Yusuf Kaka - VoIP World 2006


There are at least three leading Broadband Wireless Access (BWA) technologies capable of enabling wireless VoIP. These are Wi-Fi-Mesh, WiMAX and 3G. In order to provide the level of service that users demand, these technologies will have to meet certain requirements. These include Quality of Service (QoS) provisioning to ensure good conversation quality, mobility, which cellphone users have become accustomed to, security, high availability and low costs. In all likelihood, all of these technologies will have some role in the future of VoIP. The general consensus in the wireless industry seems to be that the next generation of mobile networks will not be focused on any particular access technology but on a "network of networks" instead. The aim is to achieve true ubiquitous networking through seamless integration of various access technologies. This vision is common among discussions on 4th generation GSM (4G), IP Multimedia Subsystem (IMS) and Generic Access Networks (GAN). Users will have the choice of using the best access technology for the situation to access any of the value added services they wish to subscribe to.

Let’s paint a picture...

In an attempt to understand the possibilities, let us imagine the perfect VoIP world. Imagine beginning the day with breakfast at home. On the table, next to your coffee is a smart-phone indicating, by means of a tiny icon that it and whoever you allow, are aware of your current context. You receive an instant message from the traffic department alerting you about traffic congestion on your route to work with a suggested alternative route around the congestion. Realizing that you might be a little late, you lookup work colleagues that are shown as being currently availability on your mobiles address book. The tiny letter “w” next to the signal strength indicator on your device is the only indication that your device is communicating with the Wi-Fi wireless router that you recently installed in order to connect to your municipal, city-wide Wi-Fi mesh network. As you leave for work, you call out the name of a colleague, whose context indicator shows that he is at work already, to your hands-free kit, which establishes a call. Your mobile, aware of your new context, routes the call over your preferred 3G network. After negotiating the morning’s traffic you arrive at the office and as soon as you're in range of the office Wi-Fi network, your mobile changes context once again and beeps indicating that it has received the day’s appointments and new email messages. At your desk, you initiate your first meeting of the day by first calling up a colleague that is somewhere in the next building, after which you call upyour client that happens to live on a remote farm serviced by a WiMAX link.

Behind the scenes a mobile operators 3G network is used, a municipalities community Wi-Fi network is used as well as a fixed line operators WiMAX network. Furthermore, features such as context awareness and instant messaging, typical of Internet-based communication platforms is also used. What then needs to happen for the scenario to become a reality?

Firstly, and perhaps the more difficult step is that various operators will need to agree on how to work across physical network boundaries. Agreements need to be put into place between traditional mobile and fixed line operators, Service Providers, Municipal and Community-Owned Wi-Fi network operators. Network Operators may look to standards such as Unlicensed Mobile Access (UMA), which provides the mechanism for linking unlicensed spectrum technologies such as Wi-Fi and Bluetooth to traditional cellular networks. A more “operator-friendly” next generation network standard that has the potential to fulfill the vision is the IP Multimedia Subsystem (IMS) standard. IMS uses a 3GPP SIP implementation to provide multimedia services to mobile and fixed telecoms operators. IMS specifies the Access network interfaces, the Core network and the Billing systems.

Critics warn however that users will probably not experience the full benefit of an open, multi-service-provider network (like the Internet) from an IMS based system, since access to service provisioning is controlled by the operator, which will have the means to allow or disallow services. Alternatively, independent third-party operators have an opportunity to integrate the various stakeholders, as demonstrated by the likes of “The Cloud” in Europe, which links 6000+ Wi-Fi hotspots together and provides in-bound roaming services to various operators including BT, O2, Skype and Vodafone Germany.

The second requirement to fulfilling the vision is that the access technologies in question need to be capable of delivering on the requirements laid out by users. They need to provide suitable QoS, Mobility, Security and Link Capacity. The following is an overview of the three competing technologies.

Wi-Fi Mesh

The IEEE 802.11, wireless LAN (Local Area Network) set of standards specifies the physical and MAC (Media Access Control) layer of wireless communication technology designed to provide high bandwidth wireless data communication between computers in typical office-like environments. The un-tethered nature of wireless LAN makes it possible to provide network services like Internet connectivity in public “hotspots” at coffee shops, libraries and airports. The brand name “Wi-Fi” describes equipment which conforms to the Wi-Fi Alliance's wireless LAN interoperability tests. The 802.11 standard defines two modes ofoperation, Infrastructure mode, in which user-devices can communicate to each other via one or more dedicated Access Points (Typical of indoor office and hotspot usage) and Ad hoc mode, in which terminals can communicate directly to each other without using an Access Point (AP).

Wi-Fi expanded into the wide-area when it became evident that the technology can operate beyond its original intended use. People began using Wi-Fi outdoors using high gain antennas and achieving distances typical of equipment designed for WAN use. The current world record for the longest non-amplified Wi-Fi link is 272km. This was demonstrated in part by press reports in 2000 describing how hackers are exploiting Wi-Fi networks from afar using crude, home-made directional antennae. Researchers and enthusiasts experimenting with Wi-Fi in ad-hoc mode, began setting up Wi-Fi mesh networks, with the aid of intelligent mesh routing protocols. Protocols such as OLSR and AODV dynamically maintain routing information between Wi-Fi nodes, thus making it possible to send information across the mesh regardless of the constantly changing environment.

As opposed to the traditional operator model, Wireless community networks, set up by enthusiasts have since emerged, with the philosophy that the network infrastructure should not be owned by any single entity but by the community instead. Communities in developing countries are often frustrated with the lack of service provided by local Telcos and view the community owned network model as a solution to the problem. While wireless community networks like CuWireless, JAWUG and Freifunk have demonstrated that it is possible to efficiently provide data services like Internet connectivity, there is a growing expectation that these networks should also be able to provide voice and multimedia services as well. In South Africa, the Meraka Institute is researching the use of Wi-Fi mesh networks and the community owned model, in providing telephony and Internet connectivity to deep rural, under-serviced areas.

Wireless mesh networks are being set up by municipalities and smaller operators in cities around the world including Seattle,San Francisco and Knysna. Internet and Voice services are seen as yet another amenity and are provided in addition to traditional amenities like water, electricity.

A strong argument against using the technology for many operators is that, although capable of meeting the requirements for awidespread VoIP access technology; this was not intended in the original 802.11 specification. Equipment therefore needs to be enhanced with customized proprietary solutions to improve QoS, Mobility, security and availability. The IEEE however, is developing standards to meet these requirements, such as the 802.11s mesh standard expected in 2008.

The 802.11 base standards cannot guarantee QoS, because no mechanisms were adopted to distinguish voice from data, an obvious requirement for packet prioritization schemes . However, the recently adopted 802.11e and WMM (Wi-Fi Multimedia) standards were designed to enable QoS provision through enhancements to the access mechanisms. Two optional enhancements are made to the original Distributed Control Function (DCF), the first is the Enhanced Distributed Control Access (EDCA) and the second is the Hybrid Controlled Channel Access (HCCA) scheme. The EDCA distinguishes between four data categories: Voice, Video, Data and Background Data. 802.11e compliant equipment has been available since 2005 and should also be backward compatible with non-802.11e equipment.

Although the 802.11 base standard does offer client hand-off between access points, the speed at which it occurs is not fast enough to meet enterprise grade voice requirements(<100ms roaming delay),when coupled with security and authentication overhead. 802.11r, expected to be finalized by 2008, aims to address fast handoffs.

The original 802.11 implemented a relatively weak RC4 encryption scheme WEP, which is easily cracked due to its use of a static 40 or 104 bit key. WPA, a scheme proposed by the Wi-Fi Alliance also utilizes RC4 encryption but implements a constantly changing 128 bit key. WPA can usually be implemented with a software upgrade. The new 802.11i standard relies on a different,stronger encryption scheme, Advanced Encryption Standard or AES. The Wi-Fi Alliance's WPA2 standard certifies 802.11i compliance. Upgrading to WPA2 will require a hardware upgrade.

The link capacity of the network is influenced by several factors including bandwidth, radio technique and modulation technique. 802.11b for example can achieve 11Mb/s using the DSSS radio technique and QPSK modulation, while 802.11a can achieve up to 54Mb/s using BPSK, QPSK, 16 and 64-QAM modulation and 52 channel OFDM radio technique. The 802.11n
standard currently being developed will support data rates of up to 289 Mb/s on a 20MHz channel and 600 Mb/s on a 40MHz channel. This will be achieved by making use of the much anticipated multiple input-multiple output (MIMO) antenna systems. It is hoped that the standard will be ratifiedby 2008.

An argument both for and against Wi-Fi is that it operates in the unlicensed industrial, scientific and medical (ISM) radio bands. This includes the 2.4GHz ISM band used by 802.11b/g and 5.8GHz used by 802.11a. Equipment operating in the unlicensed band is attractive to community owned networks as well as smaller operators where licensing fees and other regulations are prohibitive, while larger network operators will prefer equipment operating in licensed bands for which they already have licenses and which operate in much less polluted, unshared spectrum. The cost and availability of end-user equipment is also essential. One of the factors that influenced the success of Wi-Fi is thought to be Intel's mass integration of Wi-Fi into laptops.


WiMAX or Worldwide Interoperability for Microwave Access is a Metropolitan area wireless network standard defined by IEEE802.16 and promoted by the WiMAX forum that aims to ensure interoperability amongst vendors. WiMAX will be available in two flavors, Fixed Location (IEEE 802.16-2004) and Mobile WiMAX (IEEE802.16e-2005).The fixed location version is designed for carriers as a means of delivering Broadband Wireless Access, in competition with DSL and for use by cellular and Wi-Fi operators as a network back-haul technology, particularly in rural areas. There is also a potential use as a private networking technology in competition with Wi-Fi. The first WiMAX forum certified products became available earlier this year and Intel intends to have its integrated “pro-wireless” chipset supporting WiMAX shipped with laptops in 2007.

QoS is built into the WiMAX channel access scheme. The access protocol distinguishes between four types of traffic, Unsolicited Grant Service for real-time voice and video, Real TimePolling Service, similar to 802.11e's HCCA, Non Real-time Polling Service and Best Effort for data services.WiMAX was originally designed as a fixed-location wireless technology; therefore the upcoming 1stflavor of WiMAX will not have any support for mobility. However, the 802.16e or Mobile WiMAX is being designed specifically with mobility in mind. The aim is to have zero-loss hand-off at speeds of up to 100kph. The standard has been ratified and is expected to make an appearance by 2008.

The security mechanisms employed on WiMAX will depend on the operator; however the original specification promotes the use of 168-bit 3DES encryption. Additionally, there are plans to implement AES as specified by 802.11i.

One of the controversial conversations around WiMAX is the transmission rate that it is expected to achieve. A theoretical data rate of up to 70-100Mb/s is mentioned as well as distances of up to 50km. However, the high data rate is not likely to be achieved at long distances. Field trials, such asthe one carried out by The Register in the UK have demonstrated much lower rates of up to 2Mb/s. Equipment operating in unlicensed bands will also not support full duplex mode, which will reduce the data rate significantly. The mobile version of WiMAX is also expected to deliver slower data rates, of up to 15Mb/s theoretically.

WiMAX is highly standardized, including factors such as QoS, Mobility and security, which should translate into reduced costs. Another plus is the radio technique used, OFDM (orthogonal frequency division multiplexing) as opposed to the less efficient CDMA (code division multiple access) as used in current 3G offerings. OFDM sends data across multiple radio channels, each at a lower rate, combating interference effects that become unmanageable at very high data rates with systems such as CDMA. OFDM therefore becomes more compelling as data rates increase. From a cost point of view, this means that it will be easier and cheaper to upgrade the data rate capacity of WiMAX networks than3G networks. However, it must be noted that the new mobile version of the WiMAX standard will employ different radio techniques (SOFDM-A) to the fixed version and will therefore also require different base stations and user equipment than the fixed location version.

The aimed integration of WiMAX into laptops from 2006 is based on a proven strategy used by Intel to increase Wi-Fi consumer uptake. This increased uptake should result in reduced end-user device costs through better economies of scale.WiMAXoperates in unlicensed as well as licensed bands. However, initial WiMAX equipment is expected to target operators and therefore will operate in licensed bands.


The recently launched broadband wireless offerings by the Universal Mobile Telephony System (UMTS) like HSDPA (High-Speed Downlink Packet Access) may ironically allow VoIP to gain ground on cellular telephony. Many mobile networks operators are upgrading their current 3G offering with HSDPA and promoting the technology as competition to fixedline broadband like DSL. A deciding factor in this situation lies with the handset manufacturers. Although SIP enabled GSM/Wi-Fi mobile phones are starting to appear, it is still unclear whether SIP enabled 3G/Wi-Fi models will follow or if these will allow SIP traffic to be transported by the 3G network interface or whether it will be limited to the Wi-Fi interface. Phone manufacturers have been pressured in the past by network operators to hold-back on similar features. Phone manufacturers may choose to not provide the actual VoIP over 3G functionality but allow 3rdparty application developers access to sufficient resources on Smart-phones to develop their own VoIP over 3G client applications. One of the most anticipated client applications of this nature is a Symbian based Skype client and it is unclear why it has taken this long to be developed.

The suitability of 3G+ technologies in carrying VoIP thus depends more on the willingness of the mobile operators than it does on the technology itself. Many mobile operators would prefer better utilization of their GSM investment to carry voice traffic. Some are even going as far as investigating ways of impeding voice traffic on their data networks should phone manufacturers choose to take the consumers side.

HSDPA introduces a new high-speed downlink shared channel (HS-DSCH) in addition to the existing CDMA channels (Dedicated Channel, Downlink Shared Channel and Forward Access Channel). This provides a 50-100% higher cell throughput than the original 3G depending on the size of the cell. HSDPA utilizes the typical UMTS QoS classes including Conversational, Streaming, Interactive and Background. Conversational traffic will still be carried on the traditional UMTS Dedicated Channel because the HSDPA specific scheduling does not provide considerable gains. As with previous UMTS offerings, mobility is well supported by the HSDPA standard.

HSDPA can provide down-link data-rates of up to 14.4Mb/s, depending on the modulation technique used, although initial implementations have been limited to 1.8Mb/s. HSDPA uses either QPSK or 16QAM modulation techniques.

HSDPA is CDMA based, which is thought to have a data-rate ceiling lower than that of its OFDM counterparts. This means that a significant investment will have to be made in the future to upgrade to higher data rate yielding OFDM based systems. High Speed OFDM Packet Access (HSOPA)/ Long Term Evolution (LTE) are proposed as the replacement for current CDMA systems, which will utilize OFDM. The commercial deployment of HSOPA is planned for 2008.


It is apparent that 2008 will be an important year to look forward to in terms of broadband wireless access technologies. Three competing technologies will most likely be available to operators and consumers. By then, Wi-Fi would have hopefully evolved to a standards based technology capable of providing adequate QoS, mobility, security and link capacity to support widespread VoIP services. WiMAX will have made its initial appearance in the fixed location version while Mobile WiMAX will also hopefully be ready for release. Lastly, the UMTS will be introducing HSOPA to the world. As to which technology will take VoIP into widespread service?, assuming that all of these technologies live up to the expectations that they are currently creating, they will all be more-or-less on an equal footing in terms of technology. The differentiator will be user perceptions based on cost, equipment availability and ease of use. Further into the future, with the aid of technologies like software defined radio, standards like IMS and 4G and Skype-catalysis, access technologies will not be the focus anymore, but rather the services provided on an ubiquitous network.