As the sun sets on 3G networks, tackling VoLTE becomes the order of the day


The rollout of 4G was hastened by the ability to offload (via CSFB) the complicated issue of handling voice calls onto the established world of 3G. But the fix is running out of time.

If only progress were simple and straightforward, and every technology advance emerged from the oven fully baked. But in the real world, that’s not how things happen. Recent (and ongoing) network evolution underlines that lesson, nowhere more clearly than in the case of 4G voice.

Circuit switched networks (of which 3G was both the most recent and the last generation) were designed for voice and performed that role (of transporting voice) admirably. However, as the appetite for data boomed around the early 2010s among consumers and enterprises alike and thus service innovation came to depend on accommodating and enabling far greater data usage and volumes, something inevitably had to give. Data over the circuit switched network wasn’t a formula for market growth.

Switching to 4G

Change was needed and it came in the form of 4G, the first purpose-built data-first network backbone. In 4G unlike 3G network technology, IMS acted as the control platform rather than the circuit switch. This meant next generation data services could take off. So, with 4G, the first major steps into the telco future could reasonably be assured.

All good then, but what about voice? Changing the bathwater may have been necessary but throwing out the baby wasn’t going to be viable. Worse, 4G voice presented a real problem, because 4G was no more optimised for voice than 3G was for data.

To solve the problem, or at least to put a bandage on it, technology did what technology does; it found what amounted to a workaround. That came in the form of Circuit Switch Fallback (CFSB), a mechanism that enabled voice and SMS to function as normal in 4G networks using 2- and 3G systems. CSFB mean carriers could exploit the new while still continuing to leverage the old. And that, for most of the globe, is what they’ve continued to do until today.

The fly in the ointment

If progress were simple and straightforward – we’ve already noted that it isn’t – we’d be able to leave things there; business as usual. However, a fly in the ointment is rapidly emerging. There’s mounting evidence that 3G networks are reaching end of life and the pace at which they’re now shutting down is visibly increasing. Proof of this abounds; in Greece we saw Hellas begin to shut down its 3G network this summer. Others, including KPN, are following the same path.

The question is, why is this happening now and there’s an easy answer: spectrum availability. To accommodate new (now 5G) services, existing 3G spectrum must be made available and re-purposed, and this can only be achieved by sunsetting those networks.  Once their spectrum is freed up, 5G can roll out at pace. So far, so good but where does this leave voice? With the closure of 3G networks, the CSFB workaround disappears and suddenly the challenge of voice over the data network becomes unavoidable because it can no longer be offset. That’s the point in time we’re fast approaching today.

But what about voice?

To answer the question “what to do?”, we need to backtrack. CSFB was perhaps always likely to be only a temporary measure, favoured initially because it sped up the deployment of 4G networks by overcoming an immediate problem. But like many solutions designed to be short-term, it worked, and it became comfortable.

However, the voice issue was solved with the definition and introduction of VoLTE – Voice over LTE, which brought a new voice service optimised for LTE networks and with the promise of much better audio quality. From around 2012, MNOs started to deploy this, but it took time to gather pace – largely because it was complicated and there was a lack of device clients, at least in the earliest days.

However, while VoLTE is now commonplace, VoLTE roaming was largely neglected. Basically, outside of a few cases – a cadre of carriers in the US, for example – few deployed it. As a result, CSFB has remained in use for roaming. That’s why data sessions are interrupted when you receive a voice call when abroad, as you probably know – because the device switches back to 3G connections, even if those are largely over a legacy Circuit Switched voice service, carried by IP transport now.

However, things are changing. As we’ve seen, due to the sunset of 3G, that’s not going to be tenable for much longer. Today, carriers have to get to grips with VoLTE roaming because doing so can no longer be postponed (unless they want voice services to disappear from their offering altogether).  In consequence, the GSMA is leading the way to define approaches that might smooth the transition.

VoLTE architectures

The new parameters to support VoLTE roaming defined by the GSMA, include not one but three alternative reference architectures. Two are based on an approach that involves calls being broken out locally:

  • LBO Home Routing (LBO-HR)
  • LBO Visitor Routing (LBO-VR)

The third reference architecture is an interface between the Serving Gateway (SGW) and the Packet Data Network (PDN) Gateway (PGW) known as S8 Home Routing (S8HR).

Among the three options, all things aren’t created equal. S8HR features in the majority of early VoLTE Roaming implementations not because it’s inherently better but because it’s considered the easiest approach to implement. Easier, perhaps but traffic volumes have indicated that it’s also susceptible to delays with calls falling back to 3G as a result.

In contrast, the two LBO models follow more standard IMS implementations, as follows: in a VoLTE roaming call, the Proxy Call Session Control Function (P-CSCF is a SIP proxy that acts as the first point of contact for the handset in the mobile network) works in tandem with an Access Session Border Controller (SBC) which is deployed to route IMS signalling and also to anchor calls.  As a result, the Enhanced Single Radio Voice Call Continuity (eSRVCC)) server can respond as required.

These reference architectures enable local break out and calls to be routed after which the IPX Network-Network-Interface (NNI) also defined by the GSMA further specifies signalling and media IP interconnect requirements.

About Utel

Having defined the landscape and understood the theory, in our next blog, we’ll look more closely at the practice; VoLTE in action and what early deployments can teach us. Utel helps telcos to gain greater control of how their networks are performing, improve service assurance, and benefit from an enhanced ability to deliver the best possible experience to their customers.  If you’d like to learn more, please get in touch.