Abstract
IAB’s architecture gives 5G and beyond networks the ability to manage resources both centralised and decentralised: an IAB-donor, a centralised node, can efficiently coordinate resources with a network-wide view, and IAB-nodes, distributed nodes, can quickly respond to local traffic bursts multiple hops away from the IAB-donor. However, there aren’t many research on handling bursty traffic in the present IAB networking initiatives, which largely concentrate on the centralised or distributed capabilities of IAB networks. In order to manage both bursty and non-bursty traffic, a Coordinated Parallel Resource allocation technique (CPReal) is created in this study. Its main strength is that it combines a distributed system at each IAB-node for rapid response to short-term bursty traffic with a centralised scheme at the IAB-donor for effective coordination of resource allocation. The rigorous parent-to-child link between IAB-nodes defined in 5G standards is taken into account by CPReal, which complies with the specifications. IAB-nodes in CPReal may immediately agree on resource allocation in a distributed and conflict-free manner for traffic surges with the support of their parent nodes. CPReal surpasses conventional systems in end-to-end latency by up to 34.4% under different situations and in throughput by 67.3% under bursty traffic, according to extensive simulations.backhaul
Introduction
For the efficient and dense deployment of base stations (BSs) in 5G communications, integrated access and backhaul (IAB) networking is essential . To accomplish unified management of all connections and boost spectrum efficiency, backhaul and access links in IAB networks are formed via a single unified air interface. IAB-nodes and the UEs they are connected to make up an IAB network, as seen in Figure 1. A unique IAB-node, known as an IAB-donor, among all IAB-nodes, manages the whole IAB network using a central unit (CU) function and has fibre access to the core network. There is a tight parent-to-child relationship between two directly linked IAB-nodes, where the parent node administers the child node, as stated in the 5G specifications, and In this study, two IAB-nodes that are directly linked are referred to as linked IAB-nodes, whereas unlinked IAB-nodes are two IAB-nodes that do not share the same IAB link. All IAB-nodes have one or more parent nodes aside from the IAB-donor. A parent node utilises a distributed unit (DU) function to handle radio resources for the backhaul links between the parent node and its associated child nodes, while a child node connects to its parent nodes via a mobile terminal (MT) function. A parent node’s DU function can be used to create access connections with UEs in addition to controlling child IAB-nodes. The uplink traffic from each UE to the IAB-donor always comes after the sequential uplink transmissions from an MT or UE to a DU, as specified by the 5G specifications for IAB networking . Similar to uplink traffic, downlink traffic follows successive transmissions from a DU to an MT or a UE. A directed acyclic graph (DAG) architecture for routing in IAB networks is the consequence of such a design.backhaul
An IAB network is by its very nature dynamic, with bursts of traffic. Each IAB-resource node’s allocation has to be modified in order to respond to such dynamic occurrences. Since it takes time for the centralised node, or IAB-donor, to gather link and traffic data from IAB-nodes and distribute a centralised scheduling result over the entire network, a centralised scheme is ineffective for handling such situations. While a distributed system enables swift local decision-making by each IAB-node, it is also prone to disputes in the distribution of resources among IAB-nodes. An IAB-node must communicate with its nearby IAB-nodes to settle such disputes, and each of them may then communicate with other neighbours. The distributed approach is poor at resolving disagreements over resource allocation because of the sluggish chain-effect mechanism. However, the IAB-centralized donor’s coordination offers a lot of potential to speed up the conflict resolution procedure. In order to swiftly assign resources without creating resource conflicts, a coordinated distributed scheduling system between the IAB and donors is essential.backhaul
Coincidentally, resource scheduling in an IAB network also adheres to an IAB-donor coordinated mechanism, as per 3GPP standards. IAB-nodes are given time slots by the IAB-donor, and for each slot, an IAB-node chooses the specific resource blocks (RBs) for its child IAB-nodes and connected UEs. The resources that can be assigned in a slot are classified as either hard or soft depending on the various traffic types handled by the IAB network. An IAB-node can directly use all of the RBs in a slot’s designated hard resources if it wants to communicate with its child nodes and UEs. The IAB-donor must make sure that no other interfering IAB-nodes are given this time slot in order to avoid potential resource allocation conflicts. This specific slot distribution, meanwhile, is only effective in situations when traffic volume is stable. An IAB-node with bursty traffic cannot have a dedicated slot given to it. Instead, it must use a slot with soft resources, or one whose resources are shared with other interfering IAB nodes. Therefore, potential conflicts with another IAB-node may arise when an IAB-node allocates RBs in such a slot for its child IAB-nodes and UEs. All IAB-nodes assigned to the same slot participate in a sequential process of RB allocation in order to reduce such conflicts. At the top of the IAB topological tree, there is an IAB-node that serves as the starting point. When an IAB-node has finished allocating resources, the child IAB-nodes that were given the same slot continue to do so, however the child IAB-nodes are not permitted to utilise the assigned RBs. The IAB topological tree’s bottom IAB-nodes are reached by repeating this sequential operation.
Although soft resource allocation increases resource utilisation for bursting traffic, there are a number of issues. First off, in a network with several transmission hops, the sequential method causes significant resource allocation delay, especially for leaf nodes. Second, resource conflicts between two connected IAB-nodes can only be avoided by allocating resources sequentially; conflicts between two unlinked IAB-nodes continue to occur. Third, only the resources left in the soft slot are accessible to child nodes by each parent node. Resource usage is poor since there is no resource reuse.backhaul
This work introduces CPReal, an IAB-donor coordinated parallel resource allocation technique that rigorously adheres to the parent-child relationship and the specifications for soft and hard resources as described in standards, with the goal of removing the constraints in the current standards. In CPReal, the IAB-donor manages a centralised coordination scheme to coordinate each IAB-node to optimise the effectiveness of network-wide resource allocation, while each IAB-node manages a dispersed scheme in reaction to network dynamics quickly. Both physical resources and soft resources are taken into account in CPReal in order to accommodate various traffic kinds. The IAB-donor is in charge of the former during each centralised scheduling period (centralised SP), whilst the IAB-nodes are in charge of the latter during each distributed scheduling period (distributed SP). In each centralised scheduling period, network status and user needs are gathered at the start of the last frame so that there is enough time to compute and communicate scheduling choices before the following scheduling period. For guaranteed bit rate (GBR) traffic flows, the IAB-donor allots hard resources. In order to ensure effective transmission of control messages in distributed systems, it also allocates hard resources. Soft resources are entirely allocated via the distributed method in order to enable bursty traffic, hence the IAB-donor is not required to pre-allocate soft resources for IAB-nodes. In both centralised and distributed approaches, the hard and soft resources are assigned at the RB level of detail.backhaul
In order to accomplish conflict-free resource allocation for bursty traffic, CPReal is made up of two essential methods, as shown in Fig. 2. Each IAB-node seeks RBs for its local communication connections if resource allocation has to be changed as a result of network dynamics. Resource allocation decisions made by several IAB-nodes might clash. CPReal has two essential processes to seek resources and settle disputes so that it may respond to bursty traffic fast without triggering conflicts over resources. The IAB-donor manages the first mechanism. More exactly, each IAB-preferred node’s resources are chosen by the IAB-donor. Following a conflict-mitigation parallel resource allocation (CMPA) process, each IAB-node selects its necessary RBs in soft resources based on these preferences, with the result that various IAB-nodes typically choose different resources. As a result, the likelihood of resource disputes is decreased by the first mechanism. Each IAB-node delivers the results of resource selection to its parent nodes, which identify and resolve any possible conflicts of resource allocation if they still arise. The signalling broadcasts, as seen in Fig. 2, utilise the hard resources designated for control messages. There is no controversy as the IAB-donor has already decided how much resources would be allocated for signalling. When a dispute over resource allocation arises after receiving requests from child IAB-nodes, the second scheme, a parallel conflict resolution (PCR) method, is used to settle it. A conflict-pair table is used by a parent IAB-node to identify interfering child IAB-nodes in order to carry out proper conflicts detection. The network topology and the interference between IAB nodes are used by the IAB donor to construct this table, which contains all pairs of IAB nodes (both connected and unlinked) that may cause resource conflicts to one another. When a common parent IAB-node of several child IAB-nodes notices resource allocation conflicts, it can help the child IAB-nodes deallocate the resources that are in dispute. In order to guarantee proportionate justice in the distribution of resources among these child IAB-nodes, a fair allocation method is created for the parent IAB-node to identify the suitable child IAB-nodes where conflict resources are dealt with. Each distributed SP’s hard resources have completed their tasks of resource allocation and conflict resolution, and as a result, each IAB-node is free to transmit bursty traffic in the soft resources without encountering any problems.backhaul
In this study, we concentrate on developing effective plans for sub-6 GHz IAB networks, which have benefits over their millimeter-wave (mmWave) counterparts in terms of wide coverage, handling obstructions, and mobility. IAB can function in both millimeter- and sub-6 GHz frequency bands in access and backhaul networks, per 3GPP specifications . In addition to improving service coverage for rural regions, sub-6 GHz IAB also offers interim wide coverage and dependable connection in emergency situations. In spite of this, as explained in Section 5.2 and Appendix E, our techniques may also be used to mmWave IAB networks. Extensive simulations are used to assess CPReal’s performance. When there is bursty traffic, CPReal outperforms other scheduling strategies by 67.3% in terms of throughput. As a result, CPReal better adjusts to network dynamics. Additionally, CPReal considerably decreases the average end-to-end latency under diverse conditions by up to 34.4%.
The rest of the document is structured as follows. A survey of the relevant work is provided in Section 2. In Section 3, the CPReal system model is presented. Section 4 presents the dispersed scheme’s main mechanisms, while Section 5 describes the centralised coordination method. Performance data is presented in Section 6, and Section 7 brings the work to a close.backhaul
Fragments of sections
Similar work
A current area of study is resource scheduling and allocation in multi-hop networks. We concentrate our research efforts on resource allocation in IAB networks, particularly when dealing with bursty traffic.
Expected Traffic Load for IAB Networks Resource allocation and scheduling issues in multi-hop networks can be well-formulated to optimise desired performance metrics given a fixed amount of traffic load. Path selection, power allocation, and system model are influencing elements that must be considered in order to enhance the network throughput.
The general structure of CPReal is seen in Fig. 2, where temporal resources are split up into periodic centralised SPs, each of which is made up of a number of fixed-length distributed SPs. The smallest unit for scheduling data messages is the resource block (RB), which has a time slot and 12 subcarriers in the frequency dimension. A given number of OFDM symbols can be further split into each slot. A resource unit (RU),
Distributed resource allocation, and dispute resolution, is the smallest unit for scheduling control messages.
This section presents the distributed scheme’s specifics. When an IAB-node has to use soft resources to seek data RBs to manage bursty traffic, it first chooses the resources it prefers before sending the local choice to its parents for dispute resolution. The conflict-mitigation parallel resource allocation (CMPA) mechanism is intended to lower the probability that IAB-nodes will request the same data RBs in soft resources, i.e., lower the conflict probability. The simultaneous conflict
Coordinated centrally
As was previously mentioned, to increase the effectiveness of resource allocation across the entire network, the IAB-donor coordinates the distributed schemes at the IAB-nodes.
simulated environment
This section assesses the effectiveness of CPReal in various network scenarios. To assess the effectiveness of IAB networks, event-driven simulations are carried out based on prior research , and and a typical topology utilised in 3GPP standards. According to Fig. 8, the separation between two connected IAB-nodes is
The parent-to-child relationship between the IAB-nodes is shown between the arrows in the figure at m. Each IAB-coverage node’s area is 100 equally distributed points with a radius of 50 metres.
Conclusion
For IAB networks to quickly respond to network dynamics and effectively coordinate resource distribution throughout the whole network at the same time, a parallel resource allocation strategy managed by an IAB-donor was devised in this research. Both a proactive strategy and a reactive method were established in the distributed scheme with the help of the IAB-donor to ensure conflict-free resource allocation for bursty traffic. The proactive approach can increase the likelihood that IAB-nodes will select the Declaration of Competing Interest.backhaul
