LTE4you: RLC data structure

RLC Data Structure

Like any data structure, you would not completely understand this structure until you really construct this structure or analyze the structure with real data either in programming or in manually. Of course my recommendation is to analyze manually.. I will put some of real RLC data example later, but for now let's just review the diagram from 3GPP specification and just try to be familiar with them.

The first structure is as follows. It is data structure for TMD (TM mode data) structure. What kind of structure you see from the following diagram ?

It is Data only structure, not even with Header. Most of the data structure for any data communication has it's own header. IP packet has IP header. UDP packet has UDP header. TCP header has TCP header. ICMP packet has ICMP header, etc. But this packet does not have any header. As you know, the most important role of header is to carry the information as follows :

i) Who is the sender ? (sender address, sender id etc)

ii) Who is the reciepient (reciepient address, id etc)

iii) What is the size of the data ?

iv) etc..

But this data structure has none of these information. It means that the TMD packet does not add any additional (header) to the input data and does not split or combine the data coming into the RLC entity.

Then how the size of this TMD PDU is determined ? It is automatically set to be the same size as MAC PDU size.

What if the incoming data is bigger than the RLC PDU size ? It just take the initial part of data that can fit into its size and discard the rest of the data.

What if the incoming data is smaller than the RLC PDU size ? It just take the whole data and pass it to MAC layer and MAC layer add padding data at the end.

Following is a data structure for UMD (UM mode data) structure. What are the meaning of Fl, E, SN ? You should refer to other sets of diagrams later. If you remember the diagram in previous section (Fig 4.2.1.2.1-1), RLC UM entity has capability of segmentation/concatenation. It means ... if the incoming data is greater than the RLC size, it segment the data into multiple chunks and pass them one by one. In this case, you have to put some tag (number) to each of the chunks.. otherwise the receiving side cannot combine them in proper sequence. This tag (number) is SN (Sequence Number). As you see in the following diagram, there are two different types of SN. One is 5 bits and the other one is 10 bits.

What if the data coming into the RLC entity ? The simplest way is to pack the small chunk as it is (as in TM mode), but it is waste of space.

Then what do we have to do in this case ? The simplest way is to put the multiple small chunk into a single RLC packet. In this case, you need special information for each of the small chunk within the RLC packet. LI (Length Indicator) is the length value for each of the small chunks within a RLC PDU.

The last type of RLC packet type is AMD (AM mode data) structure. Overall structure is very similar to UMD structure except that it has a couple of additional field D/C, RF, P.

We saw many of small fields in RLC header and now let's try understanding the meaning of each of these fields.

FI Field : This field indicate the relative location of the RLC PDU within a RLC SDU. For example, '01' indicate the PDU is the first segment of the SDU and '10' indicates the PDU is the last segment of the SDU, and '11' indicate the PDU is in between the first and last segment. In a SDU, there is only one '01' type PDU and only one '10' type PDU. There can be one or more '11' type PDU. If FI is 00, it means the RLC PDU is same as RLC SDU. When RLC recieves '01' type, the layer start buffering the PDU within RLC memory and only when it recieves '10' PDU, RLC reassembles all the recieved PDU and transfer it to PDCP.