CDMA2000 promises to bring the first phase
of 3G services to market ahead of the mass UMTS camp, and
to do so with less disruption. But even a smoothly evolving
standard requires some changes and additions to the networks
As the first major change to the cdmaOne technology that has been implemented widely in the U.S. and Asia, it is important that CDMA2000 works. If it does, it could persuade TDMA and even some GSM operators in North America, China and Latin America to provide a CDMA2000 overlay onto their existing networks.
This first phase of the CDMA2000 upgrade, variously called 1XRTT, IMT MC1X, 3G1X, or just plain 1X, is designed to double current voice capacity and support always-on data transmission speeds ten times faster than those typically available today-some 153.6 kbit/s on both the forward and reverse links. Handset standby times should also increase by up to 50 percent. Users, meanwhile, will be able to benefit from enterprise and consumer applications that would normally require more bandwidth, including personal information management, telemetry, corporate intranet access, videoconferencing, gaming and music on demand.
Voice and data capacity
While the CDMA2000 specification allows for an evolutionary migration to later advances in CDMA2000 that use core IP networks and voice-over-IP, the current 1X migration requires relatively modest hardware and software upgrades to existing cdmaOne infrastructure. Even then, operators can upgrade to 1X without having to implement it throughout their entire cdmaOne network, which means they can upgrade certain hotspots that require voice capacity enhancements or higher data speeds.
This strategy aims to reduce technological risks by phased enhancements of networks with medium rate data services, and then later evolve to higher rate data services so as to avoid the uncertainties of return on investment that currently cloud the UMTS 3G vision. UMTS operators have in many cases over-extended themselves by paying billions of dollars for 3G spectrum to provide high-speed data services which the CDMA2000 world believes it can achieve without the same concomitant costs.
Of course, billions of dollars may well be spent by cdmaOne operators to introduce CDMA2000 and its later iterations to their customers. But that money won't (for the most part) be spent on buying spectrum, since CDMA2000 follows an in-band evolution strategy toward 3G based on existing worldwide cdmaOne band classes and infrastructure. It can thus be deployed in existing spectrum, overlaying its new feature sets and increased capacity on existing cdmaOne networks, and using an existing (and paid for) standard 1.25MHz frequency channel.
Handsets using CDMA2000 technology will also be backward compatible to existing cdmaOne networks, so current handsets and features will operate over next generation networks providing continued access both at home and while roaming. Operators which deploy 3G with 1X will still have roaming with worldwide CDMA operators on their cdmaOne networks. Dual mode handsets that allow TDMA/CDMA2000 interoperability may also augment the business case for a move by TDMA operators to CDMA2000.
The flexible migration from cdmaOne provides for a series of upgrades leading to CDMA2000's increased voice capacity and megabit data rates, allowing each operator to upgrade when its individual market requirements dictate, without having to significantly upgrade infrastructure or purchase new spectrum. The evolution of the air interface, capability of the core network (ANSI-41), and spectrum flexibility ensure this.
The business case that dictates the extent of its overlay of their cdmaOne network will, of course, differ for each carrier. This will depend on its capex spend and its capabilities, as well as varying degrees of subscriber differentiation. For some operators, the migration path may never need to reach the megabit level, while others, especially those in Asia, have a more immediate need to go as far down the 3G path as possible.
From the operator perspective, there is relatively little effort involved in upgrading a cdmaOne network to 1X. A 1X upgrade does not require any hardware 'forklifting', just some tweaking. They will be able to use most of their existing infrastructure to allow backward compatibility to IS-95 users.
To permit the advanced multimedia services that a 1X implementation promises, cdmaOne operators will have to do some retrofitting of base station channel cards with 1X cards, implement software upgrades in specific areas of their existing networks, and add some new IP-centric boxes to allow them to implement the data/voice 1X specification in part or in full. The approach to this will differ in form depending on what vendor equipment has been chosen, although most vendor approaches include ensuring backward compatibility with existing base stations with flexible future-proofing hooks for new technologies like intelligent antennas to boost capacity.
What stays relatively intact in a 1X overlay of a cdmaOne network is the MSC (mobile switching center), antenna systems, amplifiers, filters, and the existing data core of IS-95 networks called the IWF (inter working function). The IWF interfaces to the MSC via a Lambda switch and allows cdmaOne terminals access to a relatively limited packet data offering of around 9.6kbit/s. It is retained in a cdmaOne upgrade to 1X to provide this backward data compatibility. The MSC, though, requires a software upgrade.
At the base station level, new channel cards are required to support the increased capacity and enhancement in IP provisioning. New radio control software will also be required in the BS to enable the new channels, and where appropriate, software upgrades will also be necessary for base station controllers. Some operators may, however, want to implement separate boxes in the BSs in parallel with existing cdmaOne cards, primarily for cdmaOne backwards compatibility, and 1X testing imperatives.
To complete the 1X data/voice picture, operators will have to add additional hardware to define a new, comprehensive always-on high-speed IP-centric environment lacking in cdmaOne networks. Most of the new IP hardware in 1X systems is, unlike the GPRS world, off-the-shelf and does not need to be specifically designed or modified to exist in the wireless universe; this ensures investment protection.
These new components-which will exist in parallel with the IWF-include the PCF (packet control function) to act as an interface to the MSC and to the PDSN (packet data service node), another new 1X addition. The PDSN is a wireless-savvy router that acts as an interface to the internet and intranets.
Another new addition is the AAA (accounting, authentication and authorization) node which hangs off the PDSN to provide service profiles of end-user-specific data for packet data operations. The PCF was designed to allow the AAA and PDSN to be as off-the-shelf as possible as to ensure interoperability, and to allow a smooth transition to 3G services.
TDMA networks wanting to overlay 1X will require, among other things, new hybrid TDMA/CDMA2000 base stations, new BSC software, new dual mode TDMA/CDMA2000 terminals for users, a new PDSN, and new MSC software along with all the other components that make up a 1X network. Although the 1X upgrade and later 1X-EV-DO/DV iterations do not specifically require any increase in base station numbers, because of the increased fidelity required to maintain promised QoS speeds on the downlink and uplink, and the need to ensure low latency for multimedia applications, more BS and BSCs may in fact be necessary at the edge of the networks where signal to noise (S/N) ratios impact on data rates.
The necessity for improved amplifier and antenna systems to produce the required fidelity might also arise in some circumstances, although 1X promises up to 35 percent improvement in coverage using existing base stations, filters, amplifiers, and topography. It is anticipated that operators in densely populated areas will want to get the full benefit of 1X by upgrading to 1XEV where data demand is high.
Operators will also have to address voice capacity issues with other technologies like smart antennas and dynamic power control. The increase in RF capacity with 1X enables more simultaneous users, which means increase in revenue per cell site and higher return on investment, but also means that operators will need to mediate between simultaneous higher demand for voice and data.
The overlay design of 1X and its spectral efficiency protects valuable voice revenue, as the provision of medium and high rate IP-based services will not necessarily impact on voice capacity. Operators will have the flexibility to support this growing demand for voice and data without allocating bandwidth specifically for data before customer demand requires it.
Some vendors allow operators to protect their voice revenues by dedicating capacity to voice, which is usually given the highest priority in the network. This could also encompass throttling data rates up and down for users according to demand and specific business models.
Peak, symmetrical data rates of 153.6kbit/s in ideal conditions are possible, but field trails suggest that, as this is a shared resource, 130kbit/s average aggregate throughput is more likely. This will be shared by many users, which could decrease available bandwidth to individual users, much like heavy usage on an office LAN chocks individual demand. However, since usage scenarios suggest the likelihood of more bursty-type 'surfing' where all of the available bandwidth is not swallowed up in one gulp by a few users, fairly high throughputs nearing 130kbit/s are possible for individual users.
The paradigm will thus change for pure time-based billing to a combination of time and data usage. The billing scenarios may include flat-rate models that bundle talk-and-surf, or billing incorporating tiered levels of service with premium billing for QoS assurances for data and voice quality.
These include QoS support via low delay tolerance, data rate guarantee, and priority service access for multimedia applications. Operators could offer premium voice quality to certain users at a premium price using V2 voice mode (V1 is standard voice quality). They may also be able to bill for 'hot spot' coverage in convention centers, stadiums, malls and campuses, and for dedicated data rates and coverage to groups of users.
It is also very likely that they will emulate the Japanese i-mode success by taking a slice of profits for all micro-transactions successfully completed over their networks and portals. This in some cases will necessitate getting a banking partner or even a banking license to allow users to create and maintain mobile wallets for m-commerce payments on operator infrastructure. Location-based services are also likely to enhance operator revenues.
Later iterations of 1X will provide enhancements, such as increased data rates, radio enhancements of transmit diversity, smart antennas, direct spread, and multi-carrier, and inter-operation of network protocols. These include 1X EV-DO, or 1X-EV Phase One, which is a data-only enhancement to networks. Built primarily upon Qualcomm's High Data Rate (HDR) technology, it places voice and data on separate channels to provide up to 2.4Mbit/s access speeds. Operators have to install an incremental carrier in cell sites to provide these services. There's also the controversial EV-DV, or 1X-EV Phase Two, which promises stratospheric data speeds as high as three to 5Mbps. It requires a software and digital card upgrade on the existing 1X network.
1X upgrades will also include implementation of industry standard IOS V4 (version 4) interface to connect to the PDSN, giving wireless operators the flexibility of selecting the radio frequency (RF) mobile sub-system independently from the MSC.
The building blocks to 3G have been assembled. CDMA2000 is ready to go. Now it's time to put the new networks through their paces.
This article first appeared in CDMA World Focus June 2001 published by Informa Telecoms. For more information see www.telecoms.com