”

Both Robbiati and CSL’s chief technical officer, Christian Daigneault wax lyrical about the possibilities afforded by SDR. “ZTE’s superior technology in SDR was very important in our decision making and particularly relevant for Hong Kong because spectrum here is technology neutral,” he says. “This means you can reuse 2G spectrum (900MHz in this case) for 3G services because the radio access network is software configurable. It’s the perfect technology for the Hong Kong environment because you can refarm the spectrum at zero cost of spectrum, although you need to invest in technology (infrastructure) but not in spectrum.”

Robbiati is critical of the regulation binding technology to spectrum across the rest of the world, saying it, “Needs to evolve.” He continues:

“Most licenses that were issued in Europe around 900MHz mean that if you got the spectrum you were only allowed to offer 2G services. It was forbidden to use that spectrum for 3G. It may be that governments want to sell another licence, a 3G licence to operators, but the reality is that spectrum is finite and you need to use it in the most efficient way possible. So you need to relax the regulation around the spectrum, and licence the spectrum to operators so they can use it in whatever way they see fit,” he says, acknowledging a similar proposition from the UK’s Lord Carter in the Digital Britain report.

With CSL set to launch commercial LTE services in Hong Kong before the year is out, Robbiati expresses concern at what he calls a lack of investment in mobile infrastructure meant to offer 3G services.

“What worries me on a macro level is, if you look at the Western world over the last 25 years, infrastructure investment as a percentage of GDP has gone down, and that goes for telecoms, energy, water, airports, railroads. If you look at China they have taken a very different approach. You could argue the telecoms market is less competitive, but at least they’ve invested in infrastructure to spur economic growth. In India, the government still issues one licence at a time to create funds and as they raise funds there is less opportunity to invest. So you attract players that are not in it for the long term and not committed to an investment mandate. It’s very concerning. So you end up with infrastructure that’s not capable of handling this new type of demand and you end up with lots of consolidation.”

At this point Robbiati makes reference to the problems suffered by mobile operators O2 in the UK and AT&T in the US, which have both suffered well-documented network problems since the introduction of data hungry handsets such as the Apple iPhone. He is also quick to note the unsustainability of flat rate ‘unlimited’ data packages with the explosive adoption of mobile broadband services and dongles.

“Saying data is ‘unlimited’ is disingenuous. If you have a car that is only driving in first gear, yes you can drive it in unlimited fashion but you can’t get very far,” Robbiati says. “All the conventional networks are poor quality and poor coverage, so we’ve changed the game. In Hong Kong, lots of people in remote areas don’t have broadband, and the terrain is not easy to deploy in, so their only option is wireless. We are betting that customers value the speed we offer, and if they do value the speed then we win.”

CSL’s wireless broadband offering differs from pretty much every other offering in the world, because it’s based on speed, not usage. “We have gone with speed based pricing, which no one else in the world has done. If you want to be running a Fiat you pay one price, if you want a Ferrari you pay more,” Robbiati says. “It’s all about experiencing the different network experience without facing bill shock. Because you have a flat rate with certain bandwidth available.” The CSL chief claims that since the operator launched its Next G offering, mobile data traffic has increased 23 fold.

Again, Robbiati links CSL’s pricing model for mobile broadband to the decision to rip and replace its old network. “Traditional networks were designed for small data consumption. We were left with no choice but to provide speed based pricing,” he says.

“You need to get the customer to value the experience. If the customer is used to poor service you need to change the perception of the customer and that is a huge challenge. In Hong Kong we can sell 850 minutes for HK$35, that’s less than US$5 per month, then I’ve got bad debt and all sorts of other costs I need to recover, that’s how crazy things have become. People in the Hong Kong market have trashed it to the point where it is insane. As Confucius said “Quality is not cheap and cheap is not good”. So when you get to the point where your infrastructure is going to collapse because it has been underinvested in and that’s where the trouble starts.”

Robbiati and Daigneault argue that the notion of quality of service (QoS) management is central to LTE, that the technology has been structured and designed with QoS in mind. “Not all users require the same bandwidth, so this is where you can start differentiating, by being really granular,” he says.

This differentiation will be key going forward, and CSL must be doing something right, as Robbiati confides that more than 80 operators from around the globe have come to speak to CSL about how it runs its operations in the hope of gleaning some useful tips.

The CSL exec admits that the company is, “No longer a mobile operator.” It is also an ISP and a services provider that handles lots of M2M communications. So with the growing threats from players such as Google, Robbiati acknowledges that the disintermediation risk on network providers is very real. “So how do we address it? We can’t stand still, we have to add intelligence to network. There are a lot of things you can do, like platform services that have applications, and services in the cloud are also very interesting.” So for CSL, the future will be all about the services. “We’ve done all the hard work on the infrastructure already so now we can concentrate on the services,” Robbiati says.