Always-on access to broadband services has become a necessity for many business and residential users around the globe. The availability of mobile broadband is thus indispensable for people on the move (complementing their existing high-speed wireline connection in the office or at home) -- extending well beyond the simple voice connectivity that legacy mobile networks used to provide.

Over the years, with the above in mind, the capabilities of mobile networks have gradually been improved:

• The first cellular networks were designed to carry voice calls only -- using analog, circuit-switched technology.
• With second-generation (2G) networks (such as Code Division Multiple Access [CDMA] and Global System for Mobile [GSM]), the resources became dynamic -- allowing sharing among several simultaneous calls.
• Packet-based data services were introduced with the so-called 2.5G networks, which included General Packet Radio Service (GPRS) and CDMA 1X -- capable of delivering peak speeds up to 100 kbps.
• As mobile networks moved into their third generation (3G), higher throughput and lower latency for transporting data traffic were realized -- while the range of technology options proliferated, with Evolution Data-Optimized Revisions A/B (EvDO Rev A/B) and wideband CDMA/ High-Speed Packet Access (WCDMA/HSPA).
• Today, as Long Term Evolution (LTE) networks begin to be implemented, operators hope to benefit from considerably faster data rates (with theoretical peak download speeds of up to 173 Mbps), reduced latency, improved spectral efficiency and a simpler/flatter IP architecture -- which collectively enable the delivery of new mobile broadband services.

These significant improvements in wireless infrastructure -- and the capabilities they bring -- complement the value of existing networks; integrating wireline and wireless applications enables service providers to offer a seamless, always-on, end-user experience, and helps control infrastructure and operations costs.

As a matter of fact, there are several areas that see wireline and wireless technologies complementing each other to provide the best possible service coverage to end-users. Two important examples include: mobile backhaul of wireless macro cells, and the placement of femto cells inside residences. We explore both in this article -- linked to the evolution and promise of LTE.

Wireline/Wireless Convergence -- Boosted by LTE Adoption
Before discussing how wireline/wireless convergence works in practice, let’s take a quick look at one of the main enablers: the adoption of LTE.

In an effort to differentiate their offerings and increase profitability, mobile service providers have begun to augment traditional, low-bandwidth services like e-mail and short message service (SMS) with new mobile broadband services -- such as real-time video, gaming, music, and other rich multimedia applications. Alcatel-Lucent’s primary research on 4G service demand suggests indeed that there is strong interest in adopting entertainment-related applications (among residential users) and productivity-enhancing services (among business users). Introducing these new, LTE-optimized applications therefore creates a dramatically improved market opportunity: with an initial focus on large enterprises, and a market entry strategy founded on carefully packaged and priced applications, operators could achieve double digit growth in both consumer and enterprise revenues.

Obviously, all of this is closely linked to the creation of an LTE ecosystem -- one that is rich and open, and fosters the development of innovative mobile services and applications -- through the availability of new application programming interfaces (APIs) which allow developers to have access to specific network assets as well as enablers such as location, presence, and security.

It is quite clear, however, that as soon as mass-market LTE adoption becomes a reality, the resulting explosion of data traffic will also lead to an increased wireline/wireless network convergence through the concept of mobile backhaul.

LTE Backhaul Requirements
The backhaul network, linking base stations to the mobile transport switching office (MTSO) is the largest and most often overlooked gating factor for delivering mobile broadband services. Inadequate backhaul planning can lead to bandwidth bottlenecks and customer dissatisfaction. To support an increase in airlink capacity, faster links on the backhaul network are required. At the latest CTIA show, for instance, Verizon Wireless said that it aims to get 100 Mbps links to its cell sites as it moves to LTE.

In anticipation of future data services growth, wireless operators have started to deploy radio access equipment incorporating high-speed Ethernet interfaces that mate with scalable all-IP networks. Conveniently, wireline operators have already been deploying Ethernet/IP architectures to meet the needs of their own broadband services.

With both types of operators and networks heading down the path of all-IP, there is now significant harmony in their approaches and architectures. The ongoing migration from TDM to IP, however, is not without its challenges -- with LTE backhaul imposing additional requirements:

Higher Capacity
As LTE cell sites are rolled out, it is expected that backhaul bandwidth requirements will range on average from 50 to 100 Mbps -- a significant increase over 2G/3G systems. This evolution further reinforces the need to scale transport capacity at lower cost, and increase transport efficiency by migrating to packet backhaul. As LTE matures, several hundreds of Mbps will have to be foreseen (e.g., at cell sites serving multiple operators).

Multi-Service Transport
As operators gradually introduce LTE, they will leverage existing 2G/3G sites where possible. This site reuse means that the backhaul network must be scalable enough to support the co-existence and cumulative capacity of LTE with either CDMA or W-CDMA networks. And it means that the backhaul network must support a combination of TDM, ATM, and Ethernet/IP traffic.

Quality of Service (QoS)
For a reliable delivery of real-time services (e.g., video, VoIP), the control of QoS-related parameters (i.e., jitter/delay) needs to be enforced to meet the “deterministic behavior” of TDM circuits when transported over fully loaded packet links.

Synchronization
The migration towards IP/Ethernet networks that do not transport the clock reference transparently requires strict QoS implementation rules to keep delay and jitter within ITU-T recommendations, and to ensure recovery of the clock reference. This is accomplished via 2 types of synchronization:

1. Frequency Synchronization
Cell site equipment requires an accurate clock to set its RF frequencies in support of call hand-offs. All cellular radio systems require frequency synchronization of ±50 ppb (parts per billion). There are 2 instances when base station synchronization is provided by an external source: the base station is synchronized by GPS, or via a leased line (e.g., BITS interface).

2. Time of Day (ToD) Synchronization
With few exceptions, frequency-division duplex (FDD) radio systems require only the frequency synchronization discussed above. However, time-division duplex (TDD) radio systems also require Time of Day (ToD) synchronization for accurate framing of timeslots. Typical requirements are 1 to 2.5 µs, with ToD being delivered by either a GPS subsystem on the base station equipment or by the IEEE 1588v2 PTP (which is capable of supplying both frequency and ToD synchronization) over-the-top of the backhaul network.

Leveraging Fixed Broadband Access for Mobile Backhaul
The access portion of the mobile backhaul network can utilize fiber, copper, or microwave technologies for the Last Mile connection to the cell site. Often, the most cost-effective option is to leverage the broadband access network that has already been engineered to hit the cost points for the delivery of residential Triple Play services.

More specifically, the incremental cost of backhauling cell site traffic on the existing residential network will usually be less than building out a separate overlay network. The savings mainly derive from:

• Outside plant (re-use of existing copper pairs and gigabit passive optical network [GPON] splitters, feeder fibers, and optical line termination [OLT] ports).
• Access nodes (e.g., DSLAMs and GPON or point-to-point [PTP] fiber OLTs, providing low-cost broadband aggregation to minimize port consumption on relatively expensive switches and routers).

This is the rationale for leveraging the fixed broadband access network for mobile backhaul.

But can a residential network effectively meet the backhaul requirements for 3G and the evolution to LTE?

Fixed broadband access solutions have been deployed to deliver high-speed Internet access, (high-definition) video and (lifeline) voice services to customers around the world -- all while supporting business services such as T1/E1 and Ethernet. The stringent requirements that these services place on the access network almost completely overlap with those for mobile backhaul. So, in the course of pursuing their own broadband services business models, wireline service providers have actually been deploying a mobile-backhaul-ready infrastructure as well, with the requisite QoS and multi-service transport capabilities.

xDSL (digital subscriber line) technologies mine large bandwidths out of existing copper pairs. ADSL2+ (and its ADSL predecessor) is the most widely deployed access technology to deliver residential broadband; bonded SHDSL is the most recent symmetric technology for business access. Both have already been successfully deployed for mobile backhaul: ADSL2+ (and bonded ADSL2+) as well as bonded SHDSL are capable of providing tens of Mbps, making them excellent short- to medium-term tactical solutions for low-cost, fast time-to-market relief of the backhaul bottleneck problem.

However, to fully satisfy the bandwidth needs of 3G and the transition to LTE, we need solutions capable of supporting several hundreds of Mbps.

Again, wireline service providers are already expanding the boundaries of DSL to deliver (high-definition) IPTV services over existing copper by deploying small, fiber-fed access nodes closer to their end-customers and using VDSL2 to squeeze more bandwidth out of the shorter copper pairs. With the introduction of 2-pair-bonding, which multiplies bandwidth by a factor of 2, over 100 Mbps downstream over VDSL2 is now possible. Factoring in future vectoring and multi-pair bonding, VDSL2 will thus be capable of supporting several hundreds of Mbps downstream and about 100 Mbps upstream over distances of 500-1,000 meters.

Of course, the ultimate transport medium is optical fiber, with fiber-to-the-home (FTTH) deployments extending fat optical pipes to end-users, either over point-to-multipoint (PON) or PTP topologies. In the case of GPON, 2.5 Gbps downstream and 1.25 Gbps upstream pipes are shared by nominally 16-32 subscribers, depending on optical splitting and service take rates. But even when accounting for aggressive residential video service take rates, very high penetration of high-definition television sets in the home, and a very high proportion of unicast streams, a residential GPON will have enough bandwidth headroom to provide a full 1 Gbps (or more) to 1 (or more) cell sites located in that neighborhood. For point-to-point fiber, either 100 Mbps Fast Ethernet, or better, Gigabit Ethernet, can serve the cell site.

In short: VDSL2, and especially FTTH networks, represent long-term strategic solutions for the backhaul of LTE cell sites.

As mentioned before, another important consideration for the backhaul network is synchronization. Traditionally, when not provided by the GPS network, frequency synchronization is provided by the TDM backhaul connection. When this is retired in favor of the packet-based backhaul network, it becomes necessary for the packet network to provide it, something that packet networks were not originally designed to do. New standards have been created to solve this: Synchronous Ethernet, a physical layer timing analogous to TDM networks; and IEEE 1588v2 Precision Timing Protocol, which operates at the packet layer.

Here again, we find that fixed broadband access networks already possess standardized timing capabilities. Over copper, there is network timing reference (NTR) which is standardized for ADSL2+, SHDSL, and VDSL2. SHDSL NTR is already commercially available, while VDSL2 NTR is coming to market this year. The GPON standard requires an 8 kHz timing reference to be transported over the PON. PTP fiber will use Synchronous Ethernet. All 3 of these are physical layer timing architectures, not susceptible to packet layer impairments (e.g., congestion), and are adequate for mobile backhaul.

The main message is that today's fixed broadband access networks are already suitable for mobile backhaul, with most of the ongoing refinements involving the optimization of xDSL customer-premises equipment (CPE) or GPON optical network terminals (ONTs) located at the cell sites: the cell site gateways. One example of these ongoing innovations is taking an existing packet-based cell site gateway and enhancing it with xDSL and GPON uplink capability.

Wireline/Wireless Convergence - The Next Step: Femto Cells
Using fixed broadband access to backhaul mobile traffic from traditional (macro) cell sites is one way to support the trend towards fixed-mobile convergence. Another interesting development is the femto cell. Analogous to very small base stations, femto cells extend the coverage and capacity of a wireless network into homes, buildings, and fringe areas, enabling maximum bit rates (often much higher than when connected to a full-size base station). On top of this, they increase overall wireless network capacity by off-loading traffic from the full-size base stations onto the femto cell network.

Femto cells can provide mobile users with several advantages:

• Improved Internet access due to higher bandwidth and lower latency (the equivalent of having a base station in the house).
• Coverage for voice calls where none existed before, including seamless mobility due to hand-off coordination and the reliability of licensed spectrum.
• Local file sharing between the user's mobile and local devices -- without consuming network capacity using local IP routing.
• Blending of incoming calls between wired phone and mobile devices (for example, one-number dialing).

Although femto cells are an extension of the wireless network, when deployed in homes there is no getting around the need to backhaul them by means of a fixed-line broadband-access network - leveraging past investments.

The Way Forward
Service providers will need to focus on the specific convergence needs of particular market segments, with a commitment to adopt the right technologies that meet those needs. Often, the most cost-effective option in The Last Mile will be to leverage the existing, fixed broadband access network -- having already been engineered to hit the cost points for offering residential Triple Play services. In the core, the drive towards an all-IP network, in which many of the subsystems are identical across fixed and mobile, will ease wireline/wireless convergence. This, combined with the high speeds and low latency afforded by LTE, enables service providers to offer integrated wireline and wireless services that consumers and enterprises demand, while lowering their operations costs.

About the Authors

As senior product marketing manager, Ed Harstead leads the marketing activities for mobile backhaul, fixed-mobile convergence and application enablement for Alcatel-Lucent's Wireline Networks product division globally. For more information, email ed.harstead@alcatel-lucent.com or visit www.alcatel-lucent.com.

Hector Menendez is a member of the wireless solutions marketing team at Alcatel-Lucent. For more information, please visit www.alcatel-lucent.com/lte and www.alcatel-lucent.com/bba-mbh.

What’s your take on this subject? Leave a comment and get the conversation going.