Skip to main content

Integrated Access and Backhaul (IAB) Resource Allocation Methodologies

We have looked at Integrated Access and Backhaul, a.k.a. IAB in couple of posts earlier. The 3G4G Blog post late last year talked about what IAB is, what features will be available this year as part of Release-16 and what enhancements are planned as part of Release-17. We also looked at IAB from Ericsson Tech Review on the post here.
5G Americas released a whitepaper on Innovations in 5G Backhaul Technologies last month. There are quite a few interesting things in there that I am going to cover on this blog in this and the next few posts. There is a nice detailed explanation of IAB that complements my earlier posts very well. It should be a good starting point for anyone willing to look at IAB in detail.

One of the topic in the paper that caught my attention was IAB Resource Allocation Methodologies. If we think simplistically, then both the donor and the relay node can transmit simultaneously when transmitting in different frequencies (out-of-band). This can work really well when the backhaul part is using mmWave while the access part is using sub-7 GHz. If both the backhaul and access has to use the same frequency (in-band) then some kind of resource allocation methodology becomes important.

The following is from the whitepaper:

IAB Resource Allocation Methodologies

IAB networks will require a different approach to how 5G networks allocate resources. The following methodologies illustrate options available in developing IAB architectures.

Radio Resource sharing between access and backhaul

To mitigate the cross-link interferences for in-band backhaul, different half-duplex multiplexing schemes have been designed for IAB network, such as TDM (Time Division Multiplexing), FDM (Frequency Division Multiplexing) and SDM (Spatial Division Multiplexing).

In the case of out-of-band relaying, sub-6 GHz can be considered as an access and control channel for backhaul links due to its robustness against obstacles and wide coverage area. mmWave bands can be used for high capacity backhaul links. In this case, the IAB network can operate in full-duplex mode.

Interference can occur on both access and backhaul links for the out-of-band case. In addition, cross-link interference (between access and backhaul) can occur for the in-band case. Interference management techniques, which use the channel state information, are required in order to suppress the interference among concurrent transmissions, both in access and backhaul. For the IAB node coordination, efficient signaling exchange among the MAC layer of different IAB nodes is needed, considering the rate and latency constraints of wireless backhaul links.

Uplink-downlink interference is also introduced in case of asynchronous IAB node transmission mode. Adaptive intelligence algorithms can be implemented at the MAC layer of IAB and macro nodes to make adaptive decisions about the link and user/IAB child scheduling with fairness and half-duplex constraints. Training procedures can be centralized, for example at the donor nodes or distributed among some local IAB nodes.

Time Division Multiplexing (TDM)

In case of TDM, access and backhaul links operate at the same carrier frequency but at the different time frames. For downlink traffic, the IAB node will have to receive the signal from the parent node first before it can relay the signal further to the child node or the UEs. For uplink traffic, the IAB node will have to receive the signal from the child node or the UE first before it can relay the signal further to the parent node. In a TDM system it is necessary to divide and allocate time resources according to the resource situation of each link. Figure above shows one example where radio resources are allocated evenly for the child and parent links in the time domain. Note that a grey box represents a downlink (DL) slot, and a blue box represents an uplink (UL) slot.
To improve the efficiency of radio resource utilization, dynamic TDM can be used. The TDM slot number and location for parent backhaul link can be flexibly configured according to the backhaul transmission capacity requirement. For each specific frame configuration, the backhaul slot can be used for access link by dynamic scheduling if the backhaul transmission is not scheduled in that slot, as shown in Figure above. Simulation results show that dynamic TDM scheme brings significant performance improvement compared to static TDM, especially when the network resource utilization is low or medium. In the evaluations, the gain for DL 50% percentile user throughput was found to be more than doubled by using dynamic TDM compared to static TDM in case of 50% resource utilization.

Frequency Division Multiplexing (FDM)

If operators own enough frequency spectrum, FDM can be used to eliminate cross-link interference. Different carrier frequencies are allocated to parent backhaul, child backhaul and access links. With enough guard band between parent and child backhaul links, all the IAB nodes can transmit and receive simultaneously without introducing too much cross-link interference. 

To this end, it is necessary to divide and allocate the frequency resources independently between child and parent links. Figure 21 an the top below shows an example where the child and parent links are allocated the same amount of frequency resources.

Spatial Division Multiplexing (SDM)

In addition to multiplexing schemes in time and frequency domain, spatial division multiplexing can be used to separate the backhaul and access links. Beamforming algorithms using multiple antennas can be applied on IAB nodes to separate the backhaul and access links in space.

In an SDM solution, the child and parent links exploit the spatial separation between child and parent links to minimize interference. In this case, simultaneous transmission (or reception) at the IAB-node are allowed in the same time and frequency resources, as long as the spatial separation is enough to minimize the interference between the simultaneous transmission (or reception). This will greatly reduce end-to-end transmission delay.

Figure 22 at the top shows an example where radio resources are divided into child and parent links in the space domain.

In this case, simultaneous transmission (or reception) in both child and parent links are allowed. However, even though this solution allows full exploitation of time and frequency resources, the beams are usually not narrow enough to prevent the cross-link interference if all the IAB nodes transmit and receive at the same time. As such, TDM or FDM will have to be applied together with SDM to achieve the required signal-to-interference-plus-noise ratio (SINR) for each link.

The whitepaper has a lot more details so feel free to check it out here.

Related Posts:

Comments

Popular posts from this blog

High-level Architecture Introduction of Mobile Cellular Networks from 2G to 5G

Here is an old tutorial explaining high level mobile network architecture, starting from GSM and then looking at GPRS, UMTS, LTE & 5G. Slides and video below High-level architecture of Mobile Cellular Networks from 2G to 5G from 3G4G Related links : Free 2G, 3G, 4G & 5G Training Videos 5G (IMT-2020) Wireless 5G vs 4G: what is the difference?

IEEE 802.11be Extremely High Throughput (EHT), a.k.a. Wi-Fi 7

We have been writing about Wi-Fi for a long time, weather it's to do with indoor connectivity , competition with 5G or just a name change to something simpler. When we last wrote about WiFi 6, a.k.a. 802.11ax, we were expecting a quick adoption of the technology in the industry. We are still not there yet.  You know what's strange? None of the new @madebygoogle gadgets from yesterday support Wi-Fi 6. Not the Pixel 5, not the Pixel 4a 5G, not the Nest Audio, and not the new Chromecast. pic.twitter.com/QtJ8iB9FeO — Ry Crist (@rycrist) October 1, 2020 Take for instance the new iPhone 12 supports Wi-Fi 6 in all their models as one would expect but none of the new Google Pixel phones (4a, 4a 5G and 5) support it. In fact none of the new Google devices support it. Which is rather bizarre. While we are still looking forward to Wi-Fi 6 becoming widespread, IEEE has already started working on the successor of 802.11ax, 802.11be - Standard for Information technology--Telecommunicati

CSI-RS vs SRS Beamforming

In an issue of Signals Flash by Signals Research Group (SRG), they talked about 2 different types of MIMO. Quoting from their journal, "CSI-RS versus SRS. Those operators that have tested or made token use of MU-MIMO leverage a flavor of MU-MIMO that is based on CSI-RS. The MU-MIMO network we tested was based on SRS, which makes it far more likely to observe sixteen spatial layers (versus eight)." I reached out to Emil Björnson, Visiting Professor at KTH Royal Institute of Technology and Associate Professor at Linköping University to see if he has explained this in any of his videos. Here is what he said: " I'm not talking about 3GPP terminology in any of my videos. But you can listen to the slides that starts around 12:40 in this video (embedded below) . If you are looking for CSI-RS vs SRS based MU-MIMO, then jump to around 12:40 in this video where you can see CSI-RS being referred to as "grid of beams" and SRS is similar to the other option, which is t