Methods and systems for dynamic routing and optimization of subsets of MQTT message payloads

By working together with the MQTT broker processor and payload controller, a subset of MQTT message payloads is dynamically routed to consumer devices, solving the problem of excessive data delivery in traditional MQTT brokers and achieving efficient resource utilization and system scalability.

CN123123450APending Publication Date: 2026-11-17SIEMENS AG
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610682040.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2025-05-16
Filing Date
2026-05-18
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Traditional MQTT brokers, when processing bulk data, deliver redundant data to consumer devices that do not need the full content of each message. This leads to excessive bandwidth consumption, unnecessary memory allocation, and increased processing burden on edge devices, especially when multiple consumer devices have unique and overlapping subscription interests.

Method used

By working together with the MQTT broker processor and payload controller, a subset of MQTT message payloads is dynamically routed to consumer devices. Synchronization mechanisms and algorithms are used to extract data subsets that meet subscription requirements, generate and name dynamic topics, and deliver only the necessary data, reducing memory and processing resource consumption.

Benefits of technology

It optimizes network resource utilization, reduces unnecessary data transmission and storage, lowers processing and memory overhead for edge devices, improves system scalability and responsiveness, and simplifies message processing complexity for consumer devices.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123123450A_ABST
    Figure CN123123450A_ABST
Patent Text Reader

Abstract

A computer-implemented method and system for dynamically routing subsets of MQTT message payloads to consumer devices via message queue telemetry. The method uses an MQTT broker processor to handle batch payloads comprising multiple data points published by the publisher to an MQTT topic. A payload controller integrated with the processor determines a subset specified by the subscriber based on explicit consumer subscription requirements, generating dynamically named MQTT topics. This minimizes memory and processing load on edge devices by delivering only the relevant subset. The system supports combining data points from multiple publishers into a combined subset for customized delivery. Synchronization between the MQTT broker processor and the controller further reduces redundant data storage and optimizes memory usage. Real-time operational insights are achieved through system resource monitoring and dynamic visualization via the system controller. Dedicated hardware allocation ensures high throughput and responsiveness, supporting efficient and scalable edge deployments.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This disclosure generally relates to the field of computer networking, and more specifically to the management and routing of message payloads in message queuing protocols such as Message Queuing Telemetry Transport (MQTT). In particular, this disclosure relates to computer-implemented methods and systems for the dynamic partitioning, allocation, and delivery of subsets of MQTT message payloads to consumer devices according to a personalized subscription definition standard. Background Technology

[0002] The proliferation of interconnected devices and sensors in the IoT (Internet of Things) field has led to an exponential increase in the demand for real-time data transmission at the edge and cloud computing layers. In response, lightweight messaging protocols such as MQTT have gained widespread adoption due to their efficiency, low overhead, and ability to support publish-subscribe communication models. MQTT brokers have become a core element in IoT architectures by managing traffic between a wide range of publishers and consumer devices across various application domains.

[0003] Traditional MQTT broker implementations are typically designed to route messages as a whole from publishers to subscribers. In environments with bulk data publishing, this often results in delivering redundant data to consumer devices that don't need the full content of each message. Delivering unfiltered or unpartitioned payloads leads to excessive bandwidth consumption, unnecessary memory allocation, and increased processing burden on constrained consumer hardware, especially at the edge.

[0004] These inefficiencies are further exacerbated when multiple consumer devices have unique and overlapping subscription interests. Edge devices, gateways, and analytics systems can become overwhelmed by the need to parse, filter, and aggregate received data locally, respectively. As deployments scale, limitations in storage, processing, and network resources at the edge make efficiently handling unsplit batch message routing increasingly impractical.

[0005] To address these challenges, the industry has been exploring advanced message processing and intelligent routing strategies within the agent infrastructure itself. Such approaches promise to optimize resource utilization, reduce redundant data transmission, and enhance the scalability of IoT solutions, thereby facilitating simplified analytics, responsive control, and efficient capacity management in modern distributed sensor and device networks. Summary of the Invention

[0006] Therefore, one object of this disclosure is to provide a method and system for dynamically routing a subset of multiple Message Queuing Telemetry Transport (MQTT) message payloads to at least one consumer device based on individual subscription requirements.

[0007] The purpose of this disclosure is to achieve this by means of a computer-implemented method for dynamically routing subsets of multiple Message Queuing Telemetry Transport (MQTT) message payloads to at least one consumer device based on individual subscription requirements.

[0008] This method involves receiving bulk payloads published to MQTT topics from at least one publisher by an MQTT broker processor. The bulk payload comprises multiple data points. An MQTT broker processor is defined as a computational entity responsible for receiving, parsing, and initially storing multiple MQTT PUBLISH messages arriving from multiple publishers to multiple topics. An example of an MQTT broker processor is a Mosquitto broker instance running on an edge server in a smart manufacturing plant, processing multiple environmental sensor data streams. An MQTT topic is defined as a logical channel within the MQTT protocol used for message exchange. For example, an MQTT topic could be named "Environment / Data". A bulk payload is a message object encapsulating a grouped set of data points formatted using a serialization standard such as JSON, and includes multiple key-value pairs, such as temperature: 23.5°C, pressure: 1.01 atm, and voltage: 220V. Reception of bulk payloads published to a first MQTT topic from at least one publisher occurs via the MQTT CONNECT and PUBLISH protocol mechanisms, whereby the MQTT broker is configured to listen for incoming PUBLISH packets on a predefined TCP port, such as 1883. Multiple data points are measured sensor values, where each data point can be individually tagged or encapsulated as a field in a JSON or equivalent serialized structure. The MQTT broker is configured to parse the incoming bulk payloads, verify message integrity, and temporarily buffer multiple data points in memory for further processing.

[0009] The method further includes receiving multiple batch payloads from multiple publishers by a payload controller, each batch payload being published to at least one MQTT topic. The payload controller is defined as a programmable component, software module, or hardware integration logic explicitly delegated to analyze multiple input MQTT messages, determine consumer demand, and orchestrate payload partitioning. A real-world example of a payload controller is the PaCo (Payload Controller) plugin deployed within a Mosquitto broker, which examines multiple sensor data received from multiple weather stations. The multiple publishers are a group of devices or software agents configured to push multiple sensor readings (such as temperature, humidity, and pressure sensors), each sensor reading sending structured MQTT messages to multiple specific topics, such as “sensor / temperature,” “sensor / humidity,” and “sensor / pressure.” Receiving multiple batch payloads by the payload controller involves protocol-level interception or callback registration, such that when each MQTT PUBLISH packet arrives at the MQTT broker processor, a notification or reference to the payload is forwarded for routing considerations by the payload controller.

[0010] The method further includes synchronizing the extraction of at least one subscriber-specified subset from the received batch payload between the MQTT broker processor and the payload controller, such that memory resources are allocated for the at least one subscriber-specified subset. Synchronization is achieved through explicit state tracking and mutual memory buffer access protocols between the MQTT broker processor and the payload controller. The synchronization mechanism may include a shared memory segment, message queue, or remote procedure call (RPC) interface, ensuring that the exact timing and content of the subset extraction are coordinated between the MQTT broker processor and the payload controller. The extraction of the at least one subscriber-specified subset is achieved by applying an algorithm to the buffered batch payload data, such as at least one of a filter matching algorithm, a subscription-based parsing algorithm, or a pattern matching algorithm. The algorithm is configured to compare the fields required by the consumer with the input payload data points and select the desired subset, thereby ensuring that memory allocation is proportional only to the minimum required data points, rather than the entire original batch payload. This process allows the payload controller and the MQTT broker processor to allocate memory resources only for the requested data subset, preventing memory waste caused by unnecessarily retaining unwanted data. In one example, the MQTT broker processor and payload controller use a shared semaphore variable to coordinate concurrent access to the internal payload buffer, ensuring that fetching is triggered only after the subscriber specification is updated.

[0011] The method further includes a payload controller integrated with an MQTT broker processor determining at least one subscriber-specified subset from multiple data points based on an explicit subscription request from at least one consumer device. This determination is achieved by processing subscriber registration information, where each subscription specifies a particular field or data point of interest, such as "temperature" and "humidity" only from environmental data payloads. The payload controller is configured to parse the consumer device's explicit subscription request, match the desired fields within the input batch payloads, and select only those data points to be included in the subscriber-specified subset. The determination step is achieved by applying an algorithm to the subscription registry and payload metadata, such as at least one of a rule engine algorithm, a policy-based subscription filter algorithm, or a user profile matching algorithm. The algorithm is configured to identify the intersection between the requested fields and the available data points. For example, if the consumer device requests "humidity," the payload controller is configured to scan the input payload for the keyword "humidity" and extract the corresponding data, such as humidity: 56%. This determination step allows consumer devices to receive only the minimum necessary data, thereby reducing message processing and storage at the edge device compared to traditional methods where the entire batch payload is routed to each subscriber regardless of their individual needs.

[0012] The method further includes a payload controller determining, from at least two of multiple received batch payloads, a combined data point requested by at least one consumer device. This step involves the payload controller analyzing a subscription descriptor to identify a request where the consumer needs to combine data points from individual publishers, such as “temperature” from a weather sensor and “pressure” from a machine sensor. The determination of the combined data point is achieved by applying an algorithm to the input stream of the payload buffer, such as at least one of a multi-source aggregation algorithm, a joint operation algorithm, or a combined set intersection algorithm. The algorithm is configured to dynamically map existing publisher data to any subscriber’s explicit multi-field request. For example, if a consumer request contains independently published reports of “electricity consumption” and “ambient light,” the payload controller combines those values ​​into a single combined subset. This approach significantly simplifies the complexity of message subscription and filtering at the consumer device, thereby reducing the device-side computational resources required for multi-topic aggregation.

[0013] The method further includes generating dynamically named MQTT topics associated with a determined subset specified by at least one subscriber, according to a predetermined naming convention. The dynamic naming convention is defined as a string pattern combining the original topic identifier and a suffix or prefix representing a subset or consumer identifier, such as “temperature_subset_userA”. The payload controller is configured to construct such topic names at runtime based on an internal mapping table and naming logic. Examples of naming conventions include “sensor data / subset / temperature”, “bulk data / aggregate / humidity”, or “device123 / subsetA”. Using dynamically named MQTT topics aligned with subsets allows for an immediate and visible mapping between message content and topic names, thereby improving scalability and traceability compared to traditional static topic mappings.

[0014] The method further includes the payload controller assigning a determined subset, specified by at least one subscriber, to a dynamically named MQTT topic for routing to at least one consumer device. The payload controller is configured to create and maintain a mapping registry associating subset payload buffers with their respective dynamic topic identifiers. For example, the subset "Temp: 22.4°C" is assigned to the topic "EnvData_Subset_Temp". This assignment allows the payload controller to streamline subsequent subscription management and payload delivery.

[0015] The method further includes an MQTT broker processor outputting a determined subset specified by at least one subscriber to at least one consumer device via a dynamically named MQTT topic, thereby reducing memory and processing resource consumption at resource-constrained edge devices. The output process involves the MQTT broker processor serializing only the relevant subset, publishing it under a constructed dynamic topic, and delivering the subset to the network address of the registered subscriber. For example, if "Client 456" subscribes to "AirQuality_Subset_PM25", the MQTT broker processor outputs only PM25 data to that client, saving both bandwidth and memory usage at the consumer device. This targeted delivery is beneficial for IoT gateways with limited CPU and RAM, enabling the support of a larger number of consumer devices without degradation in throughput or response latency. In smart agriculture deployments, consumer devices assigned to irrigation management receive only a subset of soil moisture values, demonstrating the practical benefits of reduced device resource consumption.

[0016] The method further includes generating a composite subset comprising the determined composite data points by a payload controller. The payload controller is configured to listen for multi-topic subscription commands from consumer devices, buffer batches of payloads from multiple publishing sources, and merge the requested data point fields into an integrated structure. For example, the composite subset could be represented as {"Temperature": 22.1, "Pressure": 1.02, "Humidity": 60%}. The generation of the composite subset is achieved by applying an algorithm to payloads collected from different MQTT topics, such as at least one of a buffer merging algorithm, a tuple merging algorithm, or a stream-correlation algorithm. The output of the composite subset enables precise and targeted message delivery, thereby freeing the consumer device from cumbersome context management and message merging operations. For example, an environmental analysis device could subscribe to the "CompositeWeatherTopic" once and automatically receive periodic aggregations of all requested weather indicators.

[0017] The method further includes the payload controller assigning the composite subset to a further dynamically named MQTT topic. The payload controller is configured to construct an appropriate topic identifier, such as "Composite_AQ_TempPressHum," and map the resulting composite subset to that topic. The assignment step is registered in an internal topic payload registry, thereby ensuring correct routing and consumer notification for each additional subset.

[0018] The method further includes a payload controller outputting a combined subset to at least one consumer device, thereby reducing the complexity of message subscription and filtering at at least one consumer device. The payload controller serializes the combined subset and signals the MQTT broker processor for delivery, ensuring that the receiving device receives a minimal, context-rich data object. The consumer device only needs to process the last completed message. For example, in an industrial automation environment, an edge analytics client subscribes to a single topic and immediately receives a ready-to-use composite, improving real-time analytics efficiency compared to subscribing to and merging messages individually.

[0019] This method further includes minimizing redundant data storage and memory usage. This is achieved through the joint operation of the payload controller and the MQTT broker processor, eliminating unnecessary duplicate buffers and maintaining only the necessary subset of the payload in local memory at the edge. Redundancy minimization is achieved through pointer-based memory management techniques, reference counting, or explicit payload lifecycle control protocols after message delivery. Compared to traditional MQTT brokers that store a complete copy of the message for each subscriber, this approach scales efficiently with an increasing number of consumers.

[0020] The method further includes allocating a dedicated network interface queue for MQTT protocol traffic by edge device hardware. Edge device hardware is defined as a physical processor, RAM module, and network interface circuitry embedded within or co-located with the data source. The edge device hardware is configured to establish dedicated queues, such as those exclusively mapped to a network buffer pool on TCP port 1883, to provide priority and resource-isolated transmission paths for MQTT packet reception. Real-world examples of edge device hardware include a Raspberry Pi with an Ethernet interface in a smart grid monitoring station, or a custom embedded controller in a home automation system.

[0021] The method further includes the edge device hardware allocating processor time slices, at least to the MQTT broker processor and payload controller, for message processing. This step involves configuring scheduler-level settings, such as Linux cgroups or RTOS scheduler policies, to establish a minimum CPU time allocation for each internal component, thereby ensuring predictable real-time message routing and payload manipulation, even amidst transient system loads from other operations. Processor time slicing ensures that critical routines in the MQTT broker processor and payload controller remain unsaturated, which is necessary for sustained low-latency operation. For example, in a health monitoring appliance with multiple internal subsystems, the processor reserves a fixed period for MQTT traffic management to prevent message delivery stalls or delays.

[0022] The method further includes allocating a memory buffer by edge device hardware for temporary storage of MQTT payload data, thereby maintaining system throughput and latency. The edge device hardware instance creates a memory buffer within RAM, where the block size is dynamically set based on the expected payload volume. Each temporary buffer retains either the entire payload or a subset of the payload before routing, thus supporting high-throughput data bursts and mitigating traffic surges. The buffer management strategy enables direct handoff and rapid post-delivery memory reclamation, which improves system responsiveness and helps avoid bottlenecks typical of general-purpose buffers. For example, in a hospital patient monitoring network, the allocation of a dedicated MQTT payload memory buffer prevents the loss of vital signs data during peak emergency admissions.

[0023] The method further includes determining resource usage metrics by a system controller within the edge device, which includes at least processor utilization and memory consumption during MQTT message payload processing. The system controller is defined as an embedded monitoring agent or programmable coprocessor responsible for tracking time-averaged and instantaneous system resource loads, such as CPU load (e.g., 43%) and RAM usage (e.g., 67MB used) per message processing module. The determination of resource usage metrics is achieved by applying an algorithm to internally exposed performance counters, such as at least one of an OS-level utilization polling algorithm, a real-time telemetry aggregation algorithm, or a threshold-based event detection algorithm. The algorithm is configured to detect peaks, leaks, or utilization trend lines; the output includes alerts or status reports, which can then be used for operator awareness and load balancing. For example, the system controller is configured to issue an alert when processing reaches 75% of the CPU, thereby allowing proactive resource allocation.

[0024] The method further includes the system controller generating dynamic visualizations representing the determined resource usage metrics. The system controller formats the metrics—for example, “CPU: 55%, Memory: 33MB”—into structured data suitable for graphical display, generating time-series charts, utilization heatmaps, or resource dashboards for display within a local or remote management UI. The visualizations are formatted according to standard dimensions such as percentages, bytes, and time (e.g., a chart axis for time relative to RAM usage) to support rapid interpretation by technical operators.

[0025] The method further includes a performance dashboard displayed by the system controller based on dynamic visualization data, thereby facilitating operational awareness and capacity management during MQTT message routing. The system controller is configured to update the visualization panel in real time, either on-screen or via a remote network interface, displaying a summary of system status, historical usage trends, and actionable insights. For example, in a pipeline maintenance monitoring center, all edge MQTT devices are displayed on a live dashboard to immediately alert staff to anomalies in device resource usage, supporting uninterrupted monitoring of critical infrastructure.

[0026] This approach allows for the dynamic routing of only subscriber-relevant subsets of MQTT payloads, significantly reducing unnecessary data delivery to each consumer device and optimizing network usage. By receiving bulk payloads at the edge via the MQTT broker processor, the system minimizes upstream communication overhead and centralizes processing at the network edge, resulting in reduced latency and improved responsiveness. Using a payload controller to determine the subscriber-specified subset ensures that each consumer receives exactly the data required for its operation, enhancing data privacy and minimizing post-processing input on the consumer side. Dynamically generating unique MQTT topics for each subset ensures accurate routing and easy management of user-specific payload channels, contributing to scalability and systematic topic organization. Assigning subscriber-specified subsets to these dynamic topics prevents resource waste and duplicate message delivery, ensuring that only targeted data resides in the memory buffer. Outputting filtered subsets to subscribers via these dedicated topics results in a significant reduction in both processing and memory overhead for resource-constrained edge devices, supporting higher device density and more sustainable edge deployments.

[0027] Support for receiving multiple batch payloads from different publishers and enabling payload controllers to combine and extract subsets of combined data allows for the customization of robust, tailored data streams to complex user needs, eliminating the need for clients to subscribe to and filter multiple heterogeneous topics themselves. By aggregating necessary server-side data, the generation and allocation of combined subsets for further dynamically named MQTT topics further simplifies consumer logic, resulting in reduced software complexity and resource utilization on end devices. Outputting these combined subsets to subscribers offloads aggregation and filtering logic from the consumer device, thereby reducing codebase complexity, memory requirements, and the likelihood of processing errors during message subscription and filtering.

[0028] Synchronization between the MQTT broker processor and the payload controller allows for precise allocation of memory resources based solely on the needs of a subset of subscribers, minimizing unnecessary duplicate storage and avoiding memory bloat. This tightly coupled data exchange ensures no redundant allocation, reduces memory consumption, and maximizes the efficiency of edge devices, which is crucial in embedded or constrained environments.

[0029] The inclusion of system controllers that continuously track processor utilization and memory consumption during MQTT processing provides essential operational visibility, allowing operators to quickly identify bottlenecks, predict resource exhaustion, and make informed scaling decisions. Dynamic visualization and dashboards of these metrics simplify capacity management and enhance service reliability, thereby facilitating proactive tuning and troubleshooting in rapidly evolving industrial or IoT edge deployments.

[0030] Edge device hardware allocation of dedicated network interface queues, processor time slices, and memory buffers ensures that MQTT message traffic is processed with predictable throughput and low latency, avoiding contention with other system processes. This fine-grained resource allocation delivers higher system throughput, reduces jitter, and ensures message delivery deadlines are met even under varying system loads, thus supporting deterministic real-time operation of mission-critical applications.

[0031] The purpose of this disclosure is further achieved through a system for dynamically routing multiple subsets of MQTT message payloads to at least one consumer device based on individual subscription requirements.

[0032] The system includes an MQTT broker processor configured to receive at least one batch payload published to a first MQTT topic from at least one publisher, the batch payload comprising multiple data points. The definitions, formats, and examples of the MQTT broker processor, batch payload, and MQTT topic are as described above.

[0033] The system further includes a payload controller configured to receive multiple batch payloads from multiple publishers. The payload controller is configured to manage multiple input MQTT message notifications from an MQTT broker processor, catalog multiple publisher sources, and prepare data for downstream subset extraction.

[0034] The system further includes a synchronization mechanism configured to synchronize the fetching of at least one subscriber-specified subset between the MQTT broker processor and the payload controller. The synchronization mechanism is defined as a logical or physical protocol, such as signaling semaphores, mutexes, shared memory segments, or event-based triggers, for coordinating concurrent and conflict-free subset fetching, memory allocation, and message switching between the MQTT broker processor and the payload controller. For example, a shared circular buffer with a mutex lock supports synchronized fetching and efficient buffer reuse.

[0035] The system further includes a payload controller integrated with the MQTT broker processor and configured to determine at least one subscriber-specified subset from multiple data points based on explicit subscription requests from at least one consumer device. This integration occurs via a local API connection, a shared configuration table, or a plug-in interface, thereby supporting real-time, low-latency responses to changes in subscriber requests or newly received payloads.

[0036] The system further includes a payload controller configured to determine, from at least two of a plurality of batch payloads, a combined data point requested by at least one consumer device to generate a combined subset. This configuration enables the payload controller to support multi-topic or multi-field aggregation schemes, thereby facilitating deployment in advanced analytics or control applications by collecting and assembling multiple data points from entirely heterogeneous sources.

[0037] The system further includes a payload controller configured to generate dynamically named MQTT topics mapped to at least a subset specified by a subscriber, according to a predetermined naming convention. Details of the dynamic topic generation and naming convention are as previously described.

[0038] The system further includes a payload controller configured to assign the determined subset, specified by at least one subscriber, to a dynamically named MQTT topic for further routing. The payload controller is configured to reflect this assignment in an internal routing register and mark the referenced payload for delivery in the next publishing cycle.

[0039] The system further includes an MQTT broker processor configured to output a determined subset specified by at least one subscriber to at least one consumer device via dynamically named MQTT topics, thereby reducing resource utilization at the edge hardware. This configuration ensures that only relevant payloads can traverse the limited-capacity network and hardware layers, thus maintaining optimal system usage and performance.

[0040] The system further includes a payload controller configured to assign a combined subset to a further dynamically named MQTT topic. This assignment is recorded in a payload routing registry and used to direct the combined payload to consumer endpoints with matching interests.

[0041] The system further includes a payload controller configured to output a combined subset to at least one consumer device, thereby reducing the processing burden on the consumer device. For example, in a factory-wide predictive maintenance application, a single subscription is sufficient for the client to receive integrated, periodic statuses from all systems, eliminating the complexity of merging and correlating messages from multiple publishers on the device side.

[0042] The system further includes minimizing redundant data storage and memory usage. Through subset buffering and efficient memory management, duplicate payload buffers are avoided, and per-consumer memory requirements scale linearly with the subscriber count, thus avoiding the overhead associated with traditional per-subscriber message replication practices.

[0043] The system further includes edge device hardware configured to allocate multiple network interface queues dedicated to MQTT protocol traffic. This arrangement enables quality-of-service (QoS)-aware routing, prioritized channel allocation, and traffic isolation, for example, reserving low-latency queues for security-critical telemetry traffic and default queues for non-critical updates.

[0044] The system further includes edge device hardware configured to allocate processing time slices to at least the MQTT broker processor and the payload controller. Process times are prioritized according to system policies to ensure that end-to-end message flow avoids bottlenecks and supports continuous, reliable operation, even during peak processor load times.

[0045] The system further includes edge device hardware configured to allocate memory resources for temporary MQTT message storage, thereby maintaining system throughput and responsiveness. By dynamically tuning the memory budget of each routing component, the system prevents message queue crashes and fidelity loss during anticipated demand surges.

[0046] The system further includes a system controller within the edge device hardware, configured to determine at least the processor usage and memory consumption associated with MQTT message payload routing. The system controller monitors all internal messaging components and displays standardized resource metrics, enabling early detection of processing or memory saturation.

[0047] The system further includes a system controller configured to generate dynamic visualizations of data. This system controller processes real-time monitoring metrics to generate summary tables, charts, and display widgets that visualize resource usage trends and detailed component status information.

[0048] The system further includes a system controller configured to display a resource dashboard for operational visibility. This resource dashboard delivers context awareness and supports dynamic resource scaling decisions, thereby reducing system downtime and accelerating problem resolution. The availability of live operational insights for all actors and processes is an advantage not available in legacy MQTT solutions that lack embedded resource monitoring and dashboard capabilities.

[0049] The system is further configured to allow the payload controller to dynamically divide batch payloads into subsets that strictly conform to the explicit needs of each subscriber, generate and register new MQTT topics following structured naming conventions for these subsets, and orchestrate real-time delivery directly to consumer devices. For example, in smart agriculture deployments, each sensor client receives only the required environmental parameters, such as soil moisture or temperature, which saves memory and network bandwidth and thus increases the number of edge devices supported by the infrastructure compared to traditional batch routing.

[0050] It is to be clearly understood that, although in some embodiments the payload controller (PaCo) may be deployed within or co-located with the edge device hardware for local processing purposes, this disclosure is not limited thereto. In alternative embodiments, the payload controller may be implemented on a remote computing infrastructure external to the edge device hardware, such as on a centralized web server, within a cloud-based service, or as a distributed application spanning multiple compute nodes. In such a configuration, the payload controller is operatively connected to the MQTT broker processor and / or other relevant system components via a communication network and is capable of receiving, analyzing, and segmenting MQTT batch payloads routed through or accessible from the edge network. This flexible deployment model allows the payload controller to adapt to a wide variety of system architectures, from fully embedded industrial edge solutions to hybrid and cloud-centric IoT deployments, ensuring that dynamic routing, payload extraction, and subscriber-specific message delivery functions can be performed independently of the physical location of the edge hardware. Therefore, all methods and systems described herein are intended to include implementations in which the payload controller logic or functionality is executed outside of or separate from the resource-constrained edge device hardware, provided that appropriate connectivity and access to the relevant payload data streams are maintained. Attached Figure Description

[0051] The invention will be further described below with reference to the embodiments shown in the accompanying drawings, in which: Figure 1 This is a block diagram illustrating a system for dynamic MQTT payload subset routing implemented within edge device hardware 102 according to an embodiment of the present disclosure.

[0052] Figure 2 The illustration depicts exemplary sequence diagrams according to embodiments of the present disclosure, which detail the data and control message flows between multiple publishers, an MQTT broker processor, a payload controller, and multiple consumer devices; and Figure 3This is a flowchart illustrating an exemplary method for dynamically routing subsets of message queue telemetry transport (MQTT) message payloads to at least one consumer device based on individual subscription requirements, according to embodiments of the present disclosure.

[0053] Various embodiments are described with reference to the accompanying drawings, wherein the same reference numerals are used to refer to the same elements throughout the drawings. In the following description, numerous specific details are set forth for purposes of explanation in order to provide a thorough understanding of one or more embodiments. It will be apparent that such embodiments may be practiced without these specific details. Detailed Implementation

[0054] Figure 1 This is a block diagram illustrating a system 100 for dynamic MQTT payload subset routing implemented within edge device hardware 102 according to an embodiment of the present disclosure. Figure 1 In this configuration, edge device hardware 102 is communicatively connected to publisher 110 and multiple consumer devices, including a first consumer device 108A and a second consumer device 108B. Edge device hardware 102 is implemented as a physical platform, such as an embedded gateway or industrial edge server located near the sensor data source, configured to perform real-time data processing and efficient payload delivery within industrial, smart building, or other IoT deployments.

[0055] Edge device hardware 102 includes an MQTT broker processor 104, which is responsible for receiving bulk MQTT message payloads from one or more publishers, such as publisher 110. The MQTT broker processor 104 listens for incoming MQTT PUBLISH messages on predefined topics, manages network connectivity, and initially stores or buffers the incoming bulk data. For example, in a smart manufacturing scenario, the MQTT broker processor 104 might receive structured environmental sensor readings—such as temperature, pressure, and humidity—in a single bulk message on a topic such as “environment / data.”

[0056] Edge device hardware 102 may include a processor for executing instructions and managing device operations, and a memory unit for storing data and executable code. However, according to another embodiment, edge device hardware 102 may be implemented with more or fewer components than those illustrated in the components diagram.

[0057] A payload controller 106, also located within the edge device hardware 102, works in conjunction with the MQTT broker processor 104 to analyze the input payload and determine the specific subset of data required by individual consumer devices. The payload controller 106 interprets explicit subscription requests from consumer devices, parses the input batch payload, and applies algorithms to extract only the data points specified by each subscriber. For example, if consumer device 108A subscribes only to temperature updates while device 108B subscribes to humidity, the payload controller 106 accordingly partitions the batch payload and prepares targeted subset messages for each device.

[0058] Edge device hardware 102 further includes a synchronization mechanism 112 to coordinate data flow and memory allocation between the MQTT broker processor 104 and the payload controller 106. The synchronization mechanism 112 may include shared memory segments, semaphores, or event-based triggers to ensure mutual buffer access, data retrieval integrity, and real-time memory optimization. For example, it may employ a shared semaphore to signal when new payload data is available for partitioning, while preventing race conditions during concurrent access.

[0059] Furthermore, system controller 114 is integrated within edge device hardware 102 to provide real-time monitoring and visualization of resource utilization, such as processor load and memory consumption, during message processing operations. System controller 114 collects operational metrics from all internal messaging components and generates visualized data for performance dashboards. For example, it can display CPU and RAM usage charts that allow operators or remote administrators to identify overload conditions, thus facilitating rapid troubleshooting and capacity planning.

[0060] Publisher 110 is an external device or data source configured to send structured MQTT PUBLISH messages containing sensor data or telemetry information to edge device hardware 102. In a real-world scenario, publisher 110 might be a multi-sensor device in an industrial plant sending periodic status updates. First consumer device 108A and second consumer device 108B represent end-user devices or downstream systems subscribing to dynamically generated MQTT topics corresponding to their explicit data interests. For example, 108A could be a device monitoring interface that only requires voltage data, while 108B might be an air quality monitoring device that only subscribes to ambient humidity readings.

[0061] All these components, both inside and outside the edge device hardware 102, collaborate to achieve efficient, subscription-specific routing to a subset of MQTT payloads. This system ensures that each consumer device receives only the data it needs to operate, thereby reducing network and memory overhead and improving scalability in resource-constrained edge environments.

[0062] MQTT broker processor 104 is configured to receive bulk payloads published to MQTT topics from at least one publisher 110. The bulk payload comprises multiple data points. MQTT broker processor 104 is defined as a computational entity responsible for receiving, parsing, and initially storing multiple MQTT PUBLISH messages arriving from multiple publishers 110 to multiple topics. An example of MQTT broker processor 104 is a Mosquitto broker instance running on edge device hardware 102 in a smart manufacturing facility, processing multiple environmental sensor data streams. An MQTT topic is defined as a logical channel within the MQTT protocol used for message exchange. For example, an MQTT topic could be named "Environment / Data". A bulk payload is a message object encapsulating a grouped set of data points formatted using a serialization standard such as JSON, and includes multiple key-value pairs, such as temperature: 23.5°C, pressure: 1.01 atm, and voltage: 220V. Reception of bulk payloads published to a first MQTT topic from at least one publisher 110 occurs via the MQTT CONNECT and PUBLISH protocol mechanisms, whereby the MQTT broker processor 104 is configured to listen for incoming PUBLISH packets on a predefined TCP port, such as 1883. Multiple data points are measured sensor values, where each data point can be individually labeled or encapsulated as a field in a JSON or equivalent serialized structure. The MQTT broker processor 104 is configured to parse the incoming bulk payloads, verify message integrity, and temporarily buffer the multiple data points in memory for further processing.

[0063] Payload controller 106 is configured to receive multiple batch payloads from multiple publishers 110, each batch payload being published to at least one MQTT topic. Payload controller 106 is defined as a programmable component, software module, or hardware integration logic explicitly delegated to analyze multiple input MQTT messages, determine consumer demand, and orchestrate payload partitioning. A real-world example of payload controller 106 is a PaCo (Payload Controller) plugin deployed within a Mosquitto agent, which examines multiple sensor data received from multiple weather stations. Multiple publishers 110 are a group of devices or software agents configured to push multiple sensor readings—such as temperature, humidity, and pressure sensors—each sensor reading sending structured MQTT messages to multiple specific topics, such as “sensor / temperature,” “sensor / humidity,” and “sensor / pressure.” The payload controller 106 is configured to receive multiple batch payloads via protocol-level interception or callback registration, such that when each MQTT PUBLISH packet arrives at the MQTT broker processor 104, a notification or reference to the payload is forwarded for routing considerations by the payload controller.

[0064] Synchronization mechanism 112 is configured to synchronize the extraction of at least one subscriber-specified subset from the received batch payload between MQTT broker processor 104 and payload controller 106, such that memory resources are allocated for the at least one subscriber-specified subset. Synchronization mechanism 112 achieves synchronization through explicit state tracking and mutual memory buffer access protocols between MQTT broker processor 104 and payload controller 106. Synchronization mechanism 112 may include a shared memory segment, message queue, or remote procedure call (RPC) interface, ensuring that the exact timing and content of subset extraction are coordinated between MQTT broker processor 104 and payload controller 106. Extraction of the at least one subscriber-specified subset is achieved by applying an algorithm, such as a filter matching algorithm, a subscription-based parsing algorithm, or a pattern matching algorithm, to the buffered batch payload data. This algorithm is configured to compare the fields required by the consumer with the input payload data points and select the desired subset, thereby ensuring that memory allocation is proportional only to the minimum required data points, rather than the entire original batch payload. This process enables the payload controller 106 and the MQTT broker processor 104 to allocate memory resources only for the requested subset of data, thus preventing memory waste caused by unnecessarily retaining unwanted data. In one example, the MQTT broker processor 104 and the payload controller 106 use a shared semaphore variable to coordinate concurrent access to the internal payload buffer, ensuring that fetching is triggered only after the subscriber specification is updated.

[0065] The payload controller 106, integrated with the MQTT broker processor 104, is configured to determine at least one subscriber-specified subset from multiple data points based on explicit subscription requests from at least one consumer device 108A, 108B. The payload controller 106 achieves this determination by processing subscriber registration information, where each subscription specifies specific fields or data points of interest, such as "temperature" and "humidity" only from environmental data payloads. The payload controller 106 is configured to parse the explicit subscription requests from consumer devices 108A, 108B, match the desired fields within the input batch payloads, and select only those data points to include in the subscriber-specified subset. The determination step is implemented by applying an algorithm, such as a rule engine algorithm, a policy-based subscription filtering algorithm, or a user profile matching algorithm, to the subscription registry and payload metadata. This algorithm is configured to identify the intersection between the requested fields and the available data points. For example, if consumer devices 108A, 108B request "humidity," the payload controller 106 scans the input payload against the keyword "humidity" and extracts the corresponding data, such as humidity: 56%. This determination step allows consumer devices 108A and 108B to receive only the minimum necessary data, thereby reducing message processing and storage at the edge device hardware 102 compared to traditional methods where the entire batch payload is routed to each subscriber regardless of their individual needs.

[0066] Payload controller 106 is configured to determine, from at least two of a plurality of received batch payloads, a combined data point requested by at least one consumer device 108A, 108B. This step involves payload controller 106 analyzing subscription descriptors to identify requests where consumer devices 108A, 108B require merging data points from individual publishers 110, such as “temperature” from a weather sensor and “pressure” from a machine sensor. Payload controller 106 determines the combined data point by applying an algorithm, such as a multi-source aggregation algorithm, a joint operation algorithm, or a combined set intersection algorithm, to the input stream of the payload buffer. This algorithm is configured to dynamically map publisher 110 data to any subscriber’s explicit multi-field request. For example, if consumer devices 108A, 108B request a report containing independently published “electricity consumption” and “ambient light”, payload controller 106 merges those values ​​into a single combined subset. This approach simplifies message subscription and filtering complexity at consumer devices 108A, 108B, thereby reducing device-side computational resources required for multi-topic aggregation.

[0067] Payload controller 106 is configured to generate dynamically named MQTT topics associated with a determined subset specified by at least one subscriber, according to a predetermined naming convention. The dynamic naming convention is defined as a string pattern combining the original topic identifier and a suffix or prefix representing a subset or consumer identifier, such as “temperature_subset_userA”. Payload controller 106 is configured to construct such topic names at runtime based on an internal mapping table and naming logic. Examples of naming conventions include “SensorData / Subset / Temp”, “BulkData / Aggregate / Humidity”, or “Device123 / SubsetA”. Using dynamically named MQTT topics aligned with subsets allows for an immediate and visible mapping between message content and topic names, improving scalability and traceability compared to traditional static topic mappings.

[0068] The payload controller 106 is configured to assign a determined subset, specified by at least one subscriber, to a dynamically named MQTT topic for routing to at least one consumer device 108A, 108B. The payload controller 106 creates and maintains a mapping registry that associates subset payload buffers with their respective dynamic topic identifiers. For example, the subset "Temp: 22.4°C" is assigned to the topic "EnvData_Subset_Temp". This assignment allows the payload controller 106 to simplify subsequent subscription management and payload delivery.

[0069] MQTT broker processor 104 is configured to output a determined subset specified by at least one subscriber to at least one consumer device 108A, 108B via dynamically named MQTT topics, thereby reducing memory and processing resource consumption at resource-constrained edge device hardware 102. MQTT broker processor 104 serializes only the relevant subset and publishes it under a constructed dynamic topic, delivering the subset to the network address of the registered subscriber. For example, if "client 456" subscribes to "AirQuality_Subset_PM25", MQTT broker processor 104 outputs only PM25 data to that client, saving bandwidth and memory usage at consumer devices 108A, 108B. This targeted delivery is beneficial for edge device hardware 102 with limited CPU and RAM, enabling support for a larger number of consumer devices 108A, 108B without degradation in throughput or response latency. In smart agriculture deployments, consumer devices 108A assigned for irrigation management receive only a subset of soil moisture values, demonstrating the real-world benefits of reduced device resource consumption.

[0070] Payload controller 106 is configured to generate a composite subset including the determined composite data points. Payload controller 106 listens for multi-topic subscription commands from consumer devices 108A and 108B, buffers batches of payloads from multiple publishing sources, and merges the requested data point fields into an integrated structure. For example, the composite subset could be represented as {"Temperature": 22.1, "Pressure": 1.02, "Humidity": 60%}. Payload controller 106 generates the composite subset by applying at least one of the following algorithms to the collected payloads from different MQTT topics: a buffer merging algorithm, a tuple merging algorithm, or a stream correlation algorithm. The output of the composite subset enables precise and targeted message delivery, thereby freeing consumer devices 108A and 108B from cumbersome context management and message merging operations. For example, environmental analysis device 108A could subscribe to the "CompositeWeatherTopic" once and automatically receive periodic aggregations of all requested weather indicators.

[0071] The payload controller 106 is configured to assign composite subsets to further dynamically named MQTT topics. The payload controller 106 constructs appropriate topic identifiers, such as “Composite_AQ_TempPressHum”, and maps the resulting composite subsets to that topic. The assignment process is registered in an internal topic payload registry, thereby ensuring correct routing and consumer notification for each additional subset.

[0072] Payload controller 106 is configured to output a combined subset to at least one consumer device 108A, 108B, thereby reducing the complexity of message subscription and filtering at consumer devices 108A, 108B. Payload controller 106 serializes the combined subset and signals the MQTT broker processor 104 for delivery, ensuring that receiving devices 108A, 108B receive a minimum amount of context-rich data objects. Consumer devices 108A, 108B only need to process the last completed message. For example, in an industrial automation environment, consumer device 108A subscribes to a single topic and immediately receives a ready-to-use combined subset, improving real-time analytics efficiency compared to subscribing to and merging messages individually.

[0073] The payload controller 106 and MQTT broker processor 104 are configured to minimize redundant data storage and memory usage. The joint operation of the payload controller 106 and MQTT broker processor 104 eliminates unnecessary duplicate buffers, thus maintaining only the necessary subset of the payload in the local memory of the edge device hardware 102. Redundancy minimization is achieved through pointer-based memory management techniques, reference counting, or explicit payload lifecycle control protocols after message delivery. This approach scales efficiently with the number of consumer devices 108A, 108B compared to traditional MQTT brokers that store a complete copy of the message for each subscriber.

[0074] Edge device hardware 102 is configured to allocate network interface queues dedicated to MQTT protocol traffic. Edge device hardware 102 is defined as a physical processor, RAM module, and network interface circuitry embedded within or co-located with the data source. Edge device hardware 102 establishes dedicated queues, such as network buffers exclusively mapped to TCP port 1883, to provide priority for MQTT packet reception and resource-isolated transmission paths. Real-world examples of edge device hardware 102 include a Raspberry Pi with an Ethernet interface in a smart grid monitoring station, or a custom embedded controller in a home automation system.

[0075] Edge device hardware 102 is configured to allocate processor time slices to at least MQTT broker processor 104 and payload controller 106 for message processing. This step involves configuring scheduler-level settings, such as Linux cgroups or RTOS scheduler policies, to establish a minimum CPU time allocation for each internal component, thereby ensuring predictable real-time message routing and payload manipulation, even amidst transient system loads from other operations. Processor time slicing ensures that critical routines in MQTT broker processor 104 and payload controller 106 remain unsaturated, which is necessary for sustained low-latency operation. For example, in a health monitoring device with multiple internal subsystems, the processor reserves a fixed period for MQTT traffic management to prevent message delivery stalls or delays.

[0076] Edge device hardware 102 is configured to allocate storage buffers for temporary storage of MQTT payload data, thereby maintaining system throughput and latency. Edge device hardware 102 instances create storage buffers within RAM, where the block size is dynamically set based on the expected payload volume. Each temporary buffer retains either the entire payload or a subset of the payload before routing, supporting high-throughput data bursts and mitigating traffic surges. The buffer management strategy enables direct post-delivery switching and rapid memory reclamation, improving system responsiveness and helping to avoid bottlenecks typical of general-purpose buffers. For example, in a hospital patient monitoring network, the allocation of dedicated MQTT payload storage buffers by edge device hardware 102 prevents the loss of vital signs data during peak emergency admission periods.

[0077] System controller 114 within edge device hardware 102 is configured to determine resource usage metrics, including at least processor utilization and memory consumption during MQTT message payload processing. System controller 114 is defined as an embedded monitoring agent or programmable coprocessor responsible for tracking time-averaged and instantaneous system resource loads, such as CPU load (e.g., 43%) and RAM usage (e.g., 67MB used) for each message processing module. System controller 114 determines resource usage metrics by applying at least one of an algorithm, such as an OS-level utilization polling algorithm, a real-time telemetry aggregation algorithm, or a threshold-based event detection algorithm, to internally exposed performance counters. The algorithm is configured to detect peaks, leaks, or utilization trend lines; the output includes warnings or status reports, which can then be used for operator awareness and load balancing. For example, system controller 114 is configured to issue an alarm if processing reaches 75% of CPU usage, thereby allowing proactive resource allocation.

[0078] System controller 114 is configured to generate dynamic visualizations of identified resource usage metrics. System controller 114 formats metrics—such as “CPU: 55%, Memory: 33MB”—into structured data suitable for graphical display and generates time-series charts, utilization heatmaps, or resource dashboards for display within local or remote management UIs. The visualizations are formatted according to standard dimensions such as percentages, bytes, and time (e.g., a chart axis for time relative to RAM usage) to support rapid interpretation by technical operators.

[0079] System controller 114 is configured to display a performance dashboard based on dynamic visualization data, thereby facilitating operational awareness and capacity management during MQTT message routing. System controller 114 updates the visualization panel in real time on-screen or via a remote network interface, displaying a summary of system status, historical usage trends, and actionable insights. For example, in a pipeline maintenance monitoring center, all edge device hardware 102 MQTT devices are displayed on a live dashboard to immediately alert staff to anomalies in device resource usage, supporting uninterrupted monitoring of critical infrastructure.

[0080] Figure 2 The illustration depicts an exemplary sequence diagram according to an embodiment of the present disclosure, detailing the data and control message flow between multiple publishers 110, an MQTT broker processor 104, a payload controller 106, and multiple consumer devices 108A and 108B. The diagram illustrates a mechanism by which batches of sensor payloads are dynamically partitioned, routed, and individually delivered to those consumers who have explicitly requested a specific subset, all within a resource-constrained edge device environment as described in this disclosure.

[0081] During the initialization phase, publisher 110 is configured to publish a bulk payload, i.e., embedded data points {a, b, c, d}—representing aggregated sensor readings—to Topic_X via the MQTT PUBLISH protocol routed over TCP (e.g., port 1883). MQTT broker processor 104 is configured to receive the PUBLISH message, parse the encapsulated JSON or equivalent serialized payload, and trigger an internal notification to payload controller 106, signaling to Topic_X that a new payload with data points {a, b, c, d} has been received. According to this disclosure, this notification step is configured via protocol-level callback registration or direct messaging to ensure immediate awareness of new data arrival and to establish a temporary in-memory buffer for the payload to support downstream operations.

[0082] During the dynamic configuration phase, this disclosure implements subscriber-specific customization of data extraction. Consumer device 108A is configured to send a request to payload controller 106, specifying only a specific subset—data points {a, b}—by providing a formal subscription requirement (e.g., “temperature, humidity” in a larger set of environmental data). Payload controller 106 is configured to act upon this request and determine an appropriate dynamically named topic (Topic_A) based on an internal naming convention that uniquely identifies both the source topic and the subset requested by the consumer. Payload controller 106 is configured to respond to 108A with the assigned topic mapping and is configured to subscribe to Topic_A on behalf of consumer 108A, essentially orchestrating future routes for only the {a, b} subset to 108A. In parallel interaction, consumer device 108B is configured to transmit its request for data points {c, d} to payload controller 106, which is configured to respond with a customized mapping to Topic_B and establish a subscription to Topic_B on behalf of 108B. This process makes full use of the programmable logic of the payload controller 106 to maintain a topic-payload registry and maps a subset of the resident memory payload buffers to dynamic topic identifiers, as specified in this disclosure.

[0083] During the message routing phase, when a new batch payload or update arrives, the MQTT broker processor 104, cooperating with the payload controller 106, is configured to extract only a subscriber-specific subset from the buffered payload. The payload controller 106 is configured to apply a filtering matching algorithm—such as a subscription-based parsing algorithm or a pattern matching algorithm—to the payload registry, extract data points {a, b}, and instruct the MQTT broker processor 104 to publish only that subset to Topic_A. Data points {c, d} are similarly partitioned and published to Topic_B of 108B. Steps 11 and 12 further illustrate the continuous nature of the process, where new or updated subsets are repeatedly extracted and published to their corresponding dynamic topics. This ensures that each consumer is configured to receive only the minimum, necessary data, rather than the entire batch payload, based on their explicit subscription request, illustrating the resource-saving advantages described in detail in this disclosure.

[0084] Throughout all operations, synchronization mechanism 112 is configured to proactively synchronize state and memory buffer accesses between MQTT broker processor 104 and payload controller 106. In one example, synchronization mechanism 112 is configured to achieve this through a shared memory region, message queue, or semaphore-protected buffer access protocol, as described in this disclosure, ensuring the integrity of payload retrieval and consistent resource allocation.

[0085] The coordinated action of the payload controller 106 and the MQTT broker processor 104 is configured to implement a system in which memory and processor time on the edge device hardware 102 are precisely allocated—only for the requested subset of data, not for redundant or unused information. This dynamic, per-subscriber data partitioning model replaces traditional static topic routing and extensive payload delivery mechanisms, providing a scalable solution for environments such as smart manufacturing, industrial automation, or IoT deployments where edge resources are very limited and efficiency is critical.

[0086] Furthermore, edge device hardware 102 is configured to manage network interface queue allocation and processor time sharding, isolate MQTT communication paths, and guarantee predictable message delivery even under high load. Temporary memory buffers are dynamically allocated proportionally to the expected subset size and released immediately after delivery, thereby maximizing throughput and minimizing latency. In one example, edge device hardware 102 is configured to allocate a memory buffer for temporary storage of MQTT payload data, thus maintaining system throughput and latency.

[0087] System controller 114 is configured to continuously track and visualize resource utilization metrics (CPU and RAM) for each processing component. In one example, system controller 114 is configured to generate dynamic visualizations representing the identified resource utilization metrics. In one example, system controller 114 is configured to display a performance dashboard based on the dynamic visualizations, thereby facilitating runtime awareness and capacity management during MQTT message routing. In one example, system controller 114 is configured to issue an alert if processing reaches a threshold, such as 75% CPU, thereby allowing proactive resource allocation. In one example, system controller 114 is configured to update the visualization panel in real time on-screen or via a remote network interface, showing a summary of system status, historical usage trends, and actionable insights.

[0088] Figure 3 This is a flowchart illustrating an exemplary method for dynamically routing subsets of multiple Message Queuing Telemetry Transport (MQTT) message payloads to at least one consumer device based on individual subscription requirements, according to embodiments of the present disclosure.

[0089] At step 302, the MQTT broker processor receives a batch payload comprising multiple data points published from at least one publisher to a specific MQTT topic. The batch payload is typically formatted as a serialized message object, such as a JSON structure containing multiple key-value pairs representing sensor or device data.

[0090] At step 304, the payload controller, integrated with the MQTT broker processor, determines at least one subset of subscriber-specified data points from a plurality of data points. This determination is performed based on explicit subscription requests solicited from at least one consumer device. The payload controller applies a subscriber preference matching algorithm to parse the subscription requests and extracts only the required data point fields for each consumer.

[0091] At step 306, the payload controller generates dynamically named MQTT topics for a subset of the identified subscribers according to a predetermined naming convention. The naming convention can combine the original topic identifier, payload field descriptor, or subscriber-specific identifier to ensure unique and discoverable topic tags.

[0092] At step 308, the payload controller assigns the determined subset of subscriber-specified data to dynamically named MQTT topics, thereby establishing associations in the internal payload-topic mapping registry. This enables targeted routing of only relevant subsets of data points to each interested consumer device.

[0093] At step 310, the MQTT broker processor outputs the determined subset of subscribers to at least one consumer device via dynamically named MQTT topics. This targeted data delivery minimizes memory and processing resource consumption at resource-constrained edge devices because only the necessary subset—rather than the entire batch payload—is delivered to each subscriber.

[0094] This disclosure may take the form of a computer program product including program modules, such as instructions implementing an MQTT agent processor (104) and a payload controller (106), which are accessible from a computer-usable or computer-readable medium storing program code used or in conjunction with one or more computers, processors, or instruction execution systems within a resource-constrained edge device environment (102). For the purposes of this description, a computer-usable or computer-readable medium includes any means capable of containing, storing, communicating, propagating, or transmitting a program used by or in conjunction with an instruction execution system, apparatus, or device. The medium may be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or apparatus or device) and includes non-transitory physical computer-readable media such as semiconductor or solid-state memory, magnetic tape, removable computer disks, random access memory (RAM), read-only memory (ROM), hard disks, and optical disks (such as compact disc read-only memory (CD-ROM), compact disc read / write, and DVD). As those skilled in the art will recognize, the processing hardware and logic modules used to implement each aspect of this technology (including bulk MQTT message processing, subscriber-specified data extraction, and dynamic topic assignment) can be centralized within a single edge node or distributed across multiple components (or combinations thereof).

[0095] While this disclosure has been described in detail with respect to certain embodiments, it should be understood that this disclosure is not limited to those embodiments. In view of this disclosure, modifications and variations can be conceived by those skilled in the art without departing from the scope of the various embodiments described herein. The scope of this disclosure is defined by the following claims rather than the foregoing description, and all changes, modifications, and variations within the meaning and scope of equivalents of the claims are considered to be within its scope. All advantageous embodiments claimed in the method claims also apply to the system and apparatus claims.

[0096] The described disclosure offers significant advantages over traditional MQTT messaging and edge data processing methods. By enabling the MQTT broker processor (104) and payload controller (106) to dynamically extract and route only subscriber-specific subsets of data points, the system dramatically reduces unnecessary memory consumption and processing overhead, which is particularly beneficial for resource-constrained edge device hardware (102). This selective routing ensures that consumer devices (108A, 108B) receive only relevant information, thereby increasing data throughput and minimizing network congestion. This disclosure also supports scalable topic management via dynamic naming conventions, allowing seamless adaptation to new device and subscription demands without manual reconfiguration. Furthermore, real-time resource monitoring and intelligent buffering ensure predictable performance even under varying loads, making this solution particularly well-suited for industrial IoT (IIoT), smart manufacturing, and mission-critical automation environments where efficiency, reliability, and low latency are critical.

Claims

1. A computer-implemented method for dynamically routing subsets of message queue telemetry transport (MQTT) message payloads to at least one consumer device (108A, 108B) based on individual subscription requirements, comprising: The MQTT broker processor (104) receives (s302) a bulk payload published to an MQTT topic from at least one publisher (110), the bulk payload comprising multiple data points; The payload controller (106), integrated with the MQTT broker processor (104), determines (s304) at least one subset specified by a subscriber from the plurality of data points based on the explicit subscription requirements of the at least one consumer device (108A, 108B); The payload controller (106) generates (s306) dynamically named MQTT topics associated with a determined subset specified by at least one subscriber according to a predetermined naming convention; The payload controller (106) assigns (s308) the determined subset specified by at least one subscriber to the dynamically named MQTT topic for routing to the at least one consumer device (108A, 108B). The subset specified by at least one subscriber is determined by the output (s310) of the MQTT broker processor (104) to the at least one consumer device (108A, 108B) via the dynamically named MQTT topic, thereby reducing the consumption of memory and processing resources at the resource-constrained edge device (102); The payload controller (106) receives multiple batch payloads from multiple publishers (110), each batch payload being published to at least one MQTT topic; The payload controller (106) determines, from at least two of the received multiple batch payloads, a combined data point requested by at least one consumer device (108A, 108B); The payload controller (106) generates a combined subset by incorporating a corresponding subset of data points extracted from each of the at least two batch payloads into an integrated structure, the combined subset thereby comprising a combination of multiple subsets, each subset corresponding to a subset of data points from a corresponding MQTT topic; The payload controller (106) assigns the combined subset to further dynamically named MQTT topics; and The payload controller (106) outputs the combined subset to the at least one consumer device (108A, 108B), thereby reducing the complexity of message subscription and filtering at the at least one consumer device.

2. The method according to claim 1, further comprising: Between the MQTT broker processor (104) and the payload controller (106), the extraction of a subset of the at least one subscriber-specified data from the received batch payloads is synchronized, such that memory resources are allocated for the subset of the at least one subscriber-specified data, thereby minimizing redundant data storage and memory usage based on synchronization.

3. The method according to claim 1, further comprising: Resource usage metrics are determined by the system controller (114) within the edge device (102), which include at least processor utilization and memory consumption during MQTT message payload processing; The system controller (114) generates dynamic visualization data representing the determined resource usage indicators; The system controller (114) displays a performance dashboard based on the dynamic visualization data, thereby facilitating operational awareness and capacity management during MQTT message routing.

4. The method according to claim 1, further comprising: The edge device hardware (102) allocates a network interface queue dedicated to MQTT protocol traffic; The edge device hardware (102) allocates at least a processor time slice to the MQTT broker processor (104) and the payload controller (106) for message processing; A memory buffer allocated by the edge device hardware (102) for temporary storage of MQTT payload data ensures that system throughput and latency are maintained.

5. A system for dynamically routing subsets of multiple MQTT message payloads to at least one consumer device (108A, 108B) based on individual subscription requirements, comprising: An MQTT broker processor (104) is configured to receive (s302) at least one batch payload published to an MQTT topic from at least one publisher (110), the batch payload comprising multiple data points; A payload controller (106), which is integrated with the MQTT broker processor (104), is configured to determine (s304) at least one subset specified by a subscriber from the plurality of data points based on the explicit subscription requests of at least one consumer device (108A, 108B); The payload controller (106) is further configured to generate (s306) dynamically named MQTT topics mapped to a subset specified by the at least one subscriber, according to a predetermined naming convention. The payload controller (106) is further configured to assign (s308) the determined subset specified by at least one subscriber to the dynamically named MQTT topic for further routing; The MQTT broker processor (104) is further configured to output (s310) a subset determined by at least one subscriber to the at least one consumer device (108A, 108B) via the dynamically named MQTT topic, thereby reducing resource utilization at the edge hardware (102).

6. The system according to claim 5, wherein, The payload controller (106) is further configured to: Receive multiple batch payloads from multiple publishers (110); Determine a combined data point requested by at least two of the plurality of batch payloads, which is then used to generate a combined subset; The combined subset is assigned to a further dynamically named MQTT topic; as well as The combined subset is output to the at least one consumer device (108A, 108B) to reduce the processing burden on the consumer device.

7. The system according to claim 5, further comprising: A synchronization mechanism (116) is configured to synchronize the extraction of a subset specified by the at least one subscriber between the MQTT broker processor (104) and the payload controller (106), thereby minimizing redundant data storage and memory usage.

8. The system according to claim 5, further comprising: The system controller (114) within the edge device hardware (102) is configured to at least determine the processor usage and memory consumption associated with the routing of the MQTT message payload; Generate dynamic visualization data; The resource dashboard is used for runtime visibility.

9. The system according to claim 5, wherein, The edge device hardware (102) is further configured to allocate multiple network interface queues dedicated to MQTT protocol traffic, allocate processing time slices to at least the MQTT broker processor (104) and the payload controller (106), and allocate memory resources for temporary MQTT message storage, thereby maintaining system throughput and responsiveness.