欢迎来逛逛,聊聊技术,侃侃大山. 本人主要喜欢WCDMA/TD-SCDMA无线侧的技术，如果有何关于无线技术方面的问题，你可以给我留言，我们可以一起讨论，共同提高。

Something gotten from HSDPA for future wireless

2006-08-17 11:03:34

(This is my writing in other BBS before, sharing with you, and welcome your discussion)
To be considered on HSDPA, we might think out something about future wireless communication:

1. RLC(Radio Link Control) protocol: as we know, the dobule link control layers had been used in 3G, TCP(NAS),RLC(AS). It's always a difficult issue to balance the control between these two layers. I think if the throughput of future wireless communication become higher, then RLC layer could be omitted or replaced by TCP. Because we can regard the wireless link as wire link with close performance.

2. Smart or intelligent transceiver: Node B or Transceiver will bear more functions than before. The dominator will be transferred from the current RNC to Node B (at least partly functions). The optimization job will become more difficult or simple. How to understand these? Because the major of important functions had been done by Node B, thus, maybe you need optimize system with one Node B by one Node B. Thus, this job become more complex. Also the throughput of HSDPA node B is higher, maybe we can use default parameters, and no or less optimization job is needed. such as, if the perfect Node B can support 10M, and if we install the Node B, and 9M is gotten, the optimiztion is needed? Maybe we don't needed optimize it to save human resources.

3. The distance between basestation and terminals will be closer more than before. Thus, the transmitting power will be decreased with smaller size of transceivers.

4. With more understanding on wireless channel and advanced automated optimization tools, the optimization mission will be less important in network deployment replaced by more accurate network planning design.

5. I remember when we install the 1G basestation, the field engineers need do lot of thing. With the advanced DSP and ASIC technology, the digitization will be closer to RF front end. Thus, the less job need to be done by field engineers. As like as the current 3G node B, if the antenna system is ready, you just need put the basestation on and half hour is enough to test the node B, simply! The value of field engineer will be down!

ok, stop here today!

BR,

James

论坛模式推荐收藏等级(1) 编辑管理 查看(54) 评论(1)

TAG:

xiao_er发布于2006-08-18 16:31:34: Some of my concern
First time to answer questions...flrru2giq
www.cnttr.comC
z$D@8D2v,w
1, For current HSDPA deployment, due to BLER and TCP Margin(ack/nack/...), there will be 20% difference on RLC and TCP throughput. These 2 layer will be combined in the fiture, or more accurately addressing, RLC will lose its function since most of the RLC functions are transfering to NodeB, such as mobility, scheduling, retransmission etc.

2, Flat system architecture is the target for tech evolution. Just like WiMax, there's no need for RNC or BTS Controller. Everything is handled by BTS, and charging and registration is for OAM system.

,Rv%rv+dr(G_www.cnttr.com3, No Comments. That's true, that's how wireless technology evolve.r|uf/BX2AT
&i)s(N{:?XJ
4, Can't agree with you. Cell planning and performance prediction is still in link simulation phase. For HSDPA, for example, simulator just calculation the C/I for each mobile and get the throughput result from mapping table. This way is so simple that hasn't take lots of RF thing into consideration. Slow fading, fast fading, scheduler, HARQ retransmission. Honestly, i'm not quite believe the result of simulation. Optimization will still be big problem for us.be1{']-^]"?R_

r
g.m,qv\5, Maybe. But RF engineer works by his experience, not by knowledge. So whatever technology it is, RF engineers are the most precious people in wireless.

Thanks.

进入论坛，查看全部评论