Wi-Fi Calling provides seamless access to mobile voice and is now incorporated into several popular smartphones. The intention is that users should not even be aware they are using it. But the chance of a dropped or poor quality call is higher and is holding back wider deployment. We consider the issues and look at some alternative architectures that propose a way around the technical problems.
What are the issues with Wi-Fi Calling today?
It’s been possible to make and receive voice calls on your mobile device for many years. There are many smartphone Apps which use the data connection, including Skype, Viber and more recently WhatsApp. Many network operators even provide their own VoIP app, providing voice and SMS text services when out of coverage or roaming. These usually need to be accessed separately from the main phone user interface and don’t seamlessly handoff to/from the cellular network. They’re often called “2nd number” services because you aren’t always online to accept incoming calls, and are primarily used to save money, bypassing expensive roaming tariffs when abroad or making international calls.
I’ve identified four main concerns which have held back wider take-up:
- Difficulty in connecting to public Wi-Fi networks, identifying suitable candidates, entering personal details, accepting terms and conditions etc. Next Generation Hotspot should streamline and simplify that choice. Several vendors such as Devicescape and Birdstep offer proprietary solutions to select and use a subset of known good hotspots, continuously monitoring and ensuring performance is up to standard
- The need to open a separate standalone App and sometimes use a different contact list or address format (e.g. email or username instead of phone number). Wi-Fi Calling uses your existing phonebook and telephone numbers seamlessly within the same standard smartphone user interface, making this relatively painless.
- Variable call quality because the data path gets little or no special treatment from other services, resulting in dropouts. When it works, it can work very well and be very much cheaper (free) compared to a roaming voice call. When it doesn’t work, it can be distracting and worse than useless. The relatively short range of 5GHz Wi-Fi limits range.
- Dropped calls when moving out of the area. Seamless handoff to the cellular system is still in its infancy and calls can expect to be lost. Handover between Wi-Fi access points sharing the same controller (e.g. in a large building) can work well where properly engineered but are not as deterministic as with a cellular system.
Wi-Fi Calling, a more popular term for Voice over Wi-Fi (VoWiFi), seamlessly embeds the Wi-Fi voice service into the smartphone user interface – it becomes transparent and operates exactly as if on 2G, 3G or 4G cellular. The voice service is handled by the same 4G core network (IMS) as VoLTE and can support the wider range of features. Standard interfaces and protocols promise to resolve the handoff issue, but still need further refinement. Apple incorporated this in their iOS 8.3 update in April 2015 and is live with Sprint USA and EE UK. Android Nexus 6 was enabled in May, alongside the Microsoft Lumia 640.
EE UK advertisement at Kings Cross London Underground station. There is no cellular coverage here (from any operator), but Virgin Media (a UK Cable network) have installed Carrier Wi-Fi at most stations. EE (and some other mobile providers) already allow their customers to connect for free via a wholesale arrangement, but Wi-Fi calling extends that to the full range of voice and text services. Other networks (such as Three UK) offer similar facilities but through a more awkward OTT App.
But not every network can or will deploy VoLTE
Wi-Fi Calling is an added benefit of VoLTE (Voice over LTE) which normally requires a full IMS Core Network. That in itself can be bought and installed relatively easily (e.g. Cisco through its Starent acquisition has been winning a lot of business in that market, but most of it has gone to the major RAN equipment vendors). The higher cost involves replicating all of the existing 2G/3G voice services in this new 4G environment, retaining backward compatibility for use in either mode. This has delayed commercial deployment of VoLTE by some years. Only a handful of networks offer commercial VoLTE service today, and even Verizon/AT&T don’t yet have end-to-end VoLTE interoperability between their networks.
By April 2015, only 16 networks (4% of the 393 live LTE networks) had deployed VoLTE. Elsewhere voice services remain on 3G using the CSFB (Circuit Switched FallBack) feature that routes all voice calls through existing 3G radio and core networks. The GSMA report that less than 25% of LTE networks are actively planning or trialling VoLTE.
In the longer term, Wi-Fi Calling should also be able to take advantage of any RCS (Rich Communication Services) features built on top of VoLTE. These include video calling, conference calling, messaging – as already found in many competing Internet Apps. A comprehensive technical view of VoLTE/RCS evolution can be found in this 50 page white paper from 4G Americas.
Taqua offer a shortcut to Wi-Fi Calling
Taqua are a US core network vendor supplying many IP based voice switches for both fixed and mobile networks including Sprint’s CDMA small cell core network equipment. The company acquired Kineto (who pioneered an earlier Wi-Fi Calling technology called UMA). They offer a solution which [as I understand it] converts between the standard 4G VoWiFi protocols and the standard 3G core network interfaces. So the 4G Smartphone thinks it’s talking to an IMS core, while the 3G voice switch sees just another 3G connected smartphone. It’s quite a clever step without needing a full IMS core network and all that implies.
Sprint is probably one of the largest and earliest customers – they’ve delayed launching VoLTE but have rolled out VoWiFi using Taqua kit. If this is proven to work well, others may be tempted to take this route.
Networks tell you to expect dropped calls
A significant limitation of Wi-Fi calling today is whether handoff to/from cellular (e.g. as you leave/enter a building) is good enough. Sprint spokeswoman Michelle Leff Mermelstein told FierceWireless that Wi-Fi-to-Cellular handoff is “not smooth” and that the call will “likely drop”. Sprint’s own FAQ page is even more pessimistic - “An active call will disconnect if the caller moves from Wi-Fi coverage to a cellular network.”
AT&T Wireless also have grave concerns about this issue. Ralph de la Vega, CEO of Mobility and Business Solutions, is quoted speaking in Sep-2014:
“ De la Vega also stressed that the service would not roll out until AT&T was satisfied with the performance of the service in terms of handoff quality from Wi-Fi to cellular, which has been a sticking point with such services in the past, as evidenced by the performance of T-Mobile’s HotSpot@Home service, which struggled with handoffs from Wi-Fi to its older GSM/EDGE and nascent 3G network before it was discontinued in 2010 due to those longstanding issues. ”
So it’s unlikely ATT will launch Wi-Fi Calling until they are satisfied seamless handoff is fully in place.
Elsewhere, EE in the UK also warn that calls will drop and are working on a fix which may be available sometime later this year.
Clearly, there’s a lot of work going on to address this issue but it’s not yet solved.
Several people have suggested to me there may be other ways to re-use existing 3G core network solutions for Wi-Fi calling. These have the added benefit of reducing the impact on the core network, especially when ensuring close compatibility with existing voice services.
The most straightforward was called VoLGA (Voice over LTE using Generic Access), an evolution of the earlier UMA scheme and championed by Kineto. Now that Taqua has acquired Kineto, I doubt that will be resurrected.
YoctoComm, an Israeli startup, is proposing using the existing 3G Small Cell standard Iu-h interface to connect smartphones directly into a Small Cell Gateway. This benefits from needing little change to the core network but does require additional software in the smartphone. Doron Shalev, their CEO, points out that not everyone will adopt LTE globally; 3G isn’t set to peak until 2019 with 48% of users worldwide.
Their proposed architecture can use 3G soft handoff to make a much more seamless and smooth transition between the two radio technologies. Each voice call is effectively transmitted in parallel via both Wi-Fi and 3G during the transition, with the best quality packets in each direction being retained and used. Their architecture can also be used with an LTE core network, and includes options for Carrier Aggregation and wireline/wireless convergence.
VoLTE itself is still in its infancy with only 40 networks publicly committing to launch. Seamless Wi-Fi Calling is a by-product of VoLTE and beginning to be rolled out by a few large operators. A major concern is customer experience, especially dropped calls when leaving the Wi-Fi zone, which will hold back more widespread deployment.
Alternative architectures exist which make it possible to deploy Wi-Fi Calling without a full 4G IMS core network and all that entails, instead requiring software changes to smartphone devices.
These compare with a range of proven Small Cell solutions for both 3G and LTE which are compatible with today’s standard mobile devices and provide high quality, seamless service.
While Wi-Fi Calling will be attractive to some, there is some further technical work to resolve these issues before the technology becomes globally widespread.