Reader Forum: Addressing the capacity crunch

Editor’s Note: Welcome to our weekly Reader Forum section. In an attempt to broaden our interaction with our readers we have created this forum for those with something meaningful to say to the wireless industry. We want to keep this as open as possible, but maintain some editorial control so as to keep it free of commercials or attacks. Please send along submissions for this section to our editors at: [email protected] or [email protected].
The ascending prevalence of smartphones that feature large screens, geo-location capabilities and user friendly interface is set to change the wireless industry as the long awaited predictions on take up in mobile data services become reality. “All you can eat” pricing plans have further contributed to this take up. Moreover, with many markets reaching subscriber saturation, network operators are in a zero-sum game where one’s gain is another’s loss. Limited prospects for large subscriber addition are placing hope on data services and mobile Internet as the basis for increased future revenue. These trends have combined to increase the traffic carried on wireless networks. Today, news abounds of wireless network operators struggling to meet the demand for data services while meeting acceptable network performance.
As this trend is expected to continue well into the future, the wireless network operators are at a quandary of how to resolve the impending capacity crunch. So, what can be done to address this problem?
If we are to look at the wireless network for what it is, a transportation system for traffic data, we can identify three techniques to control the amount of traffic and capacity of the system:
1. Traffic offload;
2. Traffic management and regulation;
3. Increasing system capacity and efficiency.
The methods available today at the disposal of the wireless industry fall within one of the three categories above.
Traffic offload: Two emerging techniques include offloading data traffic onto Wi-Fi networks or use a “femto” base station to offload traffic from macro cells. Wireless technologies, and specially packet-based ones have already defined primitives to integrate with Wi-Fi and other wireless technologies. The migration from circuit-switched connection based protocols to packet-switched connection-less protocols (e.g. IP, Ethernet) will facilitate the inter-working between networks with different types of air interface. In Austria, for example, A1 provides Wi-Fi routers to its broadband bundle customers and software for laptops that automatically switches from 3G service to Wi-Fi when the subscriber is at home.
Femto base stations (defined as base station in a consumer premise which supports a very limited number of users) can potentially serve the purpose provided a viable business case is made through sufficiently low unit price and availability of backhaul through the DSL or other wired connection in the premise. These techniques are not without their challenges. For example, Wi-Fi offload can lead to lost revenue. Femto base stations present challenges in interference control, device management and other technical and operational aspects. One of the main tenants of “self-organizing networks” concept is to facilitate the management of femto base stations.
Traffic management and regulation: This includes manipulating the carried traffic to meet the capacity constraints of the wireless network. One technique includes using deep packet inspection (DPI) techniques to cap (limit amount of user data), block (block spam and bandwidth consuming applications), throttle (control network traffic at specific periods) or shape (control of traffic volume and flow rate) the carried traffic. Compression and transcoding/transrating are also applications that fit this category. Other techniques of traffic management include content storage at the edge which particularly applies to bandwidth consuming video traffic. Moreover, traffic management and regulation does not have to be a technical feature, but a marketing and commercial issue as well. For example, pricing plans serve to manage and regulate consumer behavior. Switching between flat-fee rate to volume and time based rates would also regulate traffic flow on the network. As wireless networks are typically designed to handle expected traffic during the busy hour with a certain amount of call blocking rate, it becomes an expensive proposition to continue scaling the whole infrastructure to meet busy-hour requirements when additional tariffs during that hour may curtail usage and reduce overall infrastructure buildout costs. (Many electric utilities now implement time-of-day billing through smart meters to even out the usage over time. Similar techniques can be applied to wireless networks as an extension to the off-peak discounts many wireless plans provide today.) Policy control and management systems allow the network operators to assign complex rules on charging and billing that involve several concatenated rules related, for example, to the subscriber grade of service, the day and time of use, the location and the type of device.
Traffic management and regulation techniques also have certain detractions. For instance, in-line-DPI techniques need to occur at line rate and introduce as little latency as possible. Moreover, certain applications are hard to detect by standard DPI techniques leading to the development of “deep flow inspection” techniques that use behavioral analysis and pattern recognition to identify applications.
Increasing system capacity and efficiency: This is perhaps the first thing one can think of when system capacity is reached. It can involve simple upgrades such as adding another frequency channel overlay in congested cell to more major upgrade such as migrating the network to a new standard with higher spectral efficiency. Hence, LTE promises large enhancements in spectral efficiency because of several physical and medium access control layer innovations that include using orthogonal frequency division multiplexing, variable modulation and coding rate, dynamic link adaptation, capacity boosting MIMO techniques, and hybrid automatic repeat request among other techniques.
Another subcategory of capacity enhancing techniques includes deployment of below-clutter pico base stations and distributed antenna systems. This in effect takes the concept of cell-splitting to its full potential as limiting harmful interference works to enhance overall network capacity. Hierarchical networks can then come into their full potential through an overlay of micro and macro cells to serve mobility users.
Finally, more spectrum is another way to increase system capacity. This has been recognized by most regulatory bodies and in particular by the Federal Communications Commission which has called to release up to 500 megahertz of spectrum by 2020 in their broadband plan, including 300 megahertz to be made available by 2015.
The examples presented above are not without their detractions. For instance, Pico base stations require backhaul whose price and availability could make the business case unattractive. DAS requires fiber optical cables between the central processing unit and the remote radio units which can be an expensive proposition in many locations. Spectrum is scarce and expensive, particularly in the sub-2 GHz frequency band where signal propagation properties are most advantageous for mobile communication applications.
The presented framework for addressing the capacity crunch can be seen through the prism of each of the three networks comprising the end-to-end wireless network: radio access (RAN), core (CN) and backhaul networks. Larger and faster routers are needed to complement a faster RAN. DPI has traditionally been placed in the core network, but there are benefits to migrate this function closer to the edge and placing it in the RAN particularly as the cost of fast processors falls. The implementation of QoS intelligence on the backhaul
network ensures end-to-end QoS for
user applications and provides an additional degree of freedom to manage capacity over what many experts consider a true bottleneck in wireless network evolution. Hence, functions that have typically been the preserve of one network can now be migrated and applied in another part of the network with potential benefits to overall performance as well as providing a great potential for future product development and differentiation to address critical issues arising in packet-switched wireless networks.
The challenges and solutions above are exactly what will make the wireless industry a very interesting one as the convergence of fixed and mobile services and increase in data traffic combine to force new innovations to address the capacity crunch. The fundamental premise of packet-switched connection-less wireless networks together with advancement in user devices and applications that drive higher network utilization will combine to reshape the industry as new solutions emerge and companies jostle to claim their space in a new industry order.