Abstract
Next-generation communication systems demand uninterrupted mobility across heterogeneous access networks, where users regularly switch among technologies (e.g., Wi-Fi, 5G, and future 6G networks). Network-Based Distributed Mobility Management (NB-DMM) scales and routes better than centralized mobility solutions, but vertical handovers across heterogeneous access domains may still cause handover latency, packet loss, signaling overhead, and out-of-order packet delivery. The Dual Buffer Signaling Aggregation (DB-SAHO) was an earlier mechanism proposed to improve handover performance in NB-DMM homogeneous environments. The paper further extends the DBSA concept to heterogeneous networks and proposes a heterogeneous H-DBSA scheme that facilitates vertical handovers across access technologies. The proposed method uses a dual buffer service and aggregates signaling across diverse access areas. During transition, a forward buffer preserves incoming messages from the previous access network while the new buffer ensures a continuous, ordered transfer of data from the target network. At the same time, signaling aggregation reduces the number of control messages transmitted during handover. The proposed scheme is tested for performance using OMNeT++ simulations at different mobility speeds, and for sensitivity analysis with different buffer sizes and signaling delays. Simulation results show that the heterogeneous H-DBSA scheme significantly decreases latency during handover, minimizes packet losses and out-of-order rearrangements, and enhances throughput more than it increases signaling costs compared to the existing NB-DMM. The sensitivity results also reveal that performance is stable throughout various operating scenarios, showing the capability of the proposed mechanism in heterogeneous distributed mobility contexts.
1 Introduction
The number of mobile users and the use of smart devices are rapidly increasing, creating a need for constant connectivity, and excellent services. Image-based streaming (video), voice communication, autonomous systems, and interactive cloud services are real-time applications that are highly responsive to handover effects, such as delays, packet losses, and packet reordering (Saoud et al., 2025). As a result, providing smooth mobility has become essential for the development of the upcoming generation wireless networks. Network-Based Distributed Mobility Management (NB-DMM) has been proposed to bypass the hindrances of centralized mobility solutions in terms of scalability and routing efficiency by locating mobility anchors nearer to the access network (Balfaqih et al., 2016). Although NB-DMM has architectural benefits, it has been found to suffer performance loss whenever a handover event occurs. Packet loss is also common because of latency in the forwarding of packets and failure of mobility anchors to coordinate properly (Mane and Khairnar, 2023; Murtadha et al., 2015). Furthermore, the packets might be out of order due to path diversity and unequal transmission delays, and this negatively impacts the performance of the transport-layer and real-time applications. Moreover, excessive handovers cause large signaling overheads, adding control-plane blockage and processing delays. The overall effect of packet loss, packet reordering, as well as excessive signaling reduces the capability of NB-DMM to support smooth mobility, especially in high mobility and heterogeneous deployment.
Several previous studies have focused on homogeneous access networks. Current mobility management also brings about extra complexity with the incorporation of heterogeneous access technologies, that is, Wi-Fi 7, 5G, and future 6G networks. Such technologies vary in the frame length, scheduling behavior, and latency profiles, which enhance the chances of the loss and rearrangement of packets across handover. This drives the desire to have a homogeneous framework of handover optimization that can manage the heterogeneous access networks and provide continuity of packets, ordering, and efficient signaling. A powerful buffering and signaling strategy are very necessary to enable credible real-time communication among the various wireless technologies.
The key contributions of the work are listed below:
It has been proposed that our previous work (Mahenthiran and Muruganandam, 2026), Dual buffering and signaling aggregation (H-DBSA), can be applied in heterogeneous environments to improve handover performance in NB-DMM networks.
An integrated dual buffering operation is implemented to reduce the amount of packet loss and out-of-order delivery of packets during handover.
A logical frame mapping scheme is created to align the packet forwarding of heterogeneous access technologies with various frame structures.
Signaling aggregation strategy is used to minimize the overhead of the control plane in times of mobility events.
We simulated the proposed framework with OMNeT++ and evaluated it. Its effectiveness is validated in its results.
The Dual Buffer Signaling Aggregation (DBSA) mechanism was initially proposed to enhance handover performance in a homogeneous environment. In this work, the concept is employed and applied to heterogeneous access networks. The proposed heterogeneous H-DBSA scheme enables orchestrating dual buffering and signal fusion across multiple radio access technologies, including Wi-Fi, 5G, and 6G. This extension introduces additional issues, including vertical handover control, cross-technology buffering, and signaling adaptation across dissimilar access regions. This paper considers the DBSA mechanism in heterogeneous mobility and performs a sensitivity analysis of buffer sizes and signaling delays. The uniqueness of this work is that it applies several access technologies to implement the DBSA mechanism, assesses it in the long term, and offers an in-depth analysis.
The rest of this paper is structured as follows. Section II is a literature survey of mobility management and NB-DMM. Section III explains the methodology, system model, and the suggested framework, and gives the performance metrics. In section IV, the results and analysis are discussed. Lastly, Section VI presents the conclusion.
2 Literature survey
The literature has recently examined various strategies to enhance handover performance in distributed mobility settings. Distribution-based cost studies of DMM schemes have revealed that optimizing handover mechanisms can save substantial signaling and packet-delivery costs relative to conventional mobility protocols (Aman et al., 2024; Balfaqih et al., 2022; Jung and Ali Wagan, 2018). Nevertheless, all these methods focus solely on the control-plane efficiency and do not address packet loss or reordering during the handover process. Equally, mobility control based on software-defined networking (SDN) has been proposed, including dynamic cell selection and centralized decision-making to enhance handover efficiency in heterogeneous networks (Dao and Vo, 2025; Hakimi et al., 2018). These solutions help with signaling and load balancing, but they usually depend on centralized control elements.
As network environments become increasingly heterogeneous, the frequency of vertical handovers between technologies rises. Recent surveys on 5G and newer mobility management highlight challenges in maintaining fast, stable, and efficient connections in these settings. They also find that vertical handovers are more difficult than traditional horizontal ones because of differences in technology, signaling, and link characteristics (Elsayed et al., 2023; Goh et al., 2022; Malathy and Muthuswamy, 2018; Satapathy and Mahapatro, 2023). While there are several strategies for handling these handovers, most focus on decision-making, selecting access points, or balancing loads, rather than ensuring packets remain continuous during the handover (Cordova et al., 2019; Khan et al., 2025).
Buffering-based handover mechanisms have also been studied to minimize packet loss due to mobility. These techniques store the incoming packets as the mobile node moves between access routers. Buffering in a distributed mobility environment is typically implemented in edge nodes or access routers to buffer packets during path switching (Davaasambuu, 2018). These techniques temporarily store packets in-flight as the mobile node moves between access routers. Buffering in a distributed mobility environment is typically implemented in edge nodes or access routers to buffer packets during path switching. Such techniques can minimize the number of dropped packets, but they are commonly based on one-buffer architectures and can cause packet reordering when buffer release is not coordinated with the new data path. Further, most buffering strategies are designed to operate in a non-heterogeneous network setting rather than to coordinate across heterogeneous access technologies. The other significant area of research is the minimization of signaling overhead during handover. In a distributed mobility environment, large numbers of handovers may occur, leading to excessive control messages that increase latency and network congestion. Several researchers have proposed signaling aggregations, predictive handover schemes, and SDN-caliber control structures to reduce signaling costs. The techniques would enhance the control-plane efficiency, but they usually do not ensure data-plane or packet continuity during the handover period. Even more recently, machine learning-based and intelligent handover methods have been introduced to reduce the number of handover choices. These methods use data-driven models, reinforcement learning, or neural networks to minimize unnecessary handoffs and enhance network performance (Zaid et al., 2024). Although these approaches improve decision accuracy and resource use, they primarily emphasize when and where handovers should occur. They do not treat packet preservation, buffering coordination, and signal aggregation in the real handover process. Based on the available literature, it is clear that most mobility solutions address either signaling optimisation, buffering, or handover decision mechanisms individually. The number of works that combine signaling aggregation with coordinated buffering is low, and even fewer analyse such mechanisms in heterogeneous environments that involve vertical handovers among different access technologies (Ezz-Eldien et al., 2020; Mohsin et al., 2023; Ullah et al., 2023). Moreover, the collaboration between buffering processes and signaling mechanisms in distributed heterogeneous networks has not been sufficiently studied (Hatipoglu et al., 2020; Kiran and Rao, 2022; Rastogi et al., 2026; Sudarsanan et al., 2025). Such constraints drive the desire for a coordinated mechanism that mitigates signaling overhead, packet loss, and packet reordering simultaneously across heterogeneous distributed mobility landscapes. Rasheed et al. (2022) introduced a dataflow strategy for IoT devices across different network types. Their approach focuses on choosing the best links to boost throughput and reduce delays. Arfeen et al. (2021) presents optimization techniques for enhancing network efficiency within heterogeneous environments. To address this gap, this paper introduces the heterogeneous implementation of the Dual Buffer Signaling Aggregation (H-DBSA) scheme, which provides coordinated dual-buffer functionality with signaling aggregation to enhance handover performance between Wi-Fi, 5G, and new 6G access networks.
3 Proposed method
3.1 System model
This section shows the system model and network structure that the proposed Heterogeneous Dual Buffering and Signaling Aggregation (H-DBSA) framework uses. The architecture under consideration is built upon Network-Based Distributed Mobility Management (NB-DMM) and is expanded to accommodate the heterogeneous wireless access technologies, including Wi-Fi, 5G, and 6G networks of the future. The H-DBSA framework that has been suggested has the following important network entities: mobile node, MAAR, CMD, and CN.
3.2 Proposed architecture
The H-DBSA architecture proposed is a further development of the NB-DMM framework with coordination of the buffering and signaling aggregation process on the MAAR level. Clear separation of both the control-plane and data-plane operations is done to enhance scalability and performance. The CMD in the control plane handles mobility binding management and aggregation of signaling messages sent during the handover, and thus, signaling redundancy is minimized. MAARs exchange messages with the CMD to receive updated mobility context knowledge without exchanging messages too much. Each MAAR in the data plane has two logically different buffers to handle packet forwarding in case of handover. These buffers work together to maintain the delivery of data and the sequence of packets when the MN switches between heterogeneous access networks. The proposed architecture is shown in Figure 1.
Figure 1
3.3 Logical frame mapping
Wi-Fi, 5G, and 6G have different frame structures and timing properties in heterogeneous network situations. These differences can cause packets received during a handover to be out of time synchronization, leading to packet loss and out-of-order delivery. To solve this problem, the proposed H-DBSA framework provides a logical frame mapping mechanism. A logical frame is a shared time reference that allows packets from different networks to be synchronized on a single timeline. The logical frame does not depend on a particular technology, unlike physical frames, but rather on an abstract temporal framework for synchronizing packet delivery.
In general, the logical frame time is determined as in Equation 1
Where TLFdenotes logical frame duration, sfi indicates scaling factor to map different frame duration to a common reference time, Ti denotes duration of network i.
The logical frame in this work is constructed by combining several frames from various networks. For example, a Wi-Fi frame, two 5G slots, and four 6G frames are mapped to a logical frame and expressed as in Equation 2
n1, n2, n3 represents integer scaling values considered in this paper.
n1 = 1, n2 = 2, n3 = 4 are selected, based on the relative differences in frame durations among the considered networks. Wi-Fi typically uses longer frame intervals, while 5G and 6G use increasingly shorter transmission units. By using integer multiples, several frames of higher networks can be bundled together to correspond to the time of a single frame of a slower network, allowing time alignment across heterogeneous systems. This method makes synchronization easier, is less resource-demanding, and grants efficient buffering during handover. This ratio is scenario-specific and may vary depending on the actual network frame characteristics.
To deal with timing differences, a synchronization condition shown in Equation 3 is incorporated into the proposed method to enable packet alignment from different networks:
tarrival is the time(instant) when a packet arrives.
tLF is the reference time of the logical frame used for packet alignment, δ is the allowable tolerance.
Buffering at the receiver allows small differences in arrival times within the specified tolerance. The logical frame mapping is used with the dual buffering mechanism. Packets sent by the old network are stored in the forward data buffer, while those sent by the new network are stored in the new data buffer. Such packets are then released with the logical frame alignment, which helps to maintain proper order.
It is important to note that the logical frame does not assume the same frame durations across Wi-Fi, 5G, and 6G. Instead, it is a more abstract version of time alignment. Practical aspects such as delay variation, network traffic, and buffer restraints make packet loss unavoidable. However, the proposed solution drastically reduces packet loss and improves packet sequence when passing through the hands-off. Logical frame mapping for different access networks is shown in Figure 2.
Figure 2
3.4 Mobility framework
A linear mobility model is employed in this study, where the mobile node moves along a predetermined path between access routers. The H-DBSA framework uses this model to create a controlled and predictable environment for studying handovers. The main purpose is to determine how long it takes to hand over a packet, how many packets are lost, and how packets are ordered. The linear mobility model simplifies tracking these variables by limiting random user movement. It also helps to explain how the proposed mechanisms, including dual buffering, signaling aggregation, and logical frame mapping, affect performance.
Linear mobility is common in real life, for example, vehicles on highways, trains, or people moving in city corridors where movement tends to be predictable. For this reason, it is a useful reference for measuring handover performance. It is, however, recognized that the linear mobility model does not fully reflect the dynamism of real-world user movement, where speed and direction can change over time. Such behavior can be better modeled using more realistic mobility models, such as the Random Waypoint, Gauss-Markov, and Manhattan grid models. It is noted that the proposed H-DBSA framework is independent of the mobility model and can be applied to other mobility patterns without modification. Assessing the framework in stochastic mobility models will be future work to improve its performance in a more dynamic, realistic network setting. Figure 3 depicts the mobility and handover model.
Figure 3
3.5 Handover framework
This is a four-phase handover model in the H-DBSA framework that complies with NB-DMM principles and improved to support dual buffering.
Handover detection: the signal quality/mobility pattern of the MN is monitored by the serving MAAR. Once the quality of links falls below the threshold, a handover trigger is created. It is important to detect this early to enable buffers to be activated so that there is no loss of packets.
Handover preparation: the target MAAR receives notification of the handover that is going to take place. The CMD is updated with mobility context and buffers (Forward Data Buffer and New Data Buffer) are pre-allocated. The UDP packet forwarding policies are modified in order to have continuity in the transition.
Handover execution: the MN breaks off the serving RAT and forms a connection with the target RAT. In this stage, the FDB holds in-flight UDP packets of the serving MAAR, and the NDB initiates buffering packets that come to the target MAAR.
Handover completion: once the MN has been attached to the target RAT, FDB and NDB packets are sent to the MN sequentially without losing any packets and in the correct sequence. The active MAAR binding is updated by the CMD, and control signaling of this handover is combined to minimize overhead.
3.6 Proposed algorithm description
Step 1: handover triggering
The serving MAAR tracks the MN's link quality and mobility. If the connection worsens or a switch to another RAT is likely, a handover is triggered.
Step 2: signaling aggregation
Rather than sending multiple signaling messages, the serving MAAR sends a single combined handover request to the CMD. The CMD checks the MN's details, picks the target MAAR, and shares the control information.
Step 3: dual buffer activation
When the handover starts:
The FDB at the serving MAAR starts storing UDP packets that are still in transit.
The NDB at the target MAAR stores packets that arrive after the path is redirected.
Step 4: logical frame mapping
Packets from different RATs are put into a shared logical frame space using sequence numbers and timestamps. This approach hides the physical frame differences between Wi-Fi, 5G, and 6G.
Step 5: handover execution
The MN disconnects from the serving MAAR and connects to the target MAAR. The CMD updates the binding and forwarding information.
Step 6: coordinated packet release
Packets from the FDB and NDB are combined and sent in the correct order, ensuring loss-free, ordered delivery.
Step 7: handover completion
Once the buffers are empty, normal packet forwarding continues through the target MAAR.
3.7 Simulation environment
The proposed scheme is tested with the help of OMNeT++ and the INET framework that assists in wireless networking, mobility, and UDP traffic. The network topology comprises several MAARs that are connected to the CMD and the CN through a wired backbone. Both MAARs offer wireless services to the MN. The handover process happens when the MN leaves the coverage area of one MAAR and enters the coverage area, based on a predefined signal strength threshold. The MN follows a linear mobility model with unitary velocity to ensure reproducible, predictable behavior at handover. The MN travels straight between the serving MAAR and the target MAAR, crossing the overlap area where coverage triggers a handover. MN has a speed range of 1–20 m/s, which covers pedestrian and vehicular mobility scenarios. The MN does not stop moving at any point and maintains uniform conditions for handover across all simulation runs. The UDP constant bit rate (CBR) communication between the CN and the MN is used to generate user traffic during the simulation. The handover process does not interrupt traffic, allowing monitoring of the effects of packet loss, service interruptions, and packet reordering. Dual buffers are employed in the proposed H-DBSA scheme to maintain packets at the serving and target MAARs, keeping them intact during handover and delivering them in order after path switching. The default simulation parameter used in this evaluation is shown in Table 1.
Table 1
| Type | Parameter | Value |
|---|---|---|
| Network | Simulation area | 1000 m × 1000 m |
| Number of MAARs | 2–6 | |
| Cell radius | 150–250 m | |
| Mobility model | Linear mobility | |
| MN speed | 1–20 m/s | |
| Simulation time | 200 s | |
| Traffic | Transport protocol | UDP |
| Traffic type | CBR | |
| Packet size | 512 bytes | |
| Packet interval | 10 ms | |
| Flow direction | CN → MN | |
| Buffers | Forward buffer size | 50 packets |
| New data buffer size | 50 packets | |
| Buffer activation delay | 5 ms | |
| Buffer release delay | 10 ms | |
| Signaling | Aggregation window | 8 ms |
| CMD processing delay | 5 ms |
Simulation parameters.
3.8 Analytical justification for the proposed method
The analytical framework incorporates constraints related to finite buffer capacity and signal latency to effectively depict realistic heterogeneous handover scenarios. Consequently, the performance metrics are confined within certain limits, rather than presuming an idealized or zero-loss operational state.
3.8.1 Handover latency
According to (Balfaqih et al. 2016), the handover latency of the network-based DMM is given by Equation 4.
Algorithm 1 describes the H-DBSA Procedure for Heterogeneous Network Handover and it is mentioned below.
Algorithm 1
| Input: Mobile device (MN), serving MAAR, target MAAR, Central mobility database (CMD), Correspondent node (CN), RAT = [Wi-Fi, 5G, 6G] |
| Output: Smooth handover with packets delivered in order, reduced HO Latency, PL |
| 1. Monitor the link quality at the Mobile Node (MN). |
| 2. If the link quality drops below the set threshold, follow these steps: |
| 3. Initiate the handover |
| 4. The MN sends a handover request to the target MAAR. |
| 5. The target MAAR authenticates the MN and allocates the necessary resources. |
| 6. The target MAAR updates the CMD with the MN's new location. |
| 7. CMD updates the binding information for MN, serving MAAR, and target MAAR |
| 8. Set up a forwarding path from the serving MAAR to the target MAAR. |
| 9. Enable dual buffering at both the MAARs. |
| At serving MAAR |
| 10. Initialize the Forward Data Buffer (FDB_serving) and New Data Buffer (NDB_serving). |
| 11. Store incoming packets in FDB_serving. |
| 12. Forward buffered packets from FDB_serving to the target MAAR. |
| At the target MAAR |
| 13. Initialize: Forward Data Buffer (FDB_target), New Data Buffer (NDB_target) |
| 14. Receive packets that have been transmitted from the serving MAAR |
| 15. Forwarded packets are then stored in the FDB_target |
| 16. Receive new incoming packets from the CN |
| 17. Store new packets from CN in NDB_target |
| Logical frame |
| 18. Set logical frame time T_LF |
| 19. Map packets from FDB_target and NDB_target into a logical frame using scaling factors |
| Synchronization |
| 20. For each packet |
| compute tarrival |
| determine the reference time of the logical frame tLF |
| IF |tarrival-tLF | ≤ δ THEN |
| Mark the packet as synchronized |
| Packet delivery |
| 21. Release the forwarded packets from FDB_target. |
| 22. Release the NDB_target packets according to the logical frame alignment. |
| 23. Make sure packets are delivered in the correct order. |
| 24. Send the ordered packets to the MN. |
| Handover completion |
| 25. Clear the buffers once transmission is completed. |
| 26. Stop forwarding packets from the serving MAAR. |
| 27. Resume normal communication using the target MAAR. |
| 28. End the handover process |
H-DBSA procedure for heterogeneous network handover.
Proposed
Where indicates the reduced delay.
In the proposed method, signaling aggregation reduces the number of signaling messages and it is mentioned in Equation 5. As a result, the handover latency of the H-DBSA is lower than that of the existing network-based DMM.
HL with bound
The bound equation shows the lowest and highest latencies that can happen in different conditions as mentioned in Equation 7.
3.8.2 Packet loss
The traditional methods, such as NB-DMM, do not provide sufficient buffering, so packet loss occurs whenever packets are received during interruptions. Thus, packet loss is directly proportional to the handover latency. The PL formula is shown in Equation 8.
In the proposed H-DBSA framework, two buffers are provided at the MAARs to buffer packets temporarily during handover. The loss of a packet can then be represented as in Equation 9
Buf sizeeffective be the total buffering capacity at which packets can be stored.
Moreover, the suggested H-DBSA framework will minimize the loss of packets in two ways: Dual Buffering and reduced handover latency due to signal aggregation. In this way, the improvement of the buffer capacity, as well as the decrease in the handover latency, helps to reduce the packet loss in the proposed method. It is shown in Equation 10.
In real-world networks, it is not possible to eliminate packet loss completely due to delay variations and buffer limits. However, the H-DBSA framework reduces packet loss much more than the default NB-DMM solution.
3.8.3 Out-of-order packets
The difference between the delay of the packets coming in the old path and the new path is the major cause of out-of-order delivery of packets during handover. This can be expressed as shown in Equation 11
In the proposed method, the synchronization condition is used to ensure the difference in the delay within a tolerance and it is mentioned in Equation 12
The Equation 13 shows that the proposed approach minimizes delay mismatch and greatly minimizes out-of-order delivery of packets when compared to the baseline NB-DMM approach.
3.8.4 Throughput
Throughput is the amount of information transmitted successfully in a given period of time.
The throughput for the baseline NB-DMM is expressed in Equation 14
The proposed method reduces packet loss due to buffering and handover latency due to signaling aggregation. Therefore, the proposed system achieves a higher throughput compared to the baseline NB-DMM and it is shown in Equations 15 and 16
3.8.5 Signaling cost
In the baseline NB-DMM method, signaling operations, including binding updates, acknowledgments, and tunnel establishment, are performed separately, resulting in more control message exchanges and increased signaling overhead. Consequently, the cost of signaling increases and depends on the signaling steps involved in the handover process. It is shown in Equation 17.
Signaling aggregation is proposed in the H-DBSA framework, which means that several control messages are likely to be sent as fewer messages. Rather than sending each signaling message individually, signaling operations that are related to each other are bundled together, thus decreasing the number of messages sent.
α indicates reduction factor.
Smaller values of alpha lead to higher signaling aggregation and greater reduction in signaling cost. Since the signaling aggregation reduces the number of control messages, the signaling cost of proposed is lower when compared to NB-DMM. Equations 17 and 18 indicate the signaling cost of the proposed method.
4 Results and discussion
This section examines how the proposed H-DBSA scheme performs better than NB-DMM. We look at handover latency, packet loss, out-of-order packets, throughput, and signaling cost. All metrics are analyzed over the handover period to see how mobility affects service continuity directly. The analysis shows small amounts of packet loss and less packet reordering, which more closely reflect how buffering limits and signaling delays work in the real world than idealized models. Table 2 compares performance metrics across variable mobility speeds.
Table 2
| Latency | Packet loss | OOP | Throughput | |||||
|---|---|---|---|---|---|---|---|---|
| Mobile node speed (m/s) | NB-DMM | Proposed HDBSA | NB-DMM | Proposed HDBSA | NB-DMM | Proposed HDBSA | NB-DMM | Proposed HDBSA |
| 1 | 86 | 43 | 10 | 3 | 2.3 | 0.7 | 371 | 403 |
| 5 | 99 | 49 | 12 | 4 | 2.9 | 1.1 | 365 | 396 |
| 10 | 116 | 56 | 14 | 5 | 3.5 | 1.5 | 356 | 389 |
| 15 | 133 | 64 | 17 | 6 | 4.1 | 1.9 | 346 | 381 |
| 20 | 151 | 73 | 20 | 8 | 4.7 | 2.6 | 335 | 371 |
Performance comparison under variable speed.
4.1 Speed-based comparison
Handover latency refers to the period of service interruption experienced by a mobile node during transitions between two MAARs. The baseline NB-DMM protocol requires more signaling exchanges and path reconfiguration pauses, leading to increased latency as node speed rises. Figure 4 show that the proposed method consistently reduces handover latency regardless of node speed. The adoption of signaling aggregation and dual buffers significantly mitigates service interruptions caused by the handover process. However, at faster speeds, extra delay occurs, reflecting the latency typically encountered in real-world signaling and processing.
Figure 4
Packet loss is one of the most important metrics for assessing service reliability during handover. In NB-DMM, packets that arrive during the mobility period are discarded, as there is no buffering. Figure 5 indicates that the proposed HDBSA significantly reduces data loss at variable mobility speeds. The higher the mobile node speed, the higher the packet loss in NB-DMM, as the handover frequency and speed increase. But H-DBSA exhibits much lower packet loss, demonstrating the feasibility of the dual-buffer mechanism.
Figure 5
As shown in Figure 6, in NB-DMM, there are significant out-of-order packet counts due to parallel packet forwarding during handover. H-DBSA, on the contrary, minimizes the effect. Data packets are sent in the correct order, with the controlled release of buffered packets making sure that the old path message is sent first before the new path data. This proves the usefulness of the proposed method.
Figure 6
As shown in Figure 7, H-DBSA has a higher throughput than NB-DMM. The throughput of the proposed system is fairly stable, which proves that this system can cope with changing mobility conditions.
Figure 7
Figure 8 depicts that H-DBSA consumes less signaling than NB-DMM. The reduction in signaling cost confirms the reliability of the proposed approach.
Figure 8
4.2 Sensitivity analysis
A sensitivity analysis of the buffer size is conducted (Table 3). As the buffer size increases, PL slows gradually, and OOP decreases. This improvement is mainly due to more storage capacity during the handoff. Throughput also increases when the buffer size grows. Even though there is a small amount of PL, OOP is still observed. This behavior is mainly due to signaling delays. Figures 9, 10 show that the proposed system maintains stable performance across variable buffer sizes.
Table 3
| Size of buffer | PL | OOP | Throughput |
|---|---|---|---|
| 10 | 8 | 2.2 | 360 |
| 20 | 6 | 1.7 | 373 |
| 30 | 5 | 1.3 | 383 |
| 40 | 4 | 0.9 | 393 |
| 50 | 3 | 0.8 | 403 |
Sensitivity analysis based on buffer size.
Figure 9
Figure 10
Signaling delay sensitivity analysis (Table 4) examines how increasing the CMD processing delay affects handoff performance. As signaling time rises, HL PL, and OOP increase slightly. This occurs because longer signaling delays extend the time required for the control plane to change the path and activate buffers, resulting in greater handover disturbances. The buffer may become strained during this period, occasionally causing packet drops or reordering. However, the H-DBSA scheme maintains performance within acceptable limits. These results show that signaling aggregation and dual-buffer systems can tolerate minor control-plane delays without significant performance impact. The scheme can accommodate typical variations in signaling latency, and it is shown in Figure 11.
Table 4
| Delay | Latency | PL | OOP |
|---|---|---|---|
| 2 | 51 | 3 | 1.0 |
| 4 | 54 | 3 | 1.3 |
| 6 | 58 | 4 | 1.7 |
| 8 | 61 | 4 | 1.9 |
| 10 | 66 | 5 | 2.4 |
Sensitivity analysis based on CMD delay.
Figure 11
4.3 Worst-case scenario analysis
The worst-case scenario analysis assesses the feasibility of the suggested scheme under unfavorable operating conditions, such as high mobility speed, reduced buffer size, and increased signaling delay. These situations cause high latency in handover, loss of packets, and out-of-order delivery of packets as compared to the normal operating conditions. The high-speed scenario results in the highest latency and packet loss due to the reduced handover preparation time. Equally, small buffer sizes reduce the system's capacity to buffer in-flight packets, leading to more packets being dropped. The growth of signaling delay also causes increased handover time and moderate reordering. Nevertheless, the proposed H-DBSA scheme will always have lower latency, lower packet loss, and fewer out-of-order packets than the default NB-DMM in all these worst-case scenarios. It is shown in Table 5 and Figure 12.
Table 5
| Scenario | Latency (NB-DMM) | Latency (Proposed H-DBSA) |
|---|---|---|
| Normal 10 m/s | 113 | 56 |
| High speed 20 m/s | 145 | 73 |
| Small buffer | 129 | 64 |
| High delay | 113 | 66 |
Worst-case scenario.
Figure 12
The simulations are conducted in a controlled and deterministic setting, in which the input parameters are held constant across the scenarios. Consequently, repeated experiments yield the same results by a small difference. Moreover, the sensitivity analysis and worst-case analysis further prove that the performance gains are not highly dependent on the changes in the network conditions but are rather stable.
5 Conclusion
This paper presents the performance analysis of the Dual Buffer Signaling Aggregation (DBSA) mechanism in a heterogeneous network environment. Although we have already proposed the DBSA to enhance handover performance in distributed mobility management, the present study extends the mechanism to the heterogeneous access case and evaluates its performance in practice under various network conditions.
The proposed heterogeneous scheme incorporates signaling aggregation and coordination among dissimilar access technologies via a dual-buffer scheme. The forward buffer maintains incoming packets across the handover interval, whereas the new data buffer maintains continuity and orderly data delivery after a path switch. The signaling aggregation procedure minimizes control overhead by consolidating multiple mobility messages into fewer exchanges, thereby enhancing scalability.
The simulation results indicate that heterogeneous execution of H-DBSA consistently minimizes handover delay, packet loss, and out-of-order delivery compared to the normal execution of NB-DMM across different mobility speeds. The suggested scheme improves throughput and minimizes the signaling cost. A sensitivity analysis of buffer sizes and signaling delays shows that the mechanism maintains similar performance across varying operating scenarios. Despite small residual packet loss and packet reordering at higher speeds or with limited buffers, the overall performance is much better than that of the baseline scheme.
Statements
Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.
Author contributions
PM: Conceptualization, Data curation, Investigation, Software, Visualization, Writing – original draft. DM: Funding acquisition, Investigation, Supervision, Validation, Writing – review & editing.
Funding
The author(s) declared that financial support was not received for this work and/or its publication.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Nomenclature
| Abbreviations | Full form |
|---|---|
| NB-DMM | network-based distributed mobility management |
| DBSAHO | dual buffering and signaling aggregation of handover optimization |
| H-DBSA(Proposed) | heterogeneous dual buffering and signaling aggregation |
| MAAR | mobile access and anchor routers |
| FDB | forward data buffer |
| NDB | new data buffer |
| MN | mobile node |
| CMD | centralized mobility database |
| CN | correspondent node |
| SMAAR | serving MAAR |
| TMAAR | target MAAR |
| SigCostNBDMM | signaling cost of network-based DMM |
| SigCostH−DBSA | signaling cost of the proposed |
| α | aggregation factor |
| λ | packet arrival rate |
| N | number of handoffs |
| Γ | amount of reduction |
| PL | packet loss |
| HL | handover latency |
| OOP | out of order packets |
| TL2 | delay in link transfer |
| Tpath | delay in setting up the path |
| TSignaling delay | delay in signaling |
| TObservation | observation time |
| Psize | size of the packet |
| Buf size | size of the buffer |
| Ntransmitted packets | number of packets transmitted |
References
1
AmanM.ZubairM.BadshahA.FarmanH.Ur RehmanG.KhattakS. B. A.et al. (2024). Cost analysis of IPv6 distributed mobility management protocols in comparison with TFMIPv6. PLOS ONE19:e0306132. doi: 10.1371/journal.pone.0306132
2
ArfeenZ. A.SheikhU. U.AzamM. K.HassanR.Faisal ShehzadH. M.AshrafS.et al. (2021). A comprehensive review of modern trends in optimization techniques applied to hybrid microgrid systems. Concurr. Comput. Pract. Exp. 33:e.6165. doi: 10.1002/cpe.6165
3
BalfaqihM.BalfaqihZ.ShepelevV.AlharbiS. A.JabbarW. A. (2022). An analytical framework for distributed and centralized mobility management protocols. J. Ambient Intell. Humaniz. Comput.13, 3393–3405. doi: 10.1007/s12652-020-01749-x
4
BalfaqihM.IsmailM.NordinR.BalfaqihZ.YuwonoT. (2016). “Design and evaluation of network-based distributed mobility management solution based on PFMIPv6,” in 2016 2nd International Conference on Wireless and Telematics (ICWT), 132–139. doi: 10.1109/ICWT.2016.7870867
5
CordovaR. T.GondimP. R. L.LlerenaY. P.LloretJ. (2019). SDN-DMM for intelligent mobility management in heterogeneous mobile IP networks. Int. J. Commun. Syst. 32:e4140. doi: 10.1002/dac.4140
6
DaoD.-T.VoD.-B. (2025). A software-defined networking-based load-balanced handover control scheme for distributed mobility management in vehicular communication. Univ. Danang J. Sci. Tech.23, 55–61. doi: 10.31130/ud-jst.2025.23(7)0.250E
7
DavaasambuuB. (2018). Handover with buffering for distributed mobility management in software defined mobile networks. Austral. J. Telecomm. Digit. Econ.6, 26–40. doi: 10.18080/jtde.v6n1.137
8
ElsayedM. M.HosnyK. M.FoudaM. M.KhashabaM. M. (2023). Vehicles communications handover in 5G: A survey. ICT Express9, 366–378. doi: 10.1016/j.icte.2022.01.005
9
Ezz-EldienN. A.AbdelkaderM. F.AbdallaM. I.Abdel-AttyH. M. (2020). “Handover performance improvement in heterogeneous wireless network,” in 2020 11th IEEE Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON) (Vancouver, BC), 0821–0830. doi: 10.1109/IEMCON51383.2020.9284906
10
GohM. I.YewH. T.WenD. B.KiringA. M. J.KheauC. S. (2022). Vertical handover decision making algorithms for 5G heterogeneous networks: a review. Int. J. Emerg. Techn. Adv. Eng.12, 185–199. doi: 10.46338/ijetae0122_19
11
HakimiR.LarasatiH. T.MustafaA.Abu-ArabiyaA. (2018). “Leveraging SDN for handover in distributed mobility management of 5G network,” 2018 12th International Conference on Telecommunication Systems, Services, and Applications (TSSA) (Yogyakarta), 1–7. doi: 10.1109/TSSA.2018.8708820
12
HatipogluA.BasaranM.YaziciM. A.Durak-AtaL. (2020). “Handover-based load balancing algorithm for 5G and beyond heterogeneous networks,” 2020 12th International Congress on Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT) (Brno), 7–12. doi: 10.1109/ICUMT51630.2020.9222456
13
JungL. T.Ali WaganA. (2018). “Distributed network mobility management scheme for network mobility,” 2018 4th International Conference on Computer and Information Sciences (ICCOINS) (Kuala Lumpur), 1–6. doi: 10.1109/ICCOINS.2018.8510603
14
KhanA.AhmadS.AliI.HayatB.TianY.LiuW. (2025). Dynamic mobility and handover management in software-defined networking-based fifth-generation heterogeneous networks. Int. J. Netw. Manag.35:e2268. doi: 10.1002/nem.2268
15
KiranK.RaoD. R. (2022). Analytical review and study on various vertical handover management technologies in 5G heterogeneous network. Infocommun. J.14, 28–38. doi: 10.36244/ICJ.2022.2.3
16
MahenthiranP.MuruganandamD. (2026). DB-SAHO network-based distributed mobility management: enhancing handover performance in mobile networks. Front. Commun. Netw.7:1771623. doi: 10.3389/frcmn.2026.1771623
17
MalathyE. M.MuthuswamyV. (2018). State of art: vertical handover decision schemes in next-generation wireless network. J. Commun. Infor. Netw.3, 43–52. doi: 10.1007/s41650-018-0009-2
18
ManeS. N.KhairnarD. G. (2023). “Distributed mobility management with tunneled and tunnel free approach: a compressive literature review,” 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT) (Delhi), 1–7. doi: 10.1109/ICCCNT56998.2023.10308028
19
MohsinH. S.SaadW. K.ShayeaI. (2023). “Literature review of handover decision algorithms in 5G networks,” 2023 10th International Conference on Wireless Networks and Mobile Communications (WINCOM) (Istanbul), 1–6. doi: 10.1109/WINCOM59760.2023.10322978
20
MurtadhaM. K.NoordinN. K.AliB. M.HashimF. (2015). Design and evaluation of distributed and dynamic mobility management approach based on PMIPv6 and MIH protocols. Wireless Netw.21, 2747–2763. doi: 10.1007/s11276-015-0950-z
21
RasheedZ.AshrafS.IbupotoN. A.ButtP. K.SadiqE. H. (2022). SDS: scrumptious dataflow strategy for IoT devices in heterogeneous network environment. Smart Cities5, 1115–1128. doi: 10.3390/smartcities5030056
22
RastogiK.MeghaK. B.GhoshS. K.SinghR. K. (2026). Evaluating DMM based handover solutions towards serving uRLLc in heterogeneous networks (pp. 162–176). doi: 10.1007/978-3-032-14350-1_12
23
SaoudB.ShayeaI.AlnakhliM. A.MohamadH. (2025). Mobility and handover management in 5G/6G networks: challenges, innovations, and sustainable solutions. Technologies13:352. doi: 10.3390/technologies13080352
24
SatapathyP.MahapatroJ. (2023). An adaptive context-aware vertical handover decision algorithm for heterogeneous networks. Comput. Commun.209, 188–202. doi: 10.1016/j.comcom.2023.06.029
25
SudarsananS.EhssanM.YusupovaS.JumaniyazovF.JuraevS.MishraN. (2025). Seamless handover mechanisms in heterogeneous access networks. J. Int. Serv. Infor. Secur.15, 749–759. doi: 10.58346/JISIS.2025.I2.049
26
UllahY.RosleeM.BinMitaniS. M.KhanS. A.JusohM. H. (2023). A survey on handover and mobility management in 5G HetNets: current state, challenges, and future directions. Sensors23:5081. doi: 10.3390/s23115081
27
ZaidM.KadirM. K. A.ShayeaI.MansorZ. (2024). Machine learning-based approaches for handover decision of cellular-connected drones in future networks: a comprehensive review. Eng. Sci. Tech. Int. J.55:101732. doi: 10.1016/j.jestch.2024.101732
Summary
Keywords
dual buffer, heterogeneous network, network-based DMM, signal aggregation, vertical handover
Citation
Mahenthiran P and Muruganandam D (2026) A DBSAHO scheme for smooth movement in heterogeneous network-based distributed mobility systems. Front. Comput. Sci. 8:1818011. doi: 10.3389/fcomp.2026.1818011
Received
26 February 2026
Revised
09 April 2026
Accepted
14 April 2026
Published
08 May 2026
Volume
8 - 2026
Edited by
Shahzad Ashraf, Gachon University, Republic of Korea
Reviewed by
Zeeshan Rasheed, Mir Chakar Khan Rind University, Pakistan
Wali Hussain, Dow University of Health Sciences Institute of Business and Health Management, Pakistan
Updates
Copyright
© 2026 Mahenthiran and Muruganandam.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Dinakaran Muruganandam, dinakaran.m@vit.ac.in
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.