Evolution of Voice as a Service

In the era of “data” hogging the limelight, in this post I dwell on “voice” and where we stand now on voice as a service. How do we define “Voice”? Voice as spoken by human beings is the ability to hear the words spoken by another and possibly carry on a conversation. A normal human voice falls in the frequency range of 80-250 Hz compared to the frequencies allotted for mobile communications which typically lie in the range of 700 - 2500 MHz in today’s spectrum allocations. The frequency range of human voice is about 1000,000 times lower in magnitude compared to commercial mobile spectrum.

www.guinnessworldrecords.com informs that the male voice normally goes to 180 meters in still air. For human voice to be spoken and transmitted over a distance (Here I am not talking about mike amplifiers used to address crowds) it can be done through the medium of telecom networks. The first and traditional way is the POTS phone which has a fixed wire or cable from the telephone exchange terminating in the device and it can also be done by using a mobile network. There are other ways like satellite phones or trunk radios but let me stick to voice service that an ordinary user can avail. This voice carried over any network has to have the unique property of being in “real time” so that the end user can comprehend what the other side is saying and carry out a meaningful conversation.

Voice availed through the POTs phones given by the Fixed Line service providers has evolved not only in the way the service is availed by the end user but also in the background technology. We have a fixed phone set at home or office and dial the number we need to dial, hear a ring back tone and when the other side picks up, the end to end circuit gets established enabling the users at both ends to have voice communication. Behind this simple service is an array of equipment, the central of which is the Telephone Exchange and the access side. Taking an analogy with mobile telecom systems, there is an access network side consisting of the physical media laid from the telephone exchange to the end user premises.  The voice is carried over DSL lines where voice and data can be conveyed together over digital lines on copper. Optical fibers are also being increasingly used for the last mile combining data and voice. Devices such as Splitters/IAD/MSAN is used at the customer end premises to extract the voice services and terminate in a fixed line device or telephone set. As such, the media for connecting the end user which initially had only copper cables has now got two options - copper and optical fibers. The other part - the Exchange over time has also seen technological upgrades from analog to digital systems. Even within the digital space it has changed from one version to upgraded avatars known as NGN.  ITU -T defines NGN as one “where service-related functions are independent of the underlying transport-related technologies”. Although improvements have taken place over the years in terms of the digital switching platform, there is no concrete terminology used such as 2G, 3G or 4G as used in mobile networks. But user experience has got enhanced. From the initial days when only plain basic voice was there, now the user can have add-on services like Call Forwarding and Wake-up call facilities.

Even electronic and digital systems have a shelf life no matter how advanced they might be. The service providers will no doubt try to extract the maximum utility from the installed systems and as operational costs increase in terms of repair, support, power consumption and area the equipment takes up, they need to look at technological upgrades to keep the services running.  Service providers will also need to   give increased value to the voice users. This leads to procuring newer and advanced switches and the shift from one hardware and software version to another. In current deployments Internet Multimedia Sub-Systems (IMS) have come in for the core or “exchange” end for PSTN. IMS supports SIP (IP) and the POTS phone gets transformed to IP end points. These IP end points may be SIP Voice phone, SIP Video phone or SIP Clients offering an array of services such as Call Conferencing, Call waiting, Call Forwarding among others using IP as the underlying layer.

Looking at the evolution of voice in a Mobile network, let us look at the two aspects Access and Core as we did in the case of the Fixed phones above. In the access part which the Radio part or “last mile” it is 2G, 3G or 4G to the end user who uses a mobile handset to avail the voice services. For 2G and 3G, the voice is carried over the airwaves through a Circuit switched (CS) network which connects to the mobile Network of another service provider (PLMN) or PSTN network. The 2G and 3G networks use the MSC. an equivalent of the PSTN Exchange to switch calls in different sides. Similar to the PSTN network, an end to end circuit is established from the end user mobile handset through the Radio access and Core Network onwards to the end user on the other side. This circuit is maintained till such time the call is ongoing. From the Circuit Switched network, it has evolved to Packet switched networks where there is now no dedicated path to carry voice. Voice is now carried as data packets. This is the way IP packets flow when we send across emails and other data from one laptop or desktop across cities and countries.

Voice in this new form of data packets is possible through two ways. One through VOIP which has what we have through the use of applications like Skype, WhatsApp voice and the like. These Apps for voice are free in the sense that these apps do not charge from the end users. These apps use the underlying network infrastructure of the mobile networks as an overlay layer. The app providers have their own setup consisting of servers to route the calls, maintain databases and give other value added services through the net. Although voice is apparently free, add on services may be charged. In this scenario, the network service providers who have 2G, 3G and 4G networks do not have much control and in these days of “Net Neutrality” they can ill afford to block or throttle these applications. The other approach which mobile service providers feel they can have more control over a traditional service such as Voice is to have an IMS architecture. This IMS architecture is the same one used for Fixed line voice as we saw above as IMS is agnostic to the underlying access layer whether it is Copper, Optical or 4G. But service providers need to give clear benefit to entice the user to avail voice service given by them. Using a IMS setup gives them the leverage to do so and the benefits in this game of having IMS can be giving better Quality of Service, common look and feel and avoiding the hassles of ensuring that both the callers (caller party and called party) having to use the same application. In current setup, a Skype user can only call another Skype user and not a WhatsApp user. This problem is circumvented by using IMS Voice. For a service provider deploying LTE radio access, deploying IMS network becomes mandatory as LTE by design cannot be connected to a circuit switch (Core) network like the previous generations of 2G and 3G radio to deliver voice services. Voice carried over LTE is known as Voice over LTE (VoLTE). Once a service provider introduces IMS in its network, it can also use it to give voice services over Wi-Fi (VoWiFi) or give Video services.

VOIP or Voice over IP (or Internet) allows a caller to initiate calls from his or her PC, laptop, tablet or mobile device using IP. A user making a voice call to another PC user can bypass the traditional PSTN or PLMN networks.  Depending on local policies and regulations, voice calls from VOIP to PSTN or Mobile numbers and vice versa are allowed or curtailed.

What about the actual conversion of analogue speech as we speak to digital format? Codecs are used for conversion of voice (analogue) to digital format for transmission and re-conversion back to analogue for speech at the other end. Improvements have also taken place in the quality of voice carried over telecom systems over time. For GSM (2G) it started with Full rate (FR) codecs and then progressively seen half rate (HR) codecs for pumping more information, Enhanced Full rate (EFR) codecs reducing errors leading to Adaptive Multi rate (AMR) and Wideband AMR (AMR-WB) codecs in deployments. AMR-WB for instance provides for more bandwidth than conventional 3.1 kHz and has other features for better speech quality and noise reduction.

So where do we stand on voice services in the current scenario? I take the case of India again.

• ·    Incumbent operators have got circuit switch networks that cater to 2G and 3G radio networks which are widely deployed compared to LTE. Using VoLTE presents a challenge as full and seamless LTE coverage is not there across the country currently. Having full LTE coverage will take time although service providers are increasing its coverage. In this scenario, the experience of using VoLTE in patchy LTE radio coverage areas leaves the user experience wanting.
• ·       So what do existing operators who are having LTE radio network as well as 2G and 3G network do? The mechanism of CSFB (Circuit Switch Fall Back) for voice services is a way out. This means that while in the LTE radio coverage area, data services are carried by LTE radio network while any voice service to be availed by a user, the network automatically selects the Circuit Switch Network. This fallback for voice makes sense when the service provider has a robust 2G and 3G network which is true for most such operators in India. As LTE coverage increases and expands, SRVCC can be resorted to. This mechanism uses the IMS Core for voice calls in the LTE radio network and utilizes the existing circuit core for voice calls when the user moves to a 3G or 2G radio coverage area. The difference with CSFB is that in SRVCC, the Call control is retained by the IMS Core. SRVCC is not seen in Indian networks although operators have tested the solution in their networks. CSFB remains in vogue.
• ·      Having a LTE capable handset does not always mean having a handset supporting VoLTE at the same time. It seemed as if the two aspects are separate although VoLTE runs on IMS Core with LTE radio technology in the access. As recently as a year back, there were few handsets supporting VoLTE. But Handset support for VoLTE is increasingly seen across multiple handset manufacturers and the price of these handsets are getting affordable.
• ·       To maintain their technological lead and ensure they are not lagging in the deployment of newer services, providers need to keep abreast of their peers. In the case of VoLTE too, this seems to be in the case along with the inherent need to upgrade the networks and keep pace with the developments in the market.
• ·      The Telecom Regulator (TRAI) regularly publishes the Network Quality performance indicators and considering the technological and market changes, indicators would warrant review as subscriber ramp up takes place in the IP side of voice compared to Circuit Switched voice. 
• ·         Fixed Mobile Convergence (FMC) is not much of a story in India, as of now. Service providers in India are deploying IMS solutions in their Fixed line and Mobile services but the use of one single IMS core is not visible although IMS is access agnostic. Possibly the separation is maintained because of administrative and operational reasons and not due to technological limitations.
• ·      Voice calls from PSTN or Mobile numbers to Voice over IP (VOIP) users and vice versa are not allowed under current policy/regulations. However, PC to PC or PC to International PSTN/PLMN numbers are allowed. No separate numbering scheme exists to identify VOIP users like a PSTN or Mobile number. The regulator (TRAI) had initially proposed allowing such calls and it has issued a consultation paper (2016) on the issue to solicit views of different stakeholders but as things stand, the current policy does not allow calls from VOIP users to users of fixed line and mobile numbers.
• ·     Service providers are giving affordable tariff plans which offer free unlimited local and STD voice calls bundled with data. Voice which was charged per minute or per second or multiples of the same is seeing a changeover. In current scenario, either it is not charged separately (at least in some plans) or charged at low rates.

The genesis of telecom which had voice as its core service is seeing a change and voice is now only one in the bouquet of services provided by the service providers The entry of VOIP and Reliance Jio as a provider which has no 2G and 3G networks but launching LTE networks is changing the market as far as “voice” as a service is offered. From the traditional circuit switched voice where a dedicated path is established to packet switching of calls, voice is on the path of using IP as the underlying layer, whether we term it as a IMS voice or VOIP call.