- Buyers Guide
There is little doubt that the prospect of Mobile WiMAX providing an alternative mobile broadband platform to those of the 3GPP and 3GPP2 has accelerated the competitive development of LTE. LTE and Mobile WiMAX may be technologically quite similar, indeed many vendors support both platforms, but there remain deep-seated political and commercial conflicts causing some suppliers and operators to stake their futures on one technology or the other. The early weeks of 2008 have seen a flurry of announcements from both camps, designed to emphasize the advanced nature of the roadmap towards true 4G.
The Ericsson-led LTE community is stepping more aggressively into WiMAX’ traditional TDD territory (see Wireless Watch January 15 2008), and despite ambivalence by the WiMAX Forum about making official statements, 802.16e will respond with an increased focus on FDD profiles this year. The 3GPP approved the official technology specifications for the LTE RAN, which will now be included in the forthcoming 3GPP Release 8. This approval is only three months later than originally promised, a punctuality that contrasts with the WiMAX Forum’s history of repeated delays and missed deadlines, and also highlights how timeliness is now paramount for both platforms. Operators now see the need to step up their own progress towards true mobile broadband in order to increase revenues and ARPU, and fend off new challengers.
LTE is now in the ‘change control procedure’ phase – with basic specs approved. Any further changes must be officially sanctioned, but the standard is now 80% complete and Release 8 should be frozen around year end. By that time some vendors will be ready with pre-standard test equipment for early mover carriers. Release 8 does not just focus on LTE in the RAN but is also expected to enhance HSPA+ (or HSPA Evolution), introduced in Release 7. By combining MIMO with a 64QAM modulation scheme, theoretical downlink speeds of up to 42Mbps can be reached in a 5MHz channel - similar to LTE using the same channel size – enabling HSPA operators to fully exploit their current technology before making the leap to LTE. This suggests that early adoption of LTE will be by non-3G operators, perhaps for fixed and converged usage, and for non-voice devices and applications. This brings the technology even more directly into conflict with WiMAX, which because of its OFDM head start is widely seen as being the natural choice for carriers with no W-CDMA legacy.
The LTE air interface, three years in the making, uses OFDMA (like Mobile WiMAX) in the downlink channel and SC-FDMA (Single Carrier - Frequency Division Multiple Access) in the uplink path. It takes data download rates to 100Mbps and upload rates to 50Mbps in 20MHz of spectrum while increasing network capacity and boosting performance at the cell edge. It also has a path towards 300Mbps using 4x4 MIMO.
Another element in the LTE family is System Architecture Evolution (SAE), which specifies the mobile core network and provides the migration path to all-IP. This should also be finalized by the end of 2008. In addition the LTE test specifications, which are necessary for handset manufacturers to test their devices with network elements, are "much further advanced" than normal, according to the 3GPP.
Verizon, Vodafone, and other major operators that committed to LTE as their primary next generation network are pressurizing vendors to have equipment ready for initial trials by the end of this year. Certified commercial systems will likely be available in 2010-11. This will come after the availability of Mobile WiMAX but before 802.16m, the successor to 802.16e (also known as WiMAX2, which some are dubbing with the snappier EC label WiMagic). There are multiple trials of pre-standard LTE systems now underway, with NTT DoCoMo the most advanced (working with Nokia Siemens and Panasonic). Most are taking place under the LTE/SAE Trial Initiative, in which companies such as Alcatel-Lucent, Ericsson, Orange, Nokia, NSN, Nortel, T-Mobile and Vodafone are involved.
On the WiMAX side there is renewed talk about 802.16m, which will aim to leapfrog LTE and reassure Mobile WiMAX adopters of the platform’s long term growth path. Among the parties aiming to contribute to the 802.16m air interface standard is a new initiative set up under the auspices of the European Commission’s Framework Programme 7 (FP7) for future mobile communications systems and IMT*Advanced (the successor to IMT2000, the international set of standards authorized for 3G).
The WiMagic project will incorporate seven technology partners including French WiMAX chip start-up Sequans, and six universities. WiMagic will work on a new air interface for WiMAX, which will be submitted to the IEEE 802.16m committee. WiMagic (Worldwide Interoperability for Microwave Broadband Access System for Next Generation Wireless Communications) has been accepted by the European Commission within the 7th Framework Programme for Research. The project will address physical layer functions such as MIMO, high performance synchronization and channel estimation techniques for high vehicular mobility, and MAC functions such as adaptive radio resource management. WiMagic will last for three years and sits alongside numerous FP7 projects related to next generation radio networks. Among these is MIMAX, which is looking to develop a low power, low cost MIMO platform based on a new way for combining and adaptive weighting of the antenna signals in the RF/LO front end. MIMAX also starts this month and will last for three years.
The IEEE has been careful to say that 802.16m would not necessarily equate to WiMAX - it would have to be adopted and approved by the Forum first. In theory, the technical deliberations over 802.16m should not be over-influenced by the commercial needs of the WiMAX industry, though in practice the two are virtually inseparable: WiMAX profiles are subsets of the 802.16 standards, but the Forum is the only major 'customer' for these IEEE standards. WiMAX Forum president Ron Resnick said in a recent interview that “802.16m is WiMAX and always will be." Although he would like WiMAX to work more closely with 3GPP bodies and with larger standards organizations like ETSI – especially since being approved as an IMT2000 platform last year - he does not see it converging fully with the UMTS/LTE roadmap.
One of the key enhancements 802.16m offers is greater scope in terms of channel bandwidths than 802.16e, with support for 20MHz and even 40MHz channels and a greater focus on FDD mode. Based on using enhanced OFDMA and antenna techniques (including 4x4 MIMO) in wider channels, it should be able to deliver theoretical downlink speeds of 350Mbps, with 200Mbps in the uplink, and mobility up to 350 kilometers per hour (802.16e boasts peaks of 128Mbps, 28Mbps and 120kmph respectively).
According to Wen Tong, a Nortel Fellow and holder of dozens of patents related to 3G networks, 802.16m and LTE will help address key challenges for next generation wireless networks: the need to support multiple devices per user, or what Nortel labels ‘hyperconnectivity’; and the need to use higher frequencies, which will require enhanced spectral efficiency as well as new base station formats to deliver ROI. Wen Tong told Infoweek that 802.16m will support twice the spectral efficiency of 16e, moving from current ratings of 1.2bph (bits per hertz) per second per sector, or 12Mbps in a 10MHz sector bandwidth, to 2.4bph per second per sector, or 24Mbps.
Other fundamental improvements will include enhancements in VoIP quality, mainly through multihop technology based on multiple battery- or AC-powered pico base stations filling gaps between the main stations with their backhaul. Multihopping capabilities will be integrated into .16m devices and will "unify the data rates across the nation without having to increase the grid”, delivering consistent performance regardless of distance from the macro station. Today 5MHz of spectrum can support 60-70 voice calls, but with multihop the same band can support three times that capacity, significantly increasing the quality of VoIP to match the levels delivered by circuit switched voice.