Bandwidth Part (BWP) in SA and NSA

Home               LTE              NB-IoT            5G(NR-NSA)


To learn more about 5G, please click here.


In the last section we have already learn what actually BWP is and what the need to introduce BWP in NR was. In this section we are going to learn more about BWP and in coming parts we will learn types of BWP.

So to begin with we all know that the NR is different from LTE scheduling. Also, LTE scheduling happens in each TTI (one milli sec./or ever one sub subfarme), but this is not the case with NR scheduling. And the bandwidth is too big so new concept BWP is introduced.

Bandwidth Part (BWP) in SA and NSA:

As discussed in previous part BWP is defined as a combination of a diversity of RBs (resource blocks) that are consecutive within one carrier. The main thinking behind introducing the concept of BWP is mainly for the UE to better use the carrier bandwidth. 

Bandwidth part is a contiguous set of physical resource blocks with same subcarrier spacing and thus has the same numerology. 

Thus we can say that for a defined numerology each BWP can have three following parameters: 

Ø Cyclic prefix length (CP)
Ø Subcarrier spacing
Ø Symbol duration

       

Now, the important question - How many BWP can configure for UE?

For each serving cell, after the device is connected it can be configured with up-to four uplink bandwidth and up to four downlink bandwidth but only one BWP is active for uplink and one for downlink  at a given instant of time.


BWP for FDD and TDD: 

One UE can be configured with up to four DL BWPs and four UL BWPs in an NR FDD system. Whereas in NR TDD system, one UE can be configured with up to four BWP pairs. 

Here BWP pair means that the UL BWP ID and DL BWP ID are same, and the center frequencies of the DL BWP and UL BWP are the same but the bandwidth and the subcarrier spacing may not be consistent.

The bandwidth of the UE is consistent with the bandwidth of the system, and the decoded MIB information is configured to maintain the bandwidth as far as LTE is concerned.

Where as in NR the BW of the UE can be dynamically changed. The below figure is an example to explain BWP. At the initial moment, the UE has a large amount of traffic and the system allocates a large bandwidth say BWP1 to the UE. At the second moment, the UE ha small amount of traffic and thus system configures the UE with small bandwidth say BWP2 in order to satisfy the basic configuration (We will learn more about BWP switching in upcoming part). 

Now, the demand can be at the third moment where in the system finds that there is a large range of frequency selective fading in the BW of BWP1, or the resources in the frequency range of BWP1 are scarce or BWP1 may be assign to other UE, so a new bandwidth say BWP3 is configured for the UE by the network. 



So now the UE only needs to adopt the center frequency and sampling rate of the corresponding BWP in the corresponding BWP. Moreover, each BWP is not only different in frequency and BW but each BWP can correspond to a different configuration. For instance, the subcarrier spacing, CP type, and the SSB period of each BWP can be configured differently to adapt to different services.


Within one component carrier BWP is supported on uplink and downlink. The BW of the component can be divided into many bandwidth parts. For the network point of view different BWP can be associated with different numerologies (cyclic prefix, subcarrier spacing). 

The UEs with smaller BW support capability can work within a BWP with an associated numerology, By this it means UEs with different bandwidth support capability can work on large BW component carrier.

NR supports UE bandwidth par part adaption for numerology switching and UE power saving. The network can operate on a wide BW carrier but can work over activated bandwidth parts thus optimizing the use of radio resources to the traffic demand and minimizing interference from other systems.



How network scheduler can schedule multiple use case and multiple UE at same time using different BWP?

Let’s take below example in which three different UE have large amount data to transmit and at the same time MTC and URLCC devices are also there then, just for example network Scheduler will schedule like below:
Note: In real time scenario there could be number of UE’s( GBR as well as no GBR traffic could be there) and number of MTC as well as URLCC devices scheduled at the same time. For below figure we took only three UE and one MTC as well as URLCC device example.




BWP is divided mainly in two following categories –

1) Initial BWP – this is mainly used for the UE to receive the RMSI, the OSI to initiate random access and the like.

2) Dedicated BWP – this is mainly used for data service transmission.
Note: The bandwidth of dedicated BWP is generally larger than the initial BWP.

The technical advantages of BWP have following four main aspects –

The UE does not need to support the entire bandwidth and only needs to meet the minimum bandwidth requirement. This is beneficial to the development of low coat terminals and promotes industrial development.

When UE traffic is not large enough, the UE can switch to low bandwidth operation. This helps significantly in reducing the power consumption.

5G is a forward compatible technology. So when 5G adds new technology it can directly run new technology on the new BWP. Thus ensuring the forward compatibility of the system.

Another technical advantage is that it adapts to business need and dynamically configures BWP for business.

In next document we will learn more about BWP types and BWP switching. …….. so stay tune and happy reading.


---Pinal Dobariya.






Reference:
3gpp 38.213
3gpp 38.214 
3gpp 38.321
3gpp 38.331
https://arxiv.org/ftp/arxiv/papers/1712/1712.09724.pdf

Comments

  1. A question regarding the usage of BWP. In the example above, there are more than 5 BWPs in this cell. Does this mean that a UE can be configured at most 4 BWPs, and there is no restriction on the number of BWPs of a cell?

    ReplyDelete
    Replies
    1. However, the BWP-Id ::= INTEGER (0..maxNrofBWPs), in which maxNrofBWPs = 4.

      Delete
    2. I think this maxNrofBWPs limit is per UE.

      Delete

Post a Comment

Popular posts from this blog

5G Deployment Option-3/3a/3x

PRACH Basic-3

Control Plane CIoT EPS optimization

cell selection part-2

DRX (Discontinuous Reception)

Soft Handover vs Hard Handover

DC Carrier in NR