Two Views on LTE Backhaul

You tend to hear two very conflicting views about the extent to which the rollout of Long Term Evolution (LTE) changes a mobile operator's backhaul requirements. Some present LTE as a profound game-changer that threatens to plunge unsuspecting backhaul planners into a perilous new environment they have not yet even begun to understand. Others present LTE with a shrug of the shoulders as not much more than a business-as-usual extension of the new capability that is already being deployed for 3G backhaul anyway.

So which is it? Evolution or revolution?

Actually, it's both. Or it's either one. It depends very much on what LTE deployment strategy you have in mind.

LTE does undoubtedly present profound, game-changing challenges. The X2 interface between eNodeBs is a whole new paradigm for cellular network planners. Cellular base stations have never had a direct communication path between them before. Deploying the X2 optimally and securely is a major new challenge.

The long-haul S1 interface between the eNodeBs and the core of the network is also brand new. There's no more intermediary BSC or RNC node to provide a security buffer between the cell site and the core. If the backhaul network is what the 3rd Generation Partnership Project (3GPP) defines as "untrusted," then IPSec is prescribed to protect against illegal interception – except that the prospect of managing hundreds or thousands of IPsec tunnels across the backhaul makes many mobile network planners' blood run cold.

And voice over LTE won't be a walk in the park, either. Cellular voice quality is variable enough without having to deliver it over an all-IP network – let alone one based on a newly defined standard that is still very much in its infancy.

These new features of the LTE standard are potentially very daunting for even the most IP-savvy mobile operator; there isn't much serious debate about that. Nevertheless, the alternative "business as usual" outlook is still very credible, particularly when closely linked to pragmatic assumptions about the timescales associated with these backhaul-impacting LTE features.

The bottom line is that it's possible to launch a perfectly viable LTE network in the next couple of years without actually activating the X2 interface; without implementing IPSec; and without offering voice over LTE. Operators certainly need to invest in understanding when – or in the case of IPsec, if – they will stand to benefit from deploying these capabilities and features. But there is nothing to say that they need to be deployed at launch, or even any time soon following initial launch.

As shown in my new report, "Backhaul Evolution Strategies for LTE Operators," the optimal approach for many operators will be rolling out a basic, stripped-down LTE network at launch and adding more sophisticated capabilities over time. That way, the toughest challenges can be managed as the evolution that the operators prefer – and not the revolution than many of them are right to fear.

— Patrick Donegan, Senior Analyst, Wireless, Heavy Reading

For more information about Heavy Reading's "Backhaul Evolution Strategies for LTE Operators," or to request a free executive summary of this report, please contact:

shygye75 12/5/2012 | 4:36:20 PM
re: Two Views on LTE Backhaul

Given what's currently going on in 3G data networks, will operators be able to manage a very controlled, phased launch of LTE service, or will their well-laid plans be overwhelmed by demand?

pdonegan67 12/5/2012 | 4:36:17 PM
re: Two Views on LTE Backhaul

I’d have thought by and large LTE roll-out will have to be controlled or phased. Lack of familiarity with the new technology will pretty much mandate that in many cases. Will the operators be overwhelmed? By the time most operators come to launch LTE, it’s reasonable to assume that they will be a lot further up the learning curve in terms of dealing with data traffic. But just as today, there is liable to be a big disparity between how smoothly or otherwise different operators are able to manage it, especially since many aspects of network dimensioning and traffic management will be very different in LTE.

Abundant_bits 12/5/2012 | 4:36:08 PM
re: Two Views on LTE Backhaul

I think certainly data demand is what is going to overwhelm mobile carriers, sooner than later. Some sort of phased roll-out is inevitable, IMHO, as the QoS for Voice isn't solid enough (yet) to take the leap. 

Get the Data traffic under control, then roll in voice once it's proven there are reliable solutions out there.

macster 12/5/2012 | 4:35:50 PM
re: Two Views on LTE Backhaul

I wouldn't say that the S1 (U) is brand new (think GTP). X2? Not sure how big a role this is going to play given that... There are LTE/backhaul reference architectures that take this into account.

We should understand the impact of things like offload (this does not just mean traffic at the RBS), SIPTO, the effect of femtos (which are currently just based on DSL), etc.

Mobile operators just do not know what to do. O2UK's network had issues. Korea had big issues when the iPhone was introduced there, etc.

We can say that the air interface is not an issue. Putting in capacity in backhaul links is not really a big problem. You just can't break the laws of physics (or maths). We really need to sort out the core!



Sign In