One of the biggest challenges facing wireless operators today is the explosion of wireless data growth, and how their network can cope with the demand under CAPEX and spectrum constraints. This paper describes the issues, solutions available today, and why an optimized heterogeneous network is pivotal for tomorrow's wireless networks.
Wireless service providers worldwide are struggling under the strain of unprecedented traffic growth. This traffic growth is accelerated by the proliferation of advanced devices, services and applications. The challenges facing AT&T in the U.S. market are well documented as the operator has struggled to deal with a 5,000% growth in data over the last three years (AT&T report on LTE Americas, Sept. 2009).
The traffic demands for mobile broadband services vary across a wide variety of devices that can be categorized roughly in the following categories:
■ Simple feature phones: They are devices designed primarily for voice and text-based messaging services. They play a minor role in the mobile broadband market.
■ Smartphones: Devices such as the Apple iPhone and Google Android are equipped with larger screens, longer battery life, advanced OS, and a broad range of applications ― all features that encourage higher data consumption. Cisco Visual Networking Index report estimated the capacity consumption of an iPhone is estimated to be 30 times that of a simple phone.
■ Connected computing devices: Examples include dongle and PC card connected devices, embedded laptop devices, netbook, tablet PC and MID. Because of their large form factor and rich application environments, computing devices tend to consume larger service bandwidth than mobile devices. According to research published by the Cisco Visual Networking Index report, a laptop can consume up to 450 times more bandwidth than a simple phone and 30 times more than a smartphone. As laptop and netbook adoption continue to rise, primary mobile broadband for all devices will further fuel the wireless data growth mega-trend.
Operators are well aware of these challenges. Dick Lynch, CTO of Verizon Corp, the largest wireless service provider in the US, recently projected the following at the LTE Americas Conference in Sept 2009 (see Figure 1).
With traditional network design strategies, mobile service providers essentially have three primary capacity expansion tools, including:
■ Increase macro cell site density: Each cell split would require regulatory approval, new sites and civil work.
■ Technology upgrade to OFDMA-based 4G technologies WiMAX and LTE: To achieve a 3 to 4x increase in capacity.
■ Expand radio spectrum resources: Acquiring new spectrum can be expensive, limited by availability and subject to government regulatory timelines.
In each case the backhaul transmission network may also need to be upgraded accordingly.
The traditional capacity expansion solutions require high CAPEX investment, and Intel research (technology upgrade data from 3GPP, WiMAX Forum; backhaul and spectrum data from ABI, 10GEA) shows that they will not meet the network capacity demand forecasted through 2015 (see Figure 2).
To address these challenges, service providers are increasingly adopting a heterogeneous network architecture as depicted in Figure 3.
The heterogeneous network can consist of different cell scales which range from macro to micro, pico and even femtocells, potentially sharing the same spectrum. Nodes can deploy different access technologies such as WiMAX and WiFi, over both licensed and unlicensed bands.
The biggest advantage of a heterogeneous network is the significant gains in network capacity via aggressive spatial spectrum reuse and use of unlicensed bands. For example, co-channel femtocells can provide linear gains in air interface capacity with increasing number of femto-APs in a hybrid deployment. While the macro network provides coverage, small cells (pico and femto) are better suited for capacity infill and indoor coverage.
In addition, the small cells require no tower infrastructure or low lease cost, therefore drastically cutting the operational and capital expenditures (see Figure 4).
As described in previous paragraph, the heterogeneous network consists of a garden variety of nodes, each with potentially different RF and network characteristics:
■ Coexistence of numerous access technologies
■ Different network scales: macro (thousands of users), micro (hundreds of users), pico (tens of users), femto (<10)
■ Licensed and unlicensed spectrum
■ Nodes owned and operated by service providers, enterprises and consumers
■ Wired or wireless backhaul with guaranteed QoS or best effort QoS
■ A range of multimode devices including handhelds, laptop, netbook, MID, tablets, connected consumer electronics and machine-to-machine nodes
This type of network poses a new set of challenges including:
■ Cross tier interference: For example, a dense femtocell deployment poses significant interference to macro cells. While interference to data can be addressed via intelligent use of Fractional Frequency Reuse (FFR), interference to control signals requires new mechanisms.
■ Mobility management: Handover across small cells at moderate to high speeds hogs network resources, and can degrade user experience if not managed well. Heterogeneous networks must adopt an intelligent mechanism to manage inter or intra-tier handover. We expect multimode devices themselves to play a more significant role in mobility algorithm implementation.
■ Self-Organizing Networks (SONs) are essential for consumer deployed nodes like femtocells, and are important for managing inter-tier deployment.
■ Security management between nodes of different ownership (consumer, enterprise, operators).
■ Service continuity, QoS management and delivery across multiple tiers are essential for high performance and high capacity heterogeneous networks.
The Intel Labs has been doing extensive research to tackle these challenges. The team recently submitted a contribution in the IEEE 802.16 Plenary that outlines a number of key areas the IEEE community should focus upon in the next generation of 802.16 technology development, including:
(1) Develop co-operative WiMAX and WiFi networks to create virtual WiMAX carriers by aggregating spectrum across licensed and unlicensed space.
■ Create virtual WiMAX carriers through WiFi, for example, via an integrated WiMAX/WiFi femtocell.
■ Use the additional spectrum to improve diversity, and code rates with incremental redundancy.
■ Use the additional spectrum to transmit independent data streams to increase peak throughput.
(2) On the network side, provide connectivity and service continuity between different access protocols in the network.
■ Dynamically switch between WiFi and WiMAX to minimize interference and to optimize capacity and coverage.
■ Map applications to different access technologies based upon QoS requirements.
(3) On the client side, develop client co-operation where clients jointly transmit and/or receive information in wireless environments. For example, service providers may deploy clusters of stationary/nomadic WWAN clients with WiFi P2P connectivity. The idea is to exploit client clustering and P2P communication to transmit/receive information over multiple paths between the base station and client. This creates the potential for improvement in throughput, capacity and reliability without increased infrastructure cost. This solution is particularly interesting for M2M communications.
(4) On the radio access network, deploy MIMO.
■ Network MIMO algorithms enabled by central cloud processing
■ Cooperative MIMO and Distributed Antennas across tiers in the heterogeneous network
The proposals listed above are only intended to be a starting point for the development of future wireless broadband networks. Intel is also developing and evaluating other forward looking techniques such as interference alignment and innovative radio access network architectures. By sharing ideas, Intel hopes to galvanize the ecosystem partners to join in the exploration of enhancing the performance and cost structure of next generation wireless networks to accommodate the co-existence of a variety of access standards, multimode devices and disparate spectrum.
Mobile voice service is already considered a necessity by many consumers and is quickly moving to a commodity status in mature markets. Mobile data driven by consumer demand for rich mobile Internet data is now the primary driver for both consumer purchasing and network operator deployments. Mobile subscribers desire the same rich content available on their fixed broadband Internet connections ― including streaming video ― to be available on their mobile devices. Mobile M2M connections will also increase wireless data consumption. As wireless broadband continues to be rolled out worldwide, the demand for wireless data shows all signs of accelerating demand. Heterogeneous networks hold great promise to meet consumer demand, while providing optimum total cost of ownership for the network operators. However, these networks have many technical challenges at the air interface and network layers which the wireless community is working to address in standards forums. Heterogeneous networks will enable the cost effective deployment of high performance networks in order to bring wireless broadband to every corner of the globe.
(Caroline Chan, Director of Wireless Market Development, Wireless Program Office, Intel; Geng Wu, Director of Wireless Technology, Intel Labs)