Mark 2012 in your diary. That's the year that Marcus Weldon, Alcatel-Lucent's CTO, believes that the network industry may no longer make any money. “The problem is that cost and revenues are uncorrelated,” he explains.
But it's not simply the commoditisation of bandwidth that's the issue; instead, the descent into fiscal hell is being driven by a vicious cycle of devices, demands and dollars. On the one side, subscriber are taking up more apps and demanding network access and services across ever more devices; on the other side, service providers are caught between increased CapEx and OpEx costs in provisioning and maintaining the next-gen networks required by this demand and also seeing a decreased role in the subscriber value chain, as more and more third-party players control the content layer of the equation.
In a direct parallel with the storage industry, service providers have realised that they can't keep throwing bandwidth at the problem. Neat innovative tricks (many emerging from Bell Labs) are stretching the capacity of copper and pushing the range of wireless but those aren't longterm games worth playing. Instead, the industry needs to think differently.
When, asks Weldon, bandwidth hit the wall? He believes at 30 kbs simultaneous subscriber usage. the average today is a third of that with demand rising daily…
The real problem, he explains is that “data doesn't make money – it's subsidised heavily by voice”. In his view, a typical industry statistic is that two-thirds of the cost of delivering data is subsidised by the operating profit by voice. And that is not a feasible business model.
What Alcatel-Lucent is working on to solve this dilemma is a combination of three things:
Firstly, technical innovation to squeeze bandwidth usage. So, for example, it has just announced the latest release of its Intelligence Services Access Manager (ISAM). This high-capacity GPON and point-to-point fibre capabilities of the ETSI-based 7302 and 7330 ISAM platforms allow service providers to offer a mix of access technologies, using fibre and copper, through the same node. It also supports multiple service providers offering multiple forms of content on a single network. And it also introduces an application enablement processor, enabling service providers to cache and replicate content at the node, reducing the demand for bandwidth across the network.
Next, add more intelligence to the network. So, for example, he points to a ciombination of UP and layers combined with intelligent analysis which “can save 30% on CapEx and OpEx”.
Finally, new types of content and new ways of delivering content that drive subscriber revenue. As an instance of this, he talks of Google's QoS connection based on an advanced forwarding on demand model, which derives revenue from subscribers that can be split with the service provider. Or guaranteed fast delivery of content via Kindle, for which a subscriber pays. “What is important here,” he says, “is to look at the cost of a service people didn't have before.” For example, the New York Times used to be free to read on-line but is now charged $13 per month via Kindle. You can save a physical delivery cost to a front doorstep of perhaps $500 a year at which point it makes sense to just give away the Kindle to drive people to this service.”
In other words, the way to avoid the brick wall of 2012 is now by provisioning cheap bits but by the intelligent delivery of bits. “So we should move content and apps to the point where it can be best delivered – the shift from the core to the edge closer to the user. We use introduce a delivery prioritisation of premium content and discriminate between different data flows. And, importantly, we should simplify the management layer.”
The result of all this? “The high leverage network that has an optimised cost/performance balance. The reality we need to understand in a world of explosive device growth is that the device is not the right place to do the processing. Instead, it's the place to do the rendering after media mixing on the network.”
One initiative set up by Alcatel-Lucent may help this profusion of devices. The Open API Platform is the result of what Scott Manson, Director of Application Services, calls a focus on how to help service providers create new revenue streams. Deriding the traditional network as “just a dumb pipe from which it's hard to get back your investment dollars”, his group has been working on “making the network usable for third parties”. That used, he admits, to be the jobs of telcos but “none of them are going to write the next Twitter”, he says. By using the Open API Platform as a way of delivering apps across multiple devices at a low cost, he sees a massive potential in the number of offerings that service providers can bundle. “We're dealing with an industry where there are more than 2,500 devices, more than 750 networks, more than 100 app stores and more than 14 million developers – there's a lot of potential there to monetise!”
Equally important for the creation of more intelligent networks is grasping, in the words of Jean-Pierre Lartigue, Alcatel-Lucent's VP Wireless Networking Marketing & Strategy, that “wireless and wireline can't be differentiated in terms of solutions for network roll-out. We no longer care about the boundaries in a transforming world.” In his view, the problem is simply put: “We can squeeze more with copper, but how to do this in the wireless space?” The answer is equally simple: network optimisation and network intelligence.
At least Alcatel-Lucent hope so. After all, 2012 is approaching fast!