http://www.convergedigest.com/bp-me/bp1.asp?ID=312&ctgy=
Using Pseudo-Wires for Mobile Wireless Backhaul over Carrier Ethernet by Steve Byars, Vice President of Marketing
2/13/2006
The ongoing evolution of 3G services includes technologies such as HSDPA (High Speed Downlink Packet Access), HSUPA (High Speed Uplink Packet Access) and DVB (Digital Video Broadcasting)/ DAB(Digital Audio Broadcasting). Capital investment to support each new generation of technology represents only a part of the cost of doing business for operators.
Each new generation provides new capabilities, but also requires more bandwidth – both for the air interface and in the radio access network (RAN), sometimes referred to as the backhaul network. So, each new generation of services, especially new data services like HSDPA and HSUPA, also involves additional backhaul expenses. Whereas capital equipment and spectrum licenses are one-time expenses or fixed costs, ongoing backhaul expense in the RAN is a wholly different matter – a constant and increasing expense. As more services and more subscribers are added, bandwidth requirements increase and so do the backhaul costs. Optimizing RANs to address the growing and dynamic nature of today’s traffic, while simultaneously reducing backhaul expenses, is becoming more important than ever before.
The Need to Reduce Backhaul Expenses Backhaul has always been a source of high recurrent costs for wireless operators. According to The Yankee Group, wireless operators annually spend approximately $22 billion globally to lease transmission backhaul. It has also been estimated that backhaul transmission costs can amount to as much as 40% of network OpEx in 2G networks, and 60% or more in 3G. Today’s RAN backhaul network is made up of TDM-based leased lines for the backhaul of traffic from base stations to the base station controller (BSC), which then connect to the mobile switching center (MSC) and PSTN. Traditionally, mobile operators have leased T1/E1 lines from the incumbent PTT or LEC. The cost for these T1/E1 lines varies greatly, but is always a significant component of total OpEx. A single T1/E1 typically costs between $200 and $800 per month, but can vary greatly depending on a variety of factors. In fact, in some rural areas, incumbent carriers are charging as much as $2,500 a month for a T1 line. Base station backhaul capacity requirements today are typically in the 1-4 T1/E1 range and will grow to 4-8 T1/E1 in the very near future. At many base stations, the bandwidth requirements of HSDPA are pushing beyond what can realistically be achieved using today’s T1/E1 leased lines. Operating costs in the RAN are also impacted by the expense of overlay networks. Mergers and acquisitions within the service provider industry, together with multiple standards and multiple, often incompatible, vendor equipment, have all resulted in overlay networks. There are as many as three different base stations that have been deployed at some cell sites: GSM to provide voice and low-speed data, EDGE to provide high-speed data, and UMTS to provide very high-speed data. Not only are different generations of equipment used in these overlay networks, but different generations of 2G, 2.5G, and 3G traffic each utilize their own T1/E1 leased lines. Consequently, each of these T1/E1s is often partially filled, resulting in expensive inefficiencies. These inefficiencies are a direct result of the static-mapping nature of TDM transport – there is no way to dynamically move capacity from one base station to another (i.e. voice to high-speed data) even within the same cell site. What’s more, the TDM bandwidth is dedicated on a point-to-point basis from each base station all the way back to the BSC/RNC. Additional backhaul expense results from unused bandwidth that is often stranded in the “wrong place at the wrong time.” The Emergence of Packet Access TechnologiesThe need in the RAN for modern transport technologies that are better suited to bandwidth-hungry data services and that can furnish better flexibility and economies of scale has started to be recognized. In the GSM architecture, the R4 version of the UMTS standard already incorporates ATM with IMA as an interface for the NodeB (the UMTS equivalent of the BTS). However, traffic from these ATM interfaces is typically backhauled over point-to-point E1 leased lines. In both Europe and North America, ATM transport networks are occasionally used for backhaul transport.
But the continued evolution of mobile wireless services is toward technologies that are intrinsically packet-oriented, such as HSDPA and USDPA, and indicative that packet access networks, rather than TDM-based T1s and E1s, are the logical transport for backhaul in the RAN. A number of new technologies that are cost-effective and provide flexible, high-capacity backhaul transport are emerging as the frontrunners for next-generation packet access networks These new technologies include Carrier Ethernet (also known as Metro Ethernet or Optical Ethernet), xDSL, cable HFC, EPON/GPON, and broadband packet radio (including WiMAX) The Case for RAN over Packet AccessThe savings in operating expense are readily seen in a typical example. In this example, the mobile operator needs to deploy HSDPA in 500 cell sites of a medium-sized metropolitan area. The table below shows the current cost structure with only 4 E1s per site. However, these operating costs are likely to double with HSDPA, which can easily require 8 E1 lines or more. With E1 lines typically costing between $200 and $800 per month, this example assumes an average cost of $300 per E1. In comparison, a typical Carrier Ethernet service costs just $40 per Mbps each month. So, a monthly charge for the capacity equivalent of four E1s (8 Mbps) only amounts to $320 rather than $1200.
While this is only one example, it vividly illustrates the fact that using packet access in the RAN can reduce an operator’s backhaul charges by over 70%. Put into context, with backhaul accounting for 50% to 60% of all OpEx costs, a 70% reduction in backhaul charges represents a 40% decrease in overall OpEx, which is very significant indeed. In absolute numbers, the quantity of dollars saved (which flows straight to the bottom line) continues to increase as the bandwidth requirements of each cell site increase. The quantity saved also will increase as more cell sites are added to support UMTS and HSDPA because 3G base stations cover a smaller footprint than 2G base stations.In addition to providing lower operating expense while furnishing flexible, high-capacity transport, these packet access technologies offer a number of other benefits: Equipment costs are significantly lower than legacy equipment costs. Significant economies of scale can be achieved with a unified network, not only for mobile voice and data services, but also wireline business services. In addition, provisioning costs are less and do not require equipment upgrades. While continuing to add E1s to furnish projected requirements of 25-30 Mbps is an unsustainable paradigm, Carrier Ethernet, for example, offers a wide range of speeds up to 1 Gbps, in increments as small as 1Mbps, that can be provisioned on-demand or even by the operator through a web-based tool. Optimizing Radio Access Networks & Reducing OpExThe challenge has been how to support the continuing migration to 2.5 and 3G architectures and beyond by transporting both voice and data traffic over packet-access networks. Pseudo-Wires are the powerful enabling technology that meets this challenge. Pseudo-Wire solutions meet the demanding quality, latency, and clocking/synchronization requirements needed for reliable backhaul of mobile services. This technology has now been proven in field trials and live deployments on four continents. Pseudo-Wire solutions not only enable wireless operators to use new packet networks, they also give operators a wide choice among the multiple packet-access network technologies that are available – Carrier Ethernet, xDSL, PON, cable HFC, and even WiMAX. Finally, Pseudo-Wire solutions uniquely combine circuit emulation and service emulation (TDM plus HDLC, Frame Relay, and the ATM needed for 3G UMTS) to enable backhaul of any combination of 2G, 2.5G, and 3G voice and data traffic, all over a single packet RAN. Pseudo-Wire Capabilities in the RANThe term “Pseudo-Wire” comes from the IETF’s Pseudo Wire Emulation Edge-to-Edge (PWE3) working group, which is defining various types of Pseudo-Wires to emulate traditional and emerging services over packet networks.
The various types of Pseudo-Wires can be divided into two broad categories: circuit emulation and service emulation. Circuit Emulation Pseudo-Wires (CE-PWs) furnish transparent emulation of TDM services, including both payload and synchronization. Service emulation furnishes transport for frame-based and cell-based traffic, offering network capacity gains provided by statistical multiplexing.
Figure 2. Typical AXN deployment in the RAN.
Figure 2 shows a typical Pseudo-Wire deployment in the RAN. Pseudo-Wire Access Devices are deployed at the cell sites where they perform Pseudo-Wire adaptation (circuit emulation and/or service emulation) of mobile voice and data traffic for transport over the packet network to the BSC/RNC site. Pseudo-Wire Gateways are deployed at the BSC/RNC or MSC, where they terminate the Pseudo-Wire services and function as gateways to existing voice switches, the PSTN, and FR/ATM or IP data networks. At the cell sites, the Pseudo-Wire Access Devices interface with the existing BTS/NodeB equipment using T1/E1 ports. Each port can be soft-configured for circuit emulation for voice transport or service emulation to transport frame-based data services or frame-based voice. In addition, Pseudo-Wire Access Devices can offer unique T1/E1 interfaces that support the ATM traffic used in today’s UMTS deployments. Some Pseudo-Wire Access Devices also provide customer-facing Ethernet interfaces for future services and connectivity needs, such as those specified for the R5 NodeB.At the BSC/RNC site, the Pseudo-Wire Gateways offer not only T1/E1 ports for interfacing to legacy equipment, but also channelized DS3 and OC-3/STM-1 interfaces to connect with existing TDM-based voice switches. In UMTS applications, the ATM Pseudo-Wires (that originated as T1/E1s from the NodeB) are transported across the packet network and then consolidated in the Pseudo-Wire Gateway into ATM streams on OC-3c/STM-1 or DS3/E3 interfaces to connect with the BSC/RNC. The Pseudo-Wire Gateway also furnishes Gigabit Ethernet interfaces to hand-off data services at the RNC. Benefits of Pseudo-Wires for Mobile Wireless Operators
Any Packet Network
Pseudo-Wire is the enabling technology for transporting both mobile voice and data traffic over new high-capacity, lower-cost packet networks in the RAN. Pseudo-Wire solutions not only enable mobile wireless operators to use new packet access networks, they also give operators a choice among multiple packet network technologies in the RAN, including Carrier Ethernet, xDSL, cable HFC, and even broadband packet radio. In fact, Pseudo-Wire solutions are deployed and carrying live, revenue-generating mobile traffic over Carrier Ethernet, xDSL and broadband packet radio (pre-WiMAX). All Types of Mobile Wireless TrafficPseudo-Wires also provide support for all generations of mobile wireless services. Pseudo-Wire solutions are currently carrying live, revenue-generating voice and data for CDMA, 1x-EVDO, GSM, and UMTS services. The proven capability to support all of these services ensures a smooth transition from one generation of service to the next. It allows the operator to combine voice and data services from multiple generations of services and even from both TDMA and GSM architectures – all onto a unified packet-access network. As users transition from one architecture or service to the next, network capacity is not stranded. In fact, adding incremental network capacity is only needed to support actual growth in user traffic, not just to enable migration from one service to the next. A final benefit in this scenario is that, with a unified packet-access network, adding network capacity is simply a point-and-click provisioning process. Latency and Clocking
Latency is often an issue of paramount importance in wireless networks, not only due to its impact on voice quality, but because some signaling and control protocols cannot tolerate additional delay. End-to-end delay must be minimized while accommodating delay variations in the packet network. Pseudo-Wire solutions allow network operators to control a number of parameters that affect end-to-end delay, including frame/packet size, jitter buffer size, queuing priority, and frame/packet priority marking via 802.1q/p and DiffServ.
Another requirement for reliable backhaul of mobile services is synchronization or clocking. Cell sites must utilize a clock source that is synchronized with the overall network clock. While some CDMA deployments use GPS receivers at the cell site to furnish this clock, most implementations rely on clocking to be distributed by the RAN together with the voice and data traffic. This clock must not only be very accurate, but must meet strenuous jitter and wander requirements. The delay variations inherent to packet networks induce jitter; this jitter must be filtered out of clocks derived from Pseudo-Wire services. Extensive testing by service providers confirms that Pseudo-Wire solutions are able to deliver the low latency that meets the rigorous requirements of wireless signaling and control protocols. Pseudo-Wire solutions have also been tested and certified, meeting G.823 and G.824 clocking specifications for traffic interfaces. Some solutions even meet the more stringent clocking specifications required in GSM/UMTS applications, with accuracy measured to within 16 parts per billion.
Full-Service Solution
In addition to maintaining the quality and reliability of mobile services, any new technology must retain the native features and advanced services from which operators derive value. Pseudo-Wire solutions can be fully interoperable with the equipment already installed in the BTS, the NodeB, the BSC/RNC, and/or the MSC. They can preserve complete transparency for signaling and control-plane protocols, even the proprietary extensions to mobile protocols employed by some vendors. Operators can easily add Pseudo-Wire solutions to their existing architectures without changing the configurations of equipment already installed.
ConclusionThe exponential growth of the mobile wireless market is fueled by increased voice usage, a growing reliance on mobile phones, and bandwidth hungry voice and data applications. Mobile wireless providers need to expand their coverage and increase their capacity to handle the demand for services. In today’s difficult economic times, mobile wireless providers are concerned not only about meeting the growing demand for services, but also about the return on investment of their capital expenditures and the need to reduce operational costs, particularly in the RAN, as they expand and upgrade their networks from existing 2G technology to 2.5G and 3G.
By enabling the backhaul of all generations of voice and data traffic over a single packet-based RAN, a pseudo-wire solution offers mobile wireless operators a full-service alternative to TDM access and a means to substantially reduce backhaul expenses and enhance profitability.
About the Author
Steve Byars has worked in the voice and data communications industry for over 20 years. Prior to joining Axerra Networks, Steve was Director of Carrier Infrastructure at competitive intelligence firm Current Analysis. He was previously the CTO at Netrix Corporation, where he was responsible for the company’s VoIP business strategy. His experience also includes product development for Bell Laboratories in Holmdel, New Jersey, and GTE Communications Systems in Reston, Virginia. Steve holds a Masters Degree in Electrical Engineering from Stanford University, and currently has two patents pending for implementations of bus arbitration mechanisms.
About Axerra
Axerra Networks is the leading provider of circuit emulation and service emulation solutions over packet access networks. Axerra’s Pseudo-Wire solutions enable mobile wireless operators, cable MSOs, competitive service providers, and incumbent carriers to extend IP + legacy voice and data services in native format over Ethernet, IP, and MPLS networks. The result is greater operational efficiency, new revenue opportunities, and a smooth migration strategy to a single converged network, without stranding any revenue streams from profitable legacy services. With Axerra’s Pseudo-Wire solutions, service providers can convert any packet access network (Carrier Ethernet, broadband wireless including WiMAX, cable HFC, xDSL, PON, etc.) into a full-service alternative to TDM access. For more information, please visit: http://www.axerra.com/.
Send us your response to this article.
Learn How to Get Your Column Published on this Site
Subscription Info Marketing & Advertising About Us Contact UsCopyright © 2007 Converge! Media Ventures, Inc. All rights reserved.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment