Communications Network Research Institute

Mobile Data Offload

Mark Davis

Data Tsunami on 3G Networks

Mobile broadband traffic has grown dramatically in the past two years. The main reasons for this are the advent of mobile broadband service offerings, the increasing use of smartphones, the availability of flat-rate voice and data bundles, and higher demand for entertainment services like YouTube, Apple’s iTunes and services such as video streaming from television networks. For example, in a recent market forecast Cisco Systems [1] has reported that wireless data traffic throughout the world has increased by 160 percent over the past year to 90 petabytes per month (or the equivalent of 23 million DVDs) and has predicted that by 2014 this figure will increase 39-fold to about 3.6 exabytes per month (or 3.6 billion gigabytes). This report also predicts that video will be responsible for the majority of this traffic growth and will account for 66 percent of global mobile data traffic by 2014. A single laptop can generate as much traffic as 1300 basic-feature phones, and a smartphone creates as much traffic as 10 basic-feature phones. The iPhone can generate as much traffic as 30 basic feature phones. AT&T, the exclusive carrier in the USA for the iPhone, reported recently [2] that it has seen wireless packet data on its network increase by 5,000 per cent over the last three years. This growth is expected to accelerate following the recent introduction of the 3G-capable iPad where according to AT&T between 400,000 and 500,000 iPad owners have signed up for the 3G data plans and that between 75% and 80% of these subscribers have chosen the highest data plan available.

The explosion in the growth of mobile broadband data is being described by mobile network operators (MNOs) as a “data tsunami” that is threatening to overwhelm their 3G networks [3]. Already some networks are crumbling under the stress where this extra traffic, particularly in densely populated regions, is already causing problems for consumers in the way of dropped calls and slow Internet access. Moreover, the next generation of 4G or LTE networks are unlikely to meet this exponentially growing demand for capacity. At the same time there is also greater competition for subscribers with a corresponding downward pressure on revenue per subscriber. All of which means that MNOs desperately need solutions that help them reduce network congestion while also helping them reduce costs and retain customers. One such solution is mobile data offload which holds the promise for rapidly delivering reduced congestion and significant cost reductions for operators while also improving the operator’s reach to its customers.

Mobile data offload involves the use of complementary network technologies (i.e. Wi-Fi hotspots) for delivering data originally targeted for cellular networks. Given the urgency with which additional data capacity is required, Wi-Fi offers “time-to-capacity” advantages that cannot be matched by building out cellular capacity. The availability of spectrum is often the limiting factor for mobile operators. Wi-Fi allows data traffic to be shifted from expensive licensed cellular bands to exploit the 2.4 GHz and 5 GHz unlicensed bands. With a massive installed base Wi-Fi is proving to be a natural complement to the mobile network and seems destined to play a long term role alongside LTE. A recent ABI Research report [4] states that about 16 percent of mobile data is being diverted from mobile networks today and is expected to increase to 48 percent by 2015. As data traffic itself will have grown by a factor of 30, this means that offloaded data is expected to increase a 100-fold in this period.

While mobile data offload solves the immediate problem of alleviating congestion on 3G networks, MNOs still face the challenge of delivering huge quantities of video content to their subscribers over WLAN networks. Video is particularly problematic owing to its large bandwidth requirements coupled with the QoS expectations of the end users. Delivering video traffic over WLANs has been an active research area for the last five years or so and the same challenges still remain, namely how to provision bandwidth for video traffic and how to guarantee QoS. However, mobile data offload will introduce another challenge, i.e. how to accommodate mobility as the bulk of the video content will in future be delivered to smartphone users. Gartner predicts by 2013 that the combined installed base of smartphones (and built-in browser enhanced feature phones) will be greater than the installed base of PCs [5].

Mobile Data Offload and Wi-Fi

Unlike cellular networks where the bandwidth allocation is deterministic (i.e. achieved through frequency, time, and code division multiplexing), Wi-Fi users must compete for bandwidth using a distributed contention based channel access mechanism (MAC). The MAC defined in the original IEEE 802.11 standard that defines Wi-Fi ensures that all users compete equally for access which means that it is not possible to provision bandwidth or give any QoS guarantees. The ratification of the IEEE 802.11e MAC Enhancement standard in 2005 was an attempt to address this shortcoming by introducing a differentiated services mechanism (known as EDCA) that allows for a channel access prioritization. Under the EDCA mechanism users no longer compete equally for bandwidth, but instead compete on the basis of the relative priority of their traffic. The IEEE 802.11e standard only partially solves the problem of provisioning bandwidth and guaranteeing QoS on WLANs as the EDCA really only provides for a greater degree of control over the operation of the MAC. Essentially the EDCA is an enabling mechanism for realizing bandwidth provisioning and guaranteeing QoS on WLANs [6]. Numerous schemes that employ the EDCA have been proposed that seek to deliver high quality video on WLANs [6-8]. All of these schemes are based upon appropriately configuring the EDCA control parameters to ensure that the video traffic stream obtains the bandwidth it requires in order to achieve the desired QoS.

The original IEEE 802.11 WLAN standard was intended as a low cost, easy to deploy wireless extension to cabled Ethernet LANs and as such was designed to offer a bandwidth on demand, best effort service only. Under the channel access mechanism specified in the standard, all WLAN nodes must contend for access which means that the bandwidth is shared between competing nodes irrespective of the relative priority of a node’s traffic. The introduction of a QoS mechanism into the standard (i.e. the IEEE 802.11e standard) in 2005 was intended to address this shortcoming in the original standard by allowing for a basic prioritization of data packets based on four categories - voice, video, background, and best effort. However, five years after its introduction the networking industry remains unconvinced about its ability to deliver QoS [9]. Although, all modern WLAN equipment supports the IEEE 802.11e standard (or more accurately the industry’s Wi-Fi Alliance WMM specification), few applications actually utilize it. It is used in some high end enterprise networks for deploying wireless VoIP solutions. Its poor take up within the industry is due primarily to the fact that there are too many parameters to configure and furthermore it is not clear what values should be used for these parameters. The standard recommends a set of fixed default parameter values, however the performance benefits from using default values these have been mixed.

The fundamental problem with using the IEEE 802.11e/WMM QoS enhancement mechanism is that there is a complex relationship between the various configuration parameters which is poorly understood. In particular, it needs to be recognised that the IEEE 802.11e/WMM standard provides for a QoS enabling mechanism only, it cannot in itself deliver QoS. Instead, it needs to be incorporated into a bandwidth/QoS provisioning scheme that will adaptively configure these parameters to deliver the required bandwidth and QoS. To date, this still remains one of the biggest challenges in this field, namely how to guarantee QoS and conduct efficient bandwidth allocations in IEEE 802.11e contention-based WLANs [10,11].


[1] Cisco Systems, Visual Networking Index (VNI): Global Mobile Data Traffic Forecast Update, 2009-2014, February 9, 2010.

[2] Stephen Lawson, AT&T Wi-Fi use soared 30% in Q2, in ComputerWorld, July 22, 2010.

[3] The Nielsen Company, Quantifying the Mobile Data Tsunami and its Implications, June 30, 2010.

[4] ABI Research, Mobile Network Offloading Free and Carrier Wi-Fi, Femtocells, Caching, Core Offloading, and Media Optimization, August 16, 2010.

[5] Gartner Press Release: Gartner Says Worldwide Mobile Phone Sales Grew 17 Per Cent in First Quarter 2010, May 19, 2010.

[6] Shafiqul Karim, David Green, Michael Rumsewicz, Nigel Bean, “Flexible Throughput Management in IEEE 802.11e Wireless LANs,” in Proceedings of IEEE International Conference on Networks (ICON), October 2007.

[7] Richard MacKenzie, David Hands and Timothy O’Farrell, “Packet Handling Strategies to Improve Video QoS over 802.11e WLANs,” in Proceedings of IEEE 20th International Symposium on Personal, Indoor, and Mobile Radio Communications (PIMRC), September 2009.

[8] Pilgyu Shin, Kwangsue Chung, “A Cross-Layer Based Rate Control Scheme for MPEG-4 Video Transmission by Using Efficient Bandwidth Estimation in IEEE 802.11e,” in Proceedings of International Conference on Information Networking (ICOIN), January 2008.

[9] Tim Higgins, Does Wi-Fi Multimedia WMM Really Do Anything? in Small Network Builder, 29 May 2009.

[10] Cheikh Sarr, Claude Chaudet, Guillaume Chelius, and Isabelle Guerin Lassous, “Bandwidth Estimation for IEEE 802.11-Based Ad Hoc Networks,” IEEE Transactions on Mobile Computing, Vol. 7, No. 10, October 2008, pp. 1228-1241.

[11] Yang Xiao, Frank Haizhon Li, and Bo Li, “Bandwidth Sharing Schemes for Multimedia Traffic in the IEEE 802.11e Contention-Based WLANs,” IEEE Transactions on Mobile Computing, Vol. 6, No. 7, July 2007, pp. 815-831.

If you have any comments or suggestions please feel free to contact us at cnri@cnri.dit.ie