Presentation on theme: "Fraunhofer Heinrich Hertz Institute Fraunhofer Heinrich Hertz Institute, Einsteinufer 37, 10587 Berlin www.hhi.fraunhofer.de Backhaul requirements for."— Presentation transcript:
Fraunhofer Heinrich Hertz Institute Fraunhofer Heinrich Hertz Institute, Einsteinufer 37, 10587 Berlin www.hhi.fraunhofer.de Backhaul requirements for CoMP V. Jungnickel, L. Thiele, K. Börner Fraunhofer Heinrich-Hertz-Institut
Broadband Mobile Communications Systems and Networks Implications on the network layer CoMP can be handled only for clusters of cells Dynamic clustering is most promising Down-tilt reduces the cluster size Most relevant signals come from same site Some relevant signals come from other sites Impact of the down-tilt onto clustering is yet an open issue Initial working assumptions Let us select a subset of cells out of fixed, overlapping clusters Serving cell is in the middle. Fixed clusters size is 7. Ratio between intra- and inter-site cooperation is assumed as ½. Backhaul traffic is created by inter-site cooperation only.
Broadband Mobile Communications Systems and Networks Estimation of the backhaul Assumptions: LTE-Advanced network, Operator = DTAG, 3 sectors per site Spectrum aggregation on 800 MHz, 1.8 GHz and 2.6 GHz bands 50 MHz aggregated bandwidth, CoMP is used for the whole spectrum Downlink: 8 antennas 24x64-QAM = maximal 7.2 Gbit/s Realistic data rate is much less due to interference Data rate depends on cluster size Most recent simulation results of HHI R. Irmer et al., February 2011 in IEEE Communications Magazine
Broadband Mobile Communications Systems and Networks Downlink results Assumptions: Results scale almost linearly with numbers of antennas per sector Sum rate results are computed for 2x2 MIMO Only the inter-site CoMP backhaul part is counted Intra-site CoMP charges the backplane at each site
Broadband Mobile Communications Systems and Networks Requirements for LTE Rel. 11/12 o Expected mobile backhaul requirements develop quickly (downlink) 2011 initial LTE Rollout (Rel. 8): 225 Mb/s (S1) 10 MHz, 2x2 MIMO @ 800 MHz: 2014 1 st capacity extension (Rel. 10): 2.1 Gb/s (S1) 50 MHz, 8x8 MIMO @ 0.8, 1.8, 2.6 GHz 2017 2 nd capacity extension (Rel. 11/12): 9 Gb/s (S1+X2) 50 MHz, 8x8 MIMO+CoMP with 5 BSs @ 0.8, 1.8, 2.6 GHz o CONDOR is finished in 2013, and may target Rel. 11/12 requirements Uplink requirements are ~ 2…3 times downlink requirements But uplink traffic is not as critical, compared to downlink traffic CoMP backhaul designed for 10 (DL only) or 40 Gb/s (DL+UL )
Broadband Mobile Communications Systems and Networks Conclusions for demo o Demo should exemplify key technologies for high throughput (10-40 Gb/s per base station site) low latency (<1 ms) o Design should be scalable w.r.t. bandwidth 2014: 2 Gb/s, 2017: 10 Gb/s, 2020: 40 Gb/s o Reconfigurable bandwidth More resources if network throughput is increased Semi-static allocation of optical spectrum, e.g. same wavelength Flexible signal processing (e.g. by additional OFDM subcarriers)
Broadband Mobile Communications Systems and Networks For discussion: A poor man‘s view o Basic idea: each BS has it/s own wavelength o CWDM is used for integration into metro-access network o DWDM is used locally inside CWDM bands Wavelength-controlled lasers like FlexOptics in XFP module No temperature control (low cost) 50/100 GHz spacing Tx+Rx in XFP = coherent receiver o OFDM signal processing with variable number of sub-carriers E.g. 16 subcarriers initially, then 64, then 256 Scalable sampling clock upgrade processing cards if demands grow