Thursday, July 05, 2012

Some questions and answers on LTE Radio Interface - Part2


1. Which protocol is responsible for Scheduling of user data and HARQ?


MAC


A Medium Access Control (MAC) Hybrid Automatic Repeat reQuest (HARQ) layer with fast feedback provides a means for quickly correcting most errors from the radio channel. To achieve low delay and efficient use of radio resources, the HARQ operates with a native error rate which is sufficient only for services with moderate error rate requirements such as for instance VoIP. Lower error rates are achieved by letting an
outer Automatic Repeat reQuest (ARQ) layer in the eNB handle the HARQ errors.


Name:  LTE-MME.jpg
Views: 0
Size:  68.7 KB


2. Which protocol is responsible for ciphering of user data?


PDCP


The[B] PDCP protocol maps the EPS bearer onto the E-UTRA Radio Bearer and performs [B]Robust Header Compression (ROHC).NAS messages are protected using the ciphering and integrity protection services provided by the PDCP layer.


The Packet Data Convergence Protocol supports the following functions:


• Header compression and decompression of IP data flows using the ROHC (Robust Header Compression) protocol, at the transmitting and receiving entity, respectively.
• transfer of data (user plane or control plane). This function is used for conveyance of data between users of PDCP services.
• maintenance of PDCP sequence numbers for radio bearers for radio bearers mapped on RLC acknowledged mode.
• in-sequence delivery of upper layer PDUs at Handover 
• duplicate elimination of lower layer SDUs at Handover for radio bearers mapped on RLC acknowledged mode
•[B] ciphering and deciphering of user plane data and control plane data[/B]
• integrity protection of control plane data
• timer based discard


3. How does the frequency domain structure differ in UL compared to DL?


In UL the frequency allocation must be continuous in order to preserve the single carrier properties. This is not the case in DL, where non-contiguous resource blocks  be be allocated to the same user.


4. How much can the data rate be increased with 2x2 MIMO compared to a solution without MIMO?


Up to two times


With MIMO,  multiple antennas and advanced signal processing such as spatial multiplexing, the radio channel can be separated into several
layers, or “data pipes”. Up to four layers can be utilized. This corresponds to up to four times higher data rates for a given bandwidth.


5. Explain the concepts of channel rank, layers, data rate multiplication and codebook.


The radio channel properties decide the maximum channel rank that can be used, i.e. how many layers the channel support at the moment. The number of layers that can be transmitted over the radio channel is equal to the data rate multiplication (e.g. two layers give two times the data rate compared to a solution without MIMO). The complex weights that are applied at each antenna port are selected from a finite
codebook. The codebook index is suggested and indicated by the UE.


6. How HARQ works?


Multiple simple stop-and-wait ARQ processes are processed by the HARQ entity in the MAC protocol. The operation is very fast and has a short round-trip-time thanks to the short TTI and the fact that it is located in the eNodeB, close to the radio interface. Feedback from the receiver is sent in terms of short ACK/NACK messages.


7. How to calculate the maximum theoretical physical peak data rate in LTE radio interface?


Each OFDM symbol contains, if 64-QAM is used, 6 bits per subcarrier (15kHz).


There are, if normal CP is used, 7 OFDM symbols
per slot. This ends up with 6*7 = 42 bits per slot. One slot is 0.5 ms which gives us 42/0.5ms = 84kbps per sub-carrier.


If the full bandwidth, 20MHz, is used, there are 20MHz/15kHz=1333 sub-carriers.


However, only 1200 of these are used for user data. This corresponds to 100 resource blocks.
1200*84kbps = 100,8 Mbps.


With four MIMO layers, we should be able to achieve 403.2 Mbps of raw data rate in the physical layer.


What about the user data rate? The data rates used for L1/L2 signaling, reference signals, PBCH, SCH, layer 3 signaling and
protocol headers has to be subtracted from this figure. Then we end up with approximately 320 Mbps of user data rate on RLC
level??


In UL we have approximately the same calculation, except that the gain from MIMO cannot be included, since no SU-MIMO is used in
UL. Hence, approximately 80-100 Mbps of theoretical bitrate should be possible to reach.

No comments: