Voice over Internet Protocol (VoIP) has been growing steadily in popularity over the past years. Instant messenger clients with voice chat features have given way to P2P (peer-to-peer) VoIP clients like Skype. We take a closer look at VoIP (Voice over Internet Protocol) technology, and the challenges and innovations associated with it.
If you have been on the Internet long enough, there is a good chance that you have spoken to someone over the Internet, either using a traditional chat client (like Yahoo!), or P2P software like Skype.
VoIP, or Voice over Internet Protocol, is the technology used to transmit voice over the Internet. Also known as ‘Internet telephony’, it involves breaking up analogue voice data into digital packets, and then transmitting these packets using the Internet Protocol (IP).
Factoid: A common misconception about VoIP is that it refers to the transmission of voice over only the Internet. Using VoIP, voice can be transmitted over any network that uses the Internet Protocol.
How VoIP works
To make VoIP calls, you need an Internet connection with 56 kbps or higher speeds (broadband is recommended). You also need a VoIP device that breaks your voice into digital packets and sends the packets across to the intended recipient.
The VoIP device could be an ATA (analogue telephone adapter) and a regular phone, an IP phone, or a softphone with a microphone and a headset. (A softphone is a software client (e.g., Skype), installed on your computer, that allows you to make VoIP calls.)
Depending on the type of calls you intend to make, you may need a service provider (like AT&T, Verizon or Skype).
The role of codecs
An important component of VoIP technology is a codec, short for coder-decoder. A codec is an algorithm that codes and decodes human voice, and compresses the digital data for transmission (so that the bandwidth to carry the data is reduced). A codec could be in software or hardware form. Traditional telephony uses the G.711 codec, while VoIP uses codecs like G.729, G.722.1, etc.
We asked Anil Kumar, manager—Video Telephony and Communication, Ittiam Systems, about the speech codecs used in VoIP. He says, “There are primarily five aspects of a speech codec that determine how useful it is in a particular VoIP scenario. These are: quality, processing power, delay, error resilience and compression.”
He adds, “G.711 has the best quality, processing power and the least delay and processing power. However, it lacks concealment and in-built error-resilience, and does not offer high compression. G.723 requires higher processing power, is high in delay and low in quality, but performs quite well in terms of compression, and error resilience and concealment.
“G.729 has the best balance of all these aspects, but does not scale for higher frequency bands. The codecs that may emerge in the future are the iLBC and G.722.2+ codecs. However, both these codecs require higher processing power.”
Protocols in VoIP
In VoIP, the Internet Protocol (IP) layer is used to route voice data packets between locations. Another protocol called UDP (User Datagram Protocol) is used on top of IP to send (transport) the packets. The data is carried using a protocol called RTP, the Real-time Transport Protocol, which is used for the transport of real-time data like audio and video.
RTP is used in tandem with another protocol called RTCP (Real-time Control Protocol), which is used to provide information about the Quality of Service (QoS) being provided by RTP. The signalling between the applications is handled by protocols like H.323 or SIP, SIP being specifically created for telephony (VoIP) applications.
Prof Henning Schulzrinne of Columbia University is the co-author of RTP and SIP, and numerous other RFCs. On the possible problems with the RTP protocol and developments for the future, Prof Schulzrinne says, “RTP is meant to be a stable foundation for media transmission, and is largely complete. The major current efforts that are likely to impact product design include improved QoS feedback and the addition of more media types.”
Innovations in VoIP
With VoIP, data and voice can be transmitted over a single network, which reduces hardware and operating expenses. The convergence of voice and data provides opportunities for companies to innovate solutions. For example, one company allows you to receive your voice messages as MP3 files that are sent as attachments to your e-mail address. We also look at some other innovations in the VoIP space.
Click to call: You may have already seen websites that allow you to click on a button and chat with ‘live’ support personnel. Similarly, a VoIP technology called click-to-call allows you to call people by clicking a banner or a button on a website. Click-to-call is already being offered by companies like Ifbyphone, Yellow Pages and Esqube Communication Solutions.
VQube, India’s own P2P VoIP tool: Esqube Communication Solutions Pvt Ltd, a Bangalore-based company, recently released a product called VQube, which the company calls, “…a peer to peer Voice over IP interactive communication tool that offers voice, video, text chat and voice mail.”
On the subject of how the Esqube team designed the product to address technical challenges in the Indian market, Dr K V S Hari, CEO, Esqube Communication Solutions Pvt Ltd, says, “The main technical challenges we face in India are in the areas of hardware limitations, network constraints, and bandwidth limitations.”
He reveals, “VQube uses a lightweight voice codec that needs lower CPU time (compared to other codecs), thus keeping the CPU and memory requirements to a minimum. This means that VQube works satisfactorily on older machines and on older versions of operating systems, as well as on set-top boxes installed with Linux and Windows mobile phones.”
“Since bandwidth in India is still at a premium, VQube uses a secure, lightweight protocol for transporting voice packets, which reduces the overhead and contributes to bandwidth reduction. Bandwidth usage (for voice and video calls) was also minimised using techniques like dynamic voice-codec switching, video frame rate selection, and so on,” he says.
About the network-related challenges, Dr Hari says, “In many networks, the desktop is in a private IP behind a NAT (network address translation), with the same Internet connection shared by all users. Sometimes the Internet connection is in a tight firewall (only standard ports are open to the outside world), and can only be accessed via a proxy.”
“VQube’s network discovery protocol detects the type of network, the type of proxy, and the ports that are open for voice packets to go through, and automatically routes the packet without reconfiguring the network. Also, in most conditions, packets travel P2P, thus minimising the delay,” he adds.
IP video: If you can transmit voice and data using IP, then the next step would be to add video. Companies are already working on products like IP video phones, and video conferencing systems.
About IP video phones, Kumar says, “Ittiam’s technology has been licensed and taken to market by multiple vendors, and is currently under a trial run by an Asian major. Although the IP video phone is still an emerging market, there is a clear sign of increasing interest and intent of adoption by major OEMs around the world.
The future of video telephony is bright, and it will come into our lives in multiple forms, like video phones on desktops, video phones on IPSTBs (IP Set Top Boxes), video communication embedded in ATMs, etc.”
Private networks: Since the Internet is a public network, there are inherent pitfalls (e.g., security, voice vs data traffic priority) associated with transmitting voice calls over the Internet. Some VoIP service providers use dedicated private networks that enable secure and faster transmission of voice calls. Using private networks also allows service providers to tweak parameters for better performance (for example, prioritise voice traffic over data traffic).
One example of a private network is the Multi-protocol Label Switching (MPLS) network that is used by VoIP service providers.
Challenges facing VoIP
Since VoIP is a relatively new technology, challenges are to be expected as the technology matures. We look at some of the challenges in this domain.
Security: The one challenge that seems to come up in every discussion of VoIP, is security. Addressing this point, Minhaj Zia, business development manager, Cisco India & SAARC says, “One of the most common threats is a denial of service (DoS) attack, which shuts down applications or servers. These attacks are often made against routers or Web servers, but they can also be used to target call-processing servers in IP telephony networks.
“Call tampering, which involves tampering with a phone call in progress, is also emerging as a threat. For example, the attacker can spoil the quality of the call by injecting noise packets in the communication stream. Or the attacker could withhold the delivery of packets, so that the communication becomes spotty and the participants encounter long periods of silence during a call.”
Shawn Conroy, vice president – voice networking, AT&T Inc says, “A secure VoIP deployment must include a combination of existing IP security mechanisms and VoIP-specific security mechanisms. Ultimately, the security of a customer’s VoIP services depends not only on the security measures that the vendor deploys in its network and services, but also on the security that the enterprise customer implements at its own locations.”
He adds, “Many of the security measures taken by vendors such as AT&T, to secure VoIP, must also be deployed by an enterprise on its own devices to be fully effective. Security at the customer’s premises is a critical component of end-to-end VoIP security.”
Prof Schulzrinne expresses his views on security saying, “In my opinion, most of the problems can be traced back to implementations —either sloppy implementations that fall prey to buffer overflow and call-state attacks, or lazy implementers who do not implement well-documented security techniques, such as TLS (Transport Layer Security) for signalling. The one major protocol deficiency is probably the lack of a standardised key agreement protocol for media security.”
Spam calls: Because VoIP calls are sometimes routed through public networks (like the Internet), there is a possibility of spam. Prof Schulzrinne says, “I have not heard about any real incidents of SIP-based spam, presumably because most SIP-based systems are still closed. SIP Identity is a precondition for many of the other spam-fighting techniques, such as white and blacklists, or more sophisticated versions, such as whitelists based on social networks.”
He adds, “Several other kinds of techniques have been proposed, such as detecting spam calls based on frequency or duration, or using the fact that callers start speaking immediately rather than waiting for a “Hello” greeting. Most of the techniques in the latter group have the disadvantage that they are likely to suffer from high false-positive rates, i.e., non-spam calls that are falsely labelled as spam. (For example, automated phone alerts or airline departure delay notifications are likely to be labelled spam.)”
Codecs: About the challenges in implementing codecs for VoIP-related technologies, Kumar says, “For VoIP-related technologies, several aspects of codecs and components need to be examined. The most important ones are implementation optimisation, system-friendly implementation and testing.
Since VoIP is a low-delay real-time system, the algorithms should be optimally implemented so that they introduce the least amount of processing delay. In addition to implementation optimisation, the components must be usable, and implemented so that the system has maximum flexibility in using them.”
He adds, “A core component like a codec must then be put through stringent and exhaustive tests. Examples of such tests are, a speech decoder’s behaviour if the data is corrupted, or an echo canceller’s performance when put through a non-linear path.”
Issues specific to India: On the challenges faced by VoIP in India, Zia opines, “In India, the biggest challenges were the regulatory restrictions imposed by the Indian government and interoperability issues that resulted from a lack of standardisation. Until recently, a single infrastructure for public switched telephone network (PSTN) and Closed User Group (CUG) was not allowed in India. However, the TRAI’s (Telecom Regulatory Authority of India) notification on the approval for “logical partitioning” of PSTN and CUG networks means that companies can reduce their investment expenses. This is because they would not have to manage the expense of two separate (PSTN and CUG) networks.”
About the non-technical challenges faced in India, Rajasekharan N K, executive vice-president—Business Development, Esqube says, “One challenge is the reluctance of Indian ISPs to adopt Esqube’s home-grown technology in spite of the threat of revenue loss from (foreign) service providers like Skype. The other challenge is the government’s regulatory framework related to the restriction of VoIP to PSTN gateway termination within India.”
The future of VoIP
We asked people from the industry about the future of VoIP in India. Shawn Conroy observes: “Business VoIP is picking up speed as more companies are turning to VoIP to improve functionality and increase cost efficiencies. There is a sizable growth market, particularly in India, as corporations are leveraging VoIP solutions and services.”
Rajasekharan shares a similar opinion: “The future of VoIP in India looks positive. Products like VQube, which support secure multimedia communication, in addition to VoIP, will find traction due to multi-location enterprises that are looking at improving efficiency and reducing cost. Students travelling abroad for higher education will drive the residential VoIP markets, since families and friends will want to keep in touch at low costs. In addition, social networking sites like marriage portals will find a need to adopt VoIP to attract more memberships.”
K L Narayanan, head—Business Unit, Convergence, Avaya GlobalConnect adds, “Clearly, the future of communications is VoIP. Businesses and consumers are already taking advantage of the cost savings and new features of making calls over a converged voice-data network.
“The logical next step is to take those advantages to the wireless world. The potential impact of wireless VoIP on the communications market is enormous.
“Wireless VoIP offers potential savings by allowing companies to change the way they manage their phone systems. On the lines of Web 2.0, there could be something like VoIP 2.0, which will facilitate more flexibility, customisation and powerful features. It will be the next phase of VoIP.”
The future is clearly bright for VoIP.