100 likes | 220 Views
Towards IEEE802.11 HDR in the Enterprise. Sebastien Simoens, Amitava Ghosh, Alistair Buttar, Karine Gosse, Ken Stewart Motorola Contact: simoens@motorola.com alistair.buttar@motorola.com. Future Enterprise. Enterprise data needs will continue to grow
E N D
Towards IEEE802.11 HDRin the Enterprise Sebastien Simoens, Amitava Ghosh, Alistair Buttar, Karine Gosse, Ken Stewart Motorola Contact: simoens@motorola.com alistair.buttar@motorola.com
Future Enterprise • Enterprise data needs will continue to grow • Ethernet LANs will use 1Gb/s and 10Gb/s technologies in 2005 • Deployment of WLAN APs will complement wireline and cellular infrastructure • Peak data rate of >100Mb/s is attractive to IT departments (alternative to 100Base-T) • Higher peak data rates and higher average throughput result in productivity enhancements in the enterprise • Market trends support HDR standardization effort
HDR Requirements • Extend 802.11a in the 5GHz band • Peak data rate >100Mb/s, and increased cell throughput • Broad regulatory support • Graceful degradation as the system load increases • Backwards compatible with 802.11 • Support for limited mobility • Optimized PHY/MAC solution • More efficient MAC enables to take advantage of higher rate PHY (see « WNG requirements for Home networking »)
Technology Considerations • Advanced channel coder, e.g. Turbo Codes or Turbo Trellis Coded Modulation • OFDM but with higher order QAM (e.g. 256) and higher number of sub-carriers • Hybrid ARQ • Spatial Diversity Techniques e.g. 2x2 MIMO • Link Adaptation (already supported) • Reduced Overhead / Improved Signalling • Efficient QoS support • Improved performance can also be traded-off for lower power consumption, or increased robustness
Improved Channel Coder • Use low constraint length Turbo Code eg. (K=3) R=1/3 instead of current K=7 R=1/2 convolutional code • Limit the number of iterations to 4 to reduce complexity at the terminal • 1-3dB gain can be achieved with Turbo codes compared to convolutional code
Hybrid ARQ • Self-optimizing and adjusts automatically to channel conditions • Adds redundancy only when needed • Receiver saves failed transmission attempts to help future decoding • Every transmission helps to increase the packet success probability • HARQ Scheme should support both Incremental Redundancy (IR) and Chase Combining • Enabled by N-Channel Stop-and-Wait ARQ • AT-assisted Channel Quality Indication (CQI) • Channel reciprocity needs to be considered here, although potential differences in UL/DL interference levels. • Key Challenges • Can be best accommodated using optimized MAC • Memory cost at the terminal needs to be optimized
2x2 Spatial Diversity (1/3) • 2 Transmit, 2 Receive antenna system • Already addressed in other standards (802.16.3, 3GPP) • Improves peak data rate and/or average data throughput • Reasonable additional complexity • Many alternatives to exploit the 2x2 configuration: • link budget/cell throughput improvement relying on spatial diversity: Tx space-time block coding, Rx diversity • bit-rate increase with parallel data flows: Multiple Input Multiple Output techniques (MIMO) can be considered • « Beamforming » with channel knowledge at Tx: most gain obtained but requires feedback channel or TDD channel reciprocity assumption • Need for comparisons in terms of overall cell throughput • Need for defining the level of acceptable complexity at AP and MT
2x2 Spatial diversity (2/3) QPSK, 16QAM rate ½ code Packets: 2x54 bytes Space-Time Block Coding on .11a PHY, office NLOS channel 2x2 8.7dB 1x1 2x2 shows a significant improvement of 8.7dB at PER=10-2
2x2 Spatial diversity (3/3) Based on previous slide PHY results, with path loss (n=3.8) and shadowing 10.5Mb/s A significant cell range extension can be obtained 18Mb/s 29m
Conclusion • Goal of HDR is to improve peak as well as the average throughput • Technology elements exist for efficient HDR WLAN • Standardization activity should be started before proprietary products fragment the marketplace • HDR Study Group should be created in 802.11