As with the affairs of men, there is a tide in the affairs of mobile networks. More or less once every ten years that tide comes in, bringing with it a new technological generation.
Right now the LTE tide is turning to flood, with the technology moving from definition to implementation. Commercial launches of LTE networks are expected to start in 2010, with early deployments from NTT DoCoMo, Verizon and TeliaSonera.
No new sources of revenue
The introduction of LTE is provoking sharp feelings of déjà vu. Many of the claims being made for the technology are eerily reminiscent of those made for 3G ten years ago, some of which - in particular the suggestion that higher data speeds will enable new sources of revenue for mobile operators - are no more likely to be true now than they proved to be then.
The issue of rollout strategy also has a certain ‘retro' flavour. As with 3G, there is the question of whether to go for a rapid and wide rollout so that LTE is quickly co-extensive with the existing network, or to go for a much slower and more targeted rollout, using LTE to provide increased data capacity in the locations where it is most needed.
From both a marketing and a technical perspective, faster rollout appears to make most sense. The history of mobile communications from Rabbit to Wi-Fi shows how hard it is to sell a service which requires user behaviour to overcome limited coverage; and the efficiency gains which LTE promises are undermined by the need to maintain parallel architectures associated with a gradual rollout.
Technology not yet ready for wholesale replacement of the existing mobile network
However, not all technical factors point in the same direction. LTE's support for legacy services such as voice telephony and SMS (or rather, the absence of a well-defined mechanism for such support) means that it cannot yet effect a wholesale replacement of the 3G network. The phased introduction of devices, beginning with data-only devices, tends to support the same conclusion.
Commercial and financial considerations also point in the opposite direction. A fast rollout and rapid replacement of the existing network will be very expensive, and there is little evidence of new revenue streams enabled by LTE to justify it. Investors and lenders are generally unimpressed by the business case for a rapid rollout, both as a result of the current financial climate and as a reaction to the disappointed expectations that 3G would enable new revenue streams for mobile operators.
Eventually most mobile networks will almost certainly implement LTE, just as they did 3G; only niche players with a very carefully defined strategy aimed at low-end customers can afford to stay off the LTE bandwagon permanently.
For most operators there are few advantages to early LTE deployment
For most players, there is no need to rush headlong into implementation. A few operators with specific technology migration concerns or capacity issues may have grounds for an early move to LTE; others can afford to wait until implementation and operational issues are resolved and the business case improves.
CDMA operators with LTE ambitions will have to assess what they want to do with their existing CDMA assets. This may range from shutting down the network to go to HSPA before moving to LTE, to operating two or three networks in parallel (CDMA/LTE or CDMA/HSPA/LTE).
It isn't necessary to do everything at once or soon. There are few first-mover advantages at the moment, as the entire LTE environment and business case is not yet mature. However, it's crucial to have a top-down strategy with a clear vision of what LTE is for; this will drive the business case and thus the rollout plan, which in turn is needed to drive the most pressing issue - spectrum acquisition.
Here the only certainties are partial coverage and dual-mode operation for some years to come, and eventual replacement of the 2G/3G network to ensure ROI. At the same time, all intermediate steps should factor in the eventual migration to LTE, so that investments and incremental improvements to the network are LTE-ready.