Dynamic provisioning of context-dependent, compute-based functionalities in mobile, distributed edge clouds

Description

FIELD OF THE INVENTION

[0001] The invention relates to an automated method, in particular a computer-implemented method, and a system for providing functionalities of a mobile unit for at least one function of a demand unit required remotely. The mobile unit is a mobile edge node of an edge cloud, and the demand unit is part of the edge cloud. BACKGROUND OF THE INVENTION

[0002] New business models and new developments in technology are leading to an increasing demand for computing capacity and functionality in the spatial proximity of generated sensor data or actuators in the field. This is due not only to new technical possibilities but also to technical limitations, such as the inability to transmit large amounts of data.

[0003] Typical architectures today, for example, upload all data collected by sensors to a cloud application, where the data is processed and the derived actions are passed on to the systems in the field. However, due to the huge amounts of data and prevailing bandwidth limitations or other application requirements (e.g., low latency), this is sometimes practically impossible, or simply not practical, or not feasible with current technical capabilities.

[0004] Typical application examples in the smart city / mobility environment are autonomous vehicles, where a communication infrastructure might still be possible but is of little use. These vehicles generate approximately 4,000 GB of data per day, which is primarily used for locally relevant decision-making.

[0005] Transferring this amount of data to cloud applications makes little sense, is costly, and may also violate other requirements, such as latency.

[0006] Another example is applications in an agricultural or offshore maritime environment. New data-intensive applications with data analytics, artificial intelligence, or autonomous functions can no longer be implemented with hierarchical, static cloud architectures due to a lack of communication infrastructure.

[0007] The areas of work that address these issues of providing computing resources in the field are edge and fog computing. Typically, solutions are proposed today in which the applications to be executed and the execution of the environments are planned during the system design, and if necessary, computing resources are provided in the form of additional edge nodes as needed. "Edge nodes" are devices with computing power, storage, and communication capabilities that are typically located between the end devices and a cloud and can run any application—similar to a data center.

[0008] US2019318271 A1 shows a system connected to a live environment executing a current algorithm, comprising: an algorithm data store containing information about a pool of available algorithms and a deployment platform implemented in an edge area at an industrial site connected to the real environment and coupled to the algorithm data store.

[0009] US 2019 / 065264 A1 shows a system for dynamically configuring a process associated with an application based on one or more environmental characteristics, the system comprising:a database system implemented using a server system having one or more processors, the database system configurable to process context data provided by a mobile device on which the application is usable, the context data indicating one or more environmental characteristics captured using the mobile device, the one or more environmental characteristics being associated with a physical environment in which the mobile device is or has been located.

[0010] However, this does not take into account the potential mobility of edge resources, for example vehicles with computing power such as cars, tractors, ships, airplanes, etc., nor the potential dynamics of tasks that are to be performed by the emerging dynamic edge cloud, e.g. new data collection at selected locations, communication tasks, analysis tasks, etc. SUMMARY OF THE INVENTION

[0011] The object of the invention is to provide a solution to make mobile, unused computing resources accessible to a system in a simple and cost-effective manner.

[0012] The invention is based on the features of the independent claims. Advantageous developments and refinements are the subject of the dependent claims. Further features, possible applications, and advantages of the invention will become apparent from the following description.

[0013] One aspect of the invention is to utilize the potential mobility of units, such as edge nodes, to make this functionality available to other system components based on prediction mechanisms through the (technically) cost-effective, redundant provision of rules and corresponding function templates. This makes it possible to make mobile, unused computing resources dynamically and context-dependently accessible to a system and, for example, to use them effectively for data-intensive functions in mobile edge-cloud use cases. This also enables applications that are not possible or only possible to a limited extent according to the state of the art due to communication and cost restrictions.

[0014] A similar dynamic, context-dependent solution is unknown. Current approaches solve the problem by maintaining sufficient local computing resources or providing communication connections, but these remain unused for a large portion of their operating time. Many desirable new mobile edge cloud applications are not feasible with current concepts.

[0015] A cloud-based or edge cloud-based IoT application consists of multiple virtual and physical computing units, data sources, actuators, functions, services, etc., which implement one or more cross-system applications through communication. In today's IoT systems, each subfunction (e.g., service application or analysis function) is typically dedicatedly assigned to a component within the system at design time. In the system according to the invention, such functions or subfunctions are dynamically requested at runtime by system components, for example, an agricultural operation with its farm management system on a farmer's smartphone, and then executed in a context-specific manner. The executing edge nodes are potentially mobile, and the context (especially spatial) is therefore not fixed and usually not deterministic.

[0016] In the system according to the invention, functional requirements consist of one or more rules of a context rule set (e.g., "Mobile edge node in area x,y" and "It is raining simultaneously in area x,y"), a description of the edge node properties, and the corresponding task or function in the form of, for example, a script to be executed, a function container, an application, or a network configuration. This tuple of function and rules is introduced into the system and transferred to mobile units, such as edge nodes, that are considered for this task. Mobile units are described by the functionality they provide, the hardware and operating system properties, and non-functional properties, such as security features (IT security), costs, owner, current location, etc.

[0017] The allocation of functions can be done according to fixed programmed processes or by a central or decentralized prediction component that predicts the future contexts for the system and therefore determines with sufficient reliability a set of potential mobile units on which the functions or tasks could potentially be executed in the future.

[0018] When a function or task is assigned to a mobile device, it is not executed immediately. Instead, the rules are installed on this mobile device in a rule engine via a rule management system, and the function is stored on the mobile device in a function repository or similar.

[0019] If the mobile unit, for example, the edge node, moves into the corresponding context, or if the mobile unit is already in the corresponding context, triggering the rule set for a specified task, the corresponding function is instantiated and executed on the mobile unit. Once the function / task is completed, the result is transmitted to the corresponding requestor (= demand unit) according to specified rules.

[0020] A further aspect of the invention is to maintain the rules and functions on multiple mobile units simultaneously based on the corresponding predictions, in order to achieve a sufficient probability that the task or function will be executed through such redundancy. In an extreme case, the task or function is assigned to all mobile units of the system.

[0021] Advantages of such a system with mobile units are: A provision of functions that are otherwise not available locally. A dynamic utilization of unused resources on, for example, mobile edge and IoT devices. The provision of functionalities is reduced to a memory space for storing the functionality and the rules that lead to the execution of the functions. This does not incur any overhead on the system (e.g., energy consumption due to function execution) and other functions executed on the system are not affected. A context-related deployment mechanism for functions and software components that may only have low-bandwidth communication connections. The autonomy of a device, e.g., in the event of a communication failure, is increased.

[0022] The invention claims an automated, in particular computer-implemented method for providing functionalities of at least one mobile unit for executing at least one function of a demand unit required remotely therefrom, wherein the required function is provided in the mobile unit, wherein the execution of the function is context-dependent, wherein a context rule set is provided in the mobile unit for the execution of the function, and wherein if one or more rules of the context rule set are fulfilled, the function is executed in the mobile unit.

[0023] Context in the IT environment refers to any type of information that can be used to characterize the situation of an entity in interaction with other entities. Since context depends on the specific interaction, there is no general answer to whether a piece of information is context or not. Context can be a single piece of information or a combination of many pieces of information from different sources or at different points in time. Context forms the basis for context-sensitive applications. Context sensitivity, or context-related, refers to the behavior of application programs that use information about their "context," i.e., their environment, to adapt their behavior accordingly.

[0024] Functionalities include, for example, computing capacity or communication capacity. "Function" also includes the processing of a task or a set of tasks. "Function" can also be referred to as an "application program."

[0025] In a further development, the mobile unit can be an edge node of an edge cloud and the demand unit can be a part of the edge cloud.

[0026] In a further embodiment, after the function has been executed, the result can be transmitted to the demand unit according to the context rules. This means that the return transmission to the demand unit is also rule-based.

[0027] In a further embodiment, the context rule set and the at least one function can be redundantly maintained on at least two mobile units based on a prediction model, wherein the prediction model is designed to achieve a predetermined probability of executing the function via the created redundancy.

[0028] In a further embodiment, the mobile unit can be configured to transmit to other mobile units which rules have triggered the mobile unit and which function is being executed.

[0029] In a further embodiment, the demand unit can be configured to transmit the importance and / or urgency of the execution of the function to the mobile unit.

[0030] The invention also claims a system comprising: at least one demand unit, at least one mobile unit which is configured to provide functionalities for at least one function of the demand unit which is required separately from it and to execute them as required, wherein the execution of the function is context-dependent, a rule engine module which is configured in the mobile unit and is configured to execute the function on the basis of a set of context rules stored in the rule engine module, wherein if one or more rules of the context rule set are fulfilled, the function is executed in the mobile unit or the function which has already been executed is stopped.

[0031] In a further training the system has: a function runtime module implemented in the mobile unit that is configured to instantiate the function and execute it in isolation from other functions.

[0032] In a further embodiment, the system has: a context management module trained in the mobile unit that is configured to determine the current context of the mobile unit.

[0033] In a further embodiment, the system comprises: a rule management module trained in the mobile unit that is set up to maintain the context rule set.

[0034] In a further embodiment, the system comprises: a function repository module implemented in the mobile unit that is configured to store the functions that are potentially to be executed and to make them available to the function runtime module.

[0035] In a further embodiment, the system has: a context predictor module trained in the mobile unit that is configured to predict the context of the mobile unit.

[0036] In a further form, the system has: a context sensor module formed in the mobile unit, which is configured to generate information about the context, in particular to collect data from sensors.

[0037] In further training, the system has: a movement, path, coverage and context predictor component that is configured to predict contexts, in particular movement profiles, for the components of the system and to make them available to the mobile unit.

[0038] In the embodiment, the system is designed as an edge cloud and the mobile unit as an edge node.

[0039] Further features and advantages of the invention will become apparent from the following explanations of an embodiment using schematic drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] They show: FIG. 1 shows a block diagram of an embodiment, and FIG. 2 shows a flowchart of a method for context-related provision of computing power. DETAILED DESCRIPTION OF THE INVENTION

[0041] Future IoT systems and edge cloud systems will consist of a multitude of different components. These components will contribute various properties to the system. Components will increasingly be mobile, or computing power will be available on mobile devices. Furthermore, the tasks / functions to be performed, as well as the existing communication infrastructure, may be location-dependent. Manual configuration of such a complex system is very time-consuming and error-prone. To enable reliable operation of such a complex and dynamic system topology, the mechanism according to the invention creates an environment for the automated and traceable planning, distribution, and execution of functions / tasks.

[0042] The core of the invention is the creation of the technical possibility, for example based on learning prediction mechanisms, to provide functions for execution, possibly multiple times and redundantly, together with context rules. These are then only transferred to the execution state when the context conditions are triggered. The context (see details above) can be determined by physical variables, object properties (e.g., geoposition, movement profiles, paths, and coverage of the mobile unit, for example, (edge) cloud or communications infrastructure) and / or environmental properties (e.g., the presence of certain prerequisites such as an external energy source or the geographical proximity of system participants / users).

[0043] The context can also be determined by results from calculations (e.g., predictions based on AI algorithms or models, simulations, emulations, etc.) and can also be defined by multiple dependent or independent parameters. Furthermore, business rules or user preferences can be taken into account.

[0044] FIG. 1 shows a block diagram of a system with an Edge Cloud 5 as an example. The system has the following components: Multiple mobile edge nodes 1, which can potentially perform computing, communication and / or storage functions. These have the following modules, among others: o Context Management Module 1.1; a module for determining the current context of the edge node (e.g. current motion planning) o Rule Engine Module 1.2; this module checks with each context change whether a set of rules for a function is fulfilled, or whether the rules no longer trigger for executed functions. This information is passed on to the Function Runtime Module 1.3 and the corresponding functions are either instantiated or stopped. ∘ Function Runtime Module 1.3; this module instantiates the corresponding functions and executes them in isolation from other functions. ∘ Rule Management Module 1.4; this module maintains the corresponding context rules (a set of rules) that should lead to the execution of functions on the mobile edge node 1. ∘ Function Repository Module 1.5; this module stores all functions that should potentially be executed and makes them available to the other modules, in particular the Function Runtime Module 1.3. ∘ Local Context Predictor Module 1.6; it is a local module for predicting the context on the Edge Node 1. ∘ Context History Module 1.7; a module that stores the history of all relevant data for a context prediction. ∘ Context Sensor Module 1.8; a module that generates the context information, such as data from sensors or functions. A task / function creator component 2 as a demand unit; this system component creates tasks that should be executed in the mobile Edge Cloud 3. For this purpose, in component 2, for example, a task is generated for the system from function templates, the corresponding parameters, and a set of context rules, and the execution of the task is requested in the system. The request can, for example,by a human but also automated by another automation component based on event triggers. An optional function repository component 3; it is / are one or more centrally or decentrally organized storage(s) that make possible functions in the form of, for example, containers, scripts, network functions available to other components, for example as templates. A movement, path, coverage and context predictor component 4; a component that, logically implemented centrally or decentrally, makes context predictions (e.g. movement profiles) of the mobile edge nodes 1 in the system available to the other components. This component 4 can be implemented for each individual mobile edge node 1, for each customer / tenant (= demand unit) in the system, or for the entire ecosystem. .

[0045] In a further variant, a mobile edge node 1 informs the other potential mobile edge nodes 1 which rules have triggered it and which functions it subsequently executes, as well as, if applicable, the corresponding execution status. The other mobile edge nodes 1 that maintain the same function or task can then, for example, interpret this information in such a way that they do not need to execute this function if corresponding rules trigger it, since the function is already being executed by another mobile edge node 1. Alternatively, the function can also be executed redundantly, for example.

[0046] Edge Node 1 can also use power-saving features or reduce the load on main memory when certain applications are not currently needed, for example, by stopping them, swapping them out, or running them at a reduced speed. The response to this situation can also be specified in the rules, for example. Edge Node 1 can also reject a request to execute a function, for example, if the battery level is low.

[0047] In a further development, the importance and urgency of the function execution are included in the task request. These can then be interpreted, for example, using cost functions, and the mobile edge nodes 1 can potentially be encouraged to deliberately create certain contexts or, if necessary, to maintain them for longer (e.g., accepting detours to guarantee the execution of a function). This can also be further developed by projecting these contexts directly onto a business function and, for example, offering them bindingly in an open ecosystem using smart contracts, and documenting the execution in a tamper-proof manner. Blockchain technology, for example, can be used for this purpose.

[0048] Functions for the system include: Data collection functions "Data Mule": If the mobile unit is in area a, then establish a connection to systems of type y via communication interface x, record temperature sensor data, preprocess it according to function b, and upload the resulting data to station c as it passes, which is to forward the data. Deployment functions "Update / Installation Mule": If the mobile unit passes components with the following properties x, y, upload the following update via direct communication and authenticate with certificate z. Communication gateway "Comm Mule": If the mobile unit is expected to be in area a, b for x hours, provide components c, d with a communication link e with service properties f for 60 minutes.Analytics functions "Analytics Mule": If mobile unit is expected to be in area a,b for x hours, collect data from local sensor types c,d for 60 min, analyze the data locally with model e, and send the data via communication link f to g.

[0049] FIG. 2 shows a flowchart of a method according to the invention for providing functionalities of a mobile unit, for example, a mobile edge node of an edge cloud, for a function required by a demand unit remotely. The demand unit can also be referred to as a task / function creator component.

[0050] In a first step 101, the function required by a demand unit is loaded into the mobile unit and thus made available for execution when needed. In a second step 102, a context rule set to be used for context-dependent execution of the function is loaded into the mobile unit and is thus available in the mobile unit. In a third step 103, if one or more rules of the context rule set are fulfilled, the function is executed in the mobile unit. In a final fourth step 104, the result of the execution of the function is transmitted to the demand unit. List of reference symbols

[0051] 1Edge Node 1.1Context Management Module 1.2Rule Engine Module 1.3Function Runtime Module 1.4Rule Management Module 1.5Function Repository Module 1.6Context Predictor Module 1.7Context History Module 1.8Context Sensor Module 2Task / Function Creator Component 3Function Repository Component 4Movement, Path, Coverage and Context Predictor Component 5Edge Cloud 101Loading a function 102Loading a rule set 103Executing the function 104Transferring the results

Claims

1. Automated method for providing functionalities of at least one mobile unit (1) for carrying out at least one required function of a request unit (2) separate from the mobile unit (1), - wherein the required function is provided (101) in the mobile unit (1), - wherein the function is carried out based on the context, - wherein a context rule system is provided (102) in the mobile unit (1) for carrying out the function, - wherein the function is carried out (103) in the mobile unit (1) if one or more rules of the context rule system are satisfied, and - wherein the mobile unit (1) is an edge node of an edge cloud (5) and the request unit (2) is a part of the edge cloud (5).

2. Method according to Claim 1, wherein, once the function has been carried out, the result is transmitted (104) to the request unit (2) according to the context rule system.

3. Method according to one of the preceding claims, wherein the rule system and the at least one function are provided redundantly on at least two mobile units (1) on the basis of a prediction model, wherein the prediction model is designed to achieve a predefined probability of performing the function by means of the provided redundancy.

4. Method according to one of the preceding claims, wherein the mobile unit (1) is configured to inform further mobile units (1) which rules have triggered in the mobile unit (1) and which function is carried out.

5. Method according to one of the preceding claims, wherein the request unit (2) is configured to convey the importance and / or urgency of carrying out the function to the mobile unit (1) .

6. System (5), comprising: - at least one request unit (2), - at least one mobile unit (1) which is configured to provide functionalities for at least one required function of the request unit (2) separate from the mobile unit, and to carry out same as required, wherein the function is carried out based on the context, - a rule engine module (1.2) formed in the mobile unit (1) and configured to carry out the function on the basis of a context rule system stored in the rule engine module (1.2), wherein the function is carried out in the mobile unit (1) or the function already being carried out is stopped if one or more rules of the context rule system are satisfied, characterized in that the system is formed as an edge cloud (5) and the mobile unit is formed as an edge node (1).

7. System according to Claim 6, characterized by: - a function runtime module (1.3) formed in the mobile unit (1) and configured to instantiate the function and carry it out in isolation from further functions.

8. System according to Claim 6 or 7, characterized by: - a context management module (1.1) formed in the mobile unit (1) and configured to determine the current context of the mobile unit (1).

9. System according to one of Claims 6 to 8, characterized by: - a rule management module (1.4) formed in the mobile unit (1) and configured to provide the context rule system.

10. System according to one of Claims 7 to 9, characterized by: - a function repository module (1.5) formed in the mobile unit (1) and configured to store the functions which are potentially intended to be carried out and to make them available to the function runtime module (1.3).

11. System according to one of Claims 6 to 10, characterized by: - a context predictor module (1.6) formed in the mobile unit and configured to predict the context of the mobile unit (1).

12. System according to one of Claims 6 to 11, characterized by: - a context sensor module (1.8) formed in the mobile unit (1) and configured to generate information relating to the context, in particular data from sensors.

13. System according to one of Claims 6 to 12, characterized by: - a movement, path, coverage and context predictor component (4) configured to predict contexts for the components of the system (5), in particular movement profiles, and make them available to the mobile unit (1).