A system for data analysis using local devices and cloud computing platforms

Abstract

The present invention relates to a system for performing data analysis in a network between one or more local devices (130) and a cloud computing platform (120), wherein data collected and / or stored on one or more local devices (130) and / or stored on the cloud computing platform (120) are processed by an analysis algorithm (A), the analysis algorithm being subdivided into at least two sub-algorithms (SA1, SA2), wherein one sub-algorithm (SA1) is executed on the one or more local devices (130) and another sub-algorithm (SA2) is executed on the cloud computing platform (120).

Description

Technical Field

[0001] The present invention mainly relates to a system for performing data analysis in a network between one or more local devices and a cloud computing platform, wherein an analysis algorithm (A) executes requests for processing data collected and / or stored on one or more local devices and / or stored on the cloud computing platform. Background Art

[0002] In a system based on cloud computing technology, a large number of devices are connected to the cloud computing system via the Internet. These devices may be located in a remote facility connected to the cloud computing system. For example, these devices may include or consist of equipment, sensors, actuators, robots and / or machinery in one or more industrial devices. These devices may be medical equipment and equipment in a healthcare unit. These devices may be household appliances or office appliances in a residential / commercial institution.

[0003] Cloud computing systems can enable remote configuration, monitoring, control, and maintenance of connected devices (also often referred to as "assets"). Likewise, cloud computing systems can facilitate storage of large amounts of data regularly collected from devices, analysis of large amounts of data, and provision of insights (e.g., key performance indicators, anomalies) and alerts to operators, field engineers, or device owners via a graphical user interface (e.g., a web application). These insights and alerts can enable control and maintenance of the equipment, thereby enabling efficient and fail-safe operation of the equipment. Cloud computing systems can also allow modification of parameters associated with the equipment and issuance of control commands via a graphical user interface based on insights and alerts.

[0004] A cloud computing system may include multiple servers or processors (also referred to as "cloud infrastructure") that are geographically distributed and interconnected via a network. A dedicated platform (hereinafter referred to as a "cloud computing platform") is installed on the servers / processors to provide the above-mentioned functions as a service (hereinafter referred to as a "cloud service"). The cloud computing platform may include multiple software programs executed on one or more servers or processors of the cloud computing system to implement the services requested by the device and its users.

[0005] One or more application programming interfaces (APIs) are deployed in a cloud computing system to provide various cloud services to, for example, one or more users.

[0006] There is a growing trend of industrial automation systems, assets, machines, sensors, mobile devices, etc. In all areas of industrial production, energy, transportation, and other fields such as banking, retail, hospitality and healthcare systems, they are connected to the Industrial Internet of Things (IIoT) via network connections directly or through cloud gateways. Data analysis (data mining, deep learning, artificial intelligence) is a core aspect of the entire field of connected things and generates new levels of knowledge and usability.

[0007] On the one hand, the collected data can be analyzed at the edge of the IIoT device or IIoT gateway or cloud computing platform as a central computing infrastructure. If the data analysis is performed at the edge of the local IIoT device itself, it can achieve real-time analytical operations. Since the data is executed or processed on the local device, sending the original data to the cloud is avoided, thus preserving network bandwidth and being resistant to network problems. In particular, in order to be able to use cloud applications, it is not necessary to provide a permanent, uninterrupted connection to the Internet, which has high requirements for availability, for example, in the case of devices in the automotive or mobile phone industries.

[0008] If data analysis is performed in a cloud computing platform, it allows one or more IIoT platform providers to better protect and secure the know-how and intellectual property of their analytical algorithms, since there is no need to expose these algorithms to an environment outside the cloud. Since the analytical algorithms are centrally executed in the cloud computing platform, they do not need to be deployed on local edge devices, which usually do not belong to the IIoT platform provider. In addition, other types of algorithms and software applications that require higher computing speeds and more storage space, such as image recognition and analysis in the automotive field or in the scientific or medical fields (scanning microscopy, radio diagnostics, magnetic resonance imaging), can be used on the IIoT platform, resulting in higher quality and faster data analysis.

[0009] This means that there are both advantages and disadvantages to performing data analytics at the edge on local devices or on cloud computing platforms.

[0010] An example of such an implementation is provided in patent application US 2017083312A1, in which a wearable sensing device can be configured to operate in one or more operating modes, for example, depending on the operating mode, each mode can be configured to use one or more sensors to collect and process sensor signals and generate sensor signal data, which can be processed by the wearable sensing device, or transmitted to a smartphone or center for processing, or subsequently transmitted to a cloud platform for processing. Summary of the invention

[0011] In view of the foregoing, it is therefore an object of the present invention to provide techniques that facilitate the use of both local devices' respective edge and cloud computing platforms to improve data analysis and ensure maximum utilization of the entire network.

[0012] According to a first aspect, the present invention provides a system for performing data analysis in a network between one or more local devices and a cloud computing platform, wherein data collected and / or stored on one or more local devices and / or stored on the cloud computing platform is processed by an analysis algorithm, the analysis algorithm being subdivided into at least two sub-algorithms, wherein at least one sub-algorithm is executed on one or more local devices, and at least one other sub-algorithm is executed on the cloud computing platform.

[0013] In some advantageous embodiments, at least one of the at least two sub-algorithms takes as input the result of at least another of the at least two sub-algorithms. In other words, at least one of the at least two sub-algorithms may depend on the result of at least another of the at least two sub-algorithms.

[0014] In other advantageous embodiments, the sub-algorithms work independently of each other, i.e. do not require any results from any other sub-algorithms to perform their intended functions. In this case, advantageously, all sub-algorithms can be executed simultaneously (or in other words, executed in parallel).

[0015] In any case, the (total) analysis algorithm comprises or consists of at least two sub-algorithms. The (total) analysis algorithm can be divided into these two or more sub-algorithms. For example, a user or an administrator can divide the (total) analysis algorithm into two or more sub-algorithms. Thereby, the sub-algorithms therefore serve the same purpose, or in other words, generate the same results as the (total) analysis algorithm, for example, produce the same results. The (total) analysis algorithm is modularized by dividing the analysis algorithm into two or more sub-algorithms. The sub-algorithms can be implemented by one or more software modules, i.e. by a computer, which implements autonomous tasks. The overall task of the (total) analysis algorithm can be implemented by a combination of two or more sub-algorithms.

[0016] For example, embodiments of the present invention may be used to analyze and evaluate vibration and other data in industrial plants, image data in the scientific and medical fields, drug development and clinical trial data using medical devices in the pharmaceutical field, route calculation data in the navigation field, image recognition data in the automotive field and computer games, and the like.

[0017] By decomposing the analytical algorithm into sub-algorithms, a hybrid and distributed analytical approach is provided, which can be executed on the edge of local devices to obtain real-time characteristics and edge advantages, and on cloud computing platforms to obtain cloud advantages.

[0018] Using this hybrid approach for decomposing an analysis algorithm into sub-algorithms, a serial and / or parallel execution order of the sub-algorithms can be defined. Specifically, at least the first of the multiple sub-algorithms can be assigned to an edge (e.g., a local device or gateway) for execution, while at least the second of the multiple sub-algorithms can be assigned to a cloud computing platform for execution. Subsequently, the multiple sub-algorithms can be distributed to their execution locations, i.e., deployed, according to their allocation. The division of the (total) analysis algorithm can be based on the specific application in which the (total) analysis algorithm will be used. For example, when dividing the (total) analysis algorithm into multiple sub-algorithms, the resources available on the local device and / or gateway can be considered. Additionally or alternatively, preference indicators can be determined or assigned to one or more of the (total) analysis algorithm and / or multiple sub-algorithms, as will be described below.

[0019] Furthermore, the analysis algorithm may include at least one preference indicator (or: preference indicator), wherein for at least one sub-algorithm of the analysis algorithm, or the analysis algorithm itself, the preference indicator indicates, which sub-algorithm should be executed in a theoretical manner, i.e., on the cloud computing platform, the gateway and / or one or more local devices. This information may also be specified as a preference / preference for the analysis algorithm or any sub-algorithm, respectively.

[0020] In some embodiments, the preference index of the analysis algorithm and / or at least one sub-algorithm may indicate that a particular sub-algorithm may be executed on a particular device or location. In some embodiments, the preference index of the analysis algorithm and / or at least one sub-algorithm may indicate that a particular sub-algorithm must be executed on a particular device or location.

[0021] Preferably, the analysis algorithm comprises data indicating into which sub-algorithms the analysis algorithm may be subdivided (or: separated) More preferably, for each such sub-algorithm indicated, a preference indicator is provided.

[0022] In some advantageous embodiments, the sub-algorithms can be dynamically moved or moved between the cloud, one or more gateways, and / or one or more local devices based on considerations such as likes / preferences (i.e., based on likes indicators), available resources, and computer performance. Whether a specific sub-algorithm is movable can be indicated by a corresponding mobility indicator of the sub-algorithm or the analysis algorithm itself.

[0023] Additionally, a workflow engine can be created to define sub-algorithms, assign their execution order, and assign likes / preferences for certain types of IIoT devices.

[0024] In another advantageous embodiment, sub-algorithms destined for execution on the local device may be transferred to the gateway.

[0025] According to other advantageous embodiments, the sub-algorithms are executed in a serial and / or parallel sequence.

[0026] The system may also include a software application that controls the workflow regarding the order and location of executing the sub-algorithms.

[0027] The software application may advantageously be stored in an analysis runtime module ARM of one or more local devices and / or one or more gateways and / or cloud computing platforms.

[0028] According to other advantageous embodiments, for example in a preference index, preferences are defined for the execution location and runtime of each sub-algorithm.

[0029] The definition of preferences / preferences (or in other words, preference indicators) may be based on computer performance and resources and / or availability of cloud computing platforms and / or data structures of processed data.

[0030] According to further advantageous embodiments, the analysis algorithm is configured as a clustering and / or a neural network and / or a support vector machine and / or a blockchain, which are subdivided into sub-algorithms.

[0031] In some advantageous embodiments, the processed data is collected from vibration sensors and / or acoustic sensors and / or optical sensors and / or temperature sensors and / or pressure sensors and / or chemical and / or piezoelectric sensors.

[0032] In other advantageous embodiments, a plurality of local devices are connected in a network. The preference indicator may be configured to indicate not only that the sub-algorithm is to be executed on and by the unspecified local device, but also which particular local device among the plurality of local devices the sub-algorithm is to be executed by.

[0033] One or more local devices may be configured as an industrial pump, a medical device, an imaging device, a mobile device, an automotive device, and / or an analytical scientific instrument.

[0034] According to a second aspect, the present invention provides a method for performing data analysis in a network between one or more local devices and a cloud computing platform, wherein data collected or stored on one or more local devices and / or stored on the cloud computing platform is processed by an analysis algorithm (A), and the analysis algorithm (A) is subdivided into at least two sub-algorithms (SA1, SA2), wherein at least one sub-algorithm (SA1) is executed on one or more local devices, and at least one other sub-algorithm (SA2) is executed on the cloud computing platform.

[0035] Additionally, according to a third aspect, the present invention provides a local device configured for use in a system according to the first aspect of the present invention. In particular, the local device can be configured as an industrial pump, a medical device, an imaging device, a mobile device, an automotive device and / or an analytical scientific instrument.

[0036] According to a fourth aspect, the present invention provides a cloud computing platform configured for use in the system according to the first aspect of the present invention.

[0037] According to a fifth aspect, the present invention provides a computer program product comprising executable program code configured to perform the method according to the second aspect when executed.

[0038] According to a sixth aspect, the present invention provides a non-transitory computer-readable data storage medium comprising an executable program code, which is configured to perform the method according to the second aspect when executed. The non-transitory computer-readable data storage medium may include or consist of any type of computer memory, in particular a semiconductor memory.

[0039] According to a seventh aspect, the present invention provides a data stream representing or configured to provide a program code, the program code being configured to perform the method according to the second aspect when executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Additional features, aspects and advantages of the present invention or its embodiments will become apparent from the detailed description read in conjunction with the following drawings:

[0041] Figure 1 A general overview of a system according to an embodiment of the first aspect of the invention is provided;

[0042] Figure 2 A workflow for performing data analysis according to an embodiment of the present invention is shown;

[0043] Figure 3 schematically illustrates a plurality of data analysis modules in a cloud computing platform and an exemplary industrial plant;

[0044] Figure 4 shows a schematic flow chart depicting an embodiment of a method according to an embodiment of the second aspect of the present invention;

[0045] Figure 5 schematically depicts a computer program product according to an embodiment of the fifth aspect of the invention; and

[0046] Figure 6 A non-transitory computer-readable data storage medium according to an embodiment of the sixth aspect of the invention is schematically depicted. DETAILED DESCRIPTION

[0047] In the following description, for purposes of explanation rather than limitation, specific details are set forth in order to provide a thorough understanding of the present invention. It will be apparent to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details.

[0048] Figure 1 A general overview of a system 100 for data analysis in a network of an industrial plant 110 and a cloud computing platform 120 is provided. The industrial plant 110 in this embodiment includes an industrial pump as a local device 130. However, additional local devices 130 may be added to the system 100. An acceleration sensor 140 (accelerometer) is placed on the pump 130 as an example of a mechanical component to capture rotation and vibration data, which aids in early detection of various failure modes encountered in rotating mechanical equipment.

[0049] In a simplified view, an industrial pump 130 has a mechanical system 150 and a controller 160. In the illustrated example, four accelerometers 140 are mounted on the mechanical rotating part 150 of the pump 130 to capture acceleration / vibration measurement data. The sensors 140 are connected to the pump controller 160. The pump controller 160 includes an IIoT agent 161 and is connected to an IIoT gateway 170, which in turn is connected to the IIoT cloud 120 via a network 190.

[0050] The IIoT gateway 170 may include a plurality of IIoT gateway modules 171. The collected and generated data may be stored and processed on the pump controller 160 and / or the IIoT gateway 170 and / or other one or more data storages and / or the IIoT cloud computing platform 120. The stored data may include reconstructed data, compressed data, segmented data, and other data related to the local device 130. The network 190 may include one or more wide area networks (WANs) such as the Internet, a local area network (LAN), or other networks that may facilitate data communications.

[0051] Network 190 may be divided into subnets that allow access to all other components connected to network 190. Network 190 may include encryption schemes that may be used on the public Internet. Network 190 may be configured to support wired and / or wireless configurations for data transmission.

[0052] The representation shown is intended to illustrate one possible configuration of the system 100. Other configurations may include fewer components, and in other configurations, additional assets 180 may be utilized. These changes in configuration and components may increase or alter the performance of the system 100. Particularly in the field of medical imaging devices for capturing high-resolution medical images of patients, such as magnetic resonance imaging (MRI) machines, computed tomography, x-ray, positron emission tomography, scanning microscopy, ultrasound imaging systems, sensors configured to capture image data, for example, sensors such as charge coupled devices (CCDs). Other types of sensors, such as acoustic, optical, piezoelectric, pressure, temperature, chemical sensors may also be used to collect additional data for a specific local device 130.

[0053] Figure 2 is a schematic diagram showing a workflow 200 for performing data processing and analysis on both the local device 130 and the cloud computing platform 120 of the system 100. The workflow 200 may also be performed as part of a method according to an embodiment of the second aspect of the present invention.

[0054] for Figure 1 For the vibration case shown, as an example, the analysis algorithm A is decomposed into the following sub-algorithms:

[0055] A first sub-algorithm 1 (SA1.1 and SA1.2) is generated for upsampling or downsampling the time series vibration data and normalizing the data in a standardized format.

[0056] A second sub-algorithm 2 (SA2) is generated for computing a frequency domain transform or spectrum. An example of such an algorithm is the Fast Fourier Transform (FFT).

[0057] A third sub-algorithm 3 (SA3) is generated for analyzing frequency components in the vibration spectrum to determine failure modes in the mechanical equipment.

[0058] These three sub-algorithms SA1.1+SA1.2, SA2, SA3 can be executed on separate computing units. However, more than three sub-algorithms SAn can be provided for other analysis algorithms A or other data structures. A variety of analysis algorithms A, such as clustering, neural networks, support vector machines and / or blockchains can be used in the framework of the present invention. In addition, for the workflow 200 processing sub-algorithms SA1, SA2, SA3, ... San, a network or tree structure with multiple preference levels that can be executed in a dynamic manner can be created.

[0059] In the illustrated example, it is necessary to execute SA1 in real time, but sub-algorithm SA1 is not very computationally complex and does not require a lot of computer performance / resources. In addition, from the perspective of the provider of analysis algorithm A, sub-algorithm SA1 does not contain specific knowledge or intellectual property. Therefore, its preference index may indicate that sub-algorithm SA1 can (and should) be executed on the local device 130 or pump controller 160.

[0060] However, further, in the illustrated example, SA2 requires more computer performance, but does not include any specific proprietary technology or intellectual property. Therefore, its preference index can indicate that sub-algorithm SA2 can (and should) be executed on the IIoT gateway 170 with more computing resources than the local device 130.

[0061] Further, in the illustrated example, SA3 involves analysis of frequency components, which requires specific domain knowledge, i.e., the know-how and / or intellectual property of one or more analysis providers. It may also involve benchmarking or comparison of frequency components of other similar industrial assets. Therefore, its preference index may indicate that sub-algorithm SA3 can (and should) perform this step in the IIoT cloud computing platform 120.

[0062] The workflow 200 begins with input raw data 210 and ends with the results 220 of one or more analysis algorithms A. The first row 230 lists the various sub-algorithms (SA1, SA2, ... SAn) of the one or more overall analysis algorithms A and the order in which these sub-algorithms are executed. For example, in the depicted example, SA2 cannot be executed until SA1.2 has completed execution because it requires the output of SA1.2 as input.

[0063] It should be understood that in other implementations of the system 100, the sub-algorithms SA1, SA2, ... SAn may also be executed in parallel order, and the order may also be changed dynamically according to the application. In addition, in the depicted example, only one local device 130 is used. However, there may be a network of multiple local devices 130 connected to the cloud computing platform 120 and / or there may be connections between the local devices 130.

[0064] Figure 2 The second row 240 in defines the preference / preference (i.e., the content of the corresponding preference indicator) for executing a particular SA on a specific device 130 and / or gateway 170 and / or cloud 120. A preference for the local device 130 means that the specific SA has a preference for execution on the industrial pump itself (e.g., the controller 160 of the pump 130). Similarly, the SA may have a preference for execution on the IIoT gateway 170 or the IIoT cloud computing platform 120.

[0065] Furthermore, the execution of SA may change in a dynamic manner due to the performance or availability of the IIoT cloud computing platform 120 or other criteria such as runtime or data structure of processed data.

[0066] Figure 2 The third row 250 in the table represents the feature "mobile", which defines whether a particular SA is mobile or can be moved to another possible computing unit, i.e. the content of the optional mobility indicator. If the corresponding SA is mobile, this is indicated by Figure 2 The "+" sign in the ; if the corresponding SA is not movable, this is indicated by Figure 2 For example, if the local device 130 is overloaded or does not have sufficient computing resources, the sub-algorithm SA1.2 can be moved or moved to the IIoT gateway 170. Similarly, if the IIoT gateway 170 is overloaded, SA2 can be moved or moved to the cloud computing platform 120 for execution.

[0067] Figure 3 A schematic diagram of an overall system 100 with modules configured to implement a distributed analysis method is shown.

[0068] Each part of the IIoT system 100, the cloud 120, one or more IIoT gateways 170, and one or more local devices 130, includes an analysis runtime module (ARM) 300. That is, a runtime environment is provided in which one or more different sub-algorithms are run. The ARM 300 executes different sub-algorithms SA1, SA2, and SA3. In the local device 130, the ARM 300 can also be part of the IIoT agent 161. However, other configurations are also possible. The industrial plant 110 and the cloud computing platform 120 can be separated by a firewall 105. Depending on the situation, two or more sub-algorithms can be executed in the runtime environment in the local device, gateway, or cloud computing platform. This may be because the runtime environment is the same or similar in architecture, thereby allowing the sub-algorithms to be deployed in corresponding locations. At this time, if, for example, the availability of resources of the local device and / or gateway changes, such as due to other tasks performed by the local device, it is reduced, then one or more sub-algorithms deployed at the local device and / or gateway can be reallocated to be executed in the cloud, that is, moved to the cloud. Vice versa, when resources are freed up at the local device and / or gateway, one or more sub-algorithms may be moved from the cloud to the local device or gateway.

[0069] The cloud computing platform 120 additionally includes an analysis workflow module (AWM) 310 and an analysis scheduler 320. The AWM 310 manages multiple SAs and provides a method for defining Figure 2The user interface of the SA is shown. In another embodiment, the scheduler 320 can also be a part of the IIoT gateway 170 to provide local scheduling capabilities. In addition, the cloud computing platform 120 includes a database 330, an analysis configuration database 340, a communication module 350 and other IIoT modules 360.

[0070] The steps of how to use the system are described below: An administrator or data scientist can use the user interface provided by the AWM 310 to provide (in particular define) an analysis algorithm A together with different SAs, their execution order, their preference for IIoT devices, etc. Criteria such as computer performance, relevance to know-how and intellectual property are preferably used to define the SAs and their preference for the cloud 120 or local devices 130 (preference indicators) and / or their portability indicators.

[0071] The AWM 310 saves the configuration in the analysis configuration database 340. However, another software application may be implemented in the AWM 310 that dynamically adjusts the execution order of the sub-algorithms based on criteria such as the availability of the cloud computing platform 120. This is particularly useful in situations where the removable local device 130 has limited access to the cloud 120.

[0072] After generating the analysis algorithm A and SA, the administrator triggers the deployment of algorithm A together with SA. This can also be triggered automatically based on some other analysis results or criteria.

[0073] The analysis scheduler 320 reads the analysis configuration from the analysis configuration database 340. It determines other criteria such as resource availability on the IIoT gateway 170 and the local device 130 by querying these entities. Based on the configuration rules and resource availability, the analysis scheduler 320 generates a scheduling plan for the SA.

[0074] According to the scheduling plan, the analysis scheduler 320 provides the SA definition to the analysis runtime module ARM 300 on the cloud computing platform 120, the IIoT gateway 170, or one or more local devices 130 for the execution of SA. By defining multiple sub-algorithms SA1, SA2, ... SAn, a distributed analysis framework can be created. By separating the analysis algorithm A into sub-algorithms SA1, SA2, ... SAn, the analysis data can be executed in the cloud computing platform 120 and / or on one or more local devices 130 based on business and / or performance requirements.

[0075] With this approach, the analytics provider can better preserve her / his domain knowledge or intellectual property when the domain-specific portion of the analytics algorithm A is executed under her / his control in the cloud computing platform 120. This also allows for enhanced data security of the analytics model in an IIoT approach.

[0076] Additionally, since the sub-algorithms SA1, SA2, ... SAn can be dynamically transferred between different local IIoT devices based on resource availability, this improves the performance of the IIoT system and minimizes bottlenecks.

[0077] In particular, distributed analysis methods can be used for Such an open IIoT operating system.

[0078] The above implementation of vibration data shows how to decompose the analysis algorithm A into sub-algorithms SA1, SA2, ... SAn, which can then be executed on different devices in the IIoT environment. Similarly, other analysis algorithms (clustering, neural networks, support vector machines, blockchain, etc.) can be used and decomposed into sub-algorithms SA1, SA2, ... SAn.

[0079] Figure 4 A schematic flow chart depicting a method according to an embodiment of the second aspect of the present invention is shown. Figures 1 to 3 The reference numerals partially describe Figure 4 method, although Figure 4 The method is not limited to Figures 1 to 3 On the other hand, Figure 4 The method can use reference Figures 1 to 3 The method may be performed according to any of the embodiments described, and accordingly, the method may be adapted and modified according to any variations and modifications described in the foregoing.

[0080] In step S10, algorithm A is provided (e.g., defined, received, or stored) in a network between one or more local devices 130 and the cloud computing platform 120, and the analysis algorithm is configured to process data collected and / or stored on one or more local devices 130 and / or stored on the cloud computing platform 120.

[0081] In step S20, the analysis algorithm A is subdivided (or: separated) into at least two sub-algorithms SA1, SA2. Step S20 may be performed simultaneously with step S10. For example, when defining the analysis algorithm, it may be defined together with the sub-algorithms.

[0082] In step S30, at least one of the at least two sub-algorithms SA1, SA2 is executed on (or by) one or more local devices 130. In particular, a local device 130 according to the third aspect of the invention may be provided to perform step S30.

[0083] In step S40, at least another sub-algorithm SA2 of the at least two sub-algorithms is executed on (or by) the cloud computing platform 120. Preferably, all sub-algorithms SA2 that are not executed on the local device 130 are executed on the cloud computing platform 120. In particular, the cloud computing 120 according to the fourth aspect of the present invention may be provided to perform step S40.

[0084] Figure 5 Schematically depicts a computer program product 400 comprising an executable program code 450, which executable program code is configured to perform the method according to the second aspect of the present invention when executed, in particular as described with reference to Figure 3 and / or Figure 4 as described.

[0085] Figure 6 Schematically depicts a non-transitory computer-readable data storage medium 500 comprising an executable program code 550, which executable program code is configured to perform, when executed, a method according to the second aspect of the present invention, in particular as described with reference to Figure 3 and / or Figure 4 as described.

Claims

1. A system for performing data analysis in a network between one or more local devices (130) and a cloud computing platform (120), wherein: The system includes a processor configured to process data collected and / or stored on the one or more local devices (130) and / or stored on the cloud computing platform (120) by an analysis algorithm (A), the analysis algorithm being subdivided into at least two sub-algorithms (SA1, SA2), wherein, based on a liking / preference index, at least one sub-algorithm (SA1) is executed on the one or more local devices (130), and at least another sub-algorithm (SA2) is executed on the cloud computing platform (120), wherein the liking / preference index is defined for an execution location and a runtime of each sub-algorithm based on a data structure of the processed data, wherein the execution location is about allocating the sub-algorithm to its execution location according to a deployment, wherein the at least one sub-algorithm (SA1) designated for execution on the one or more local devices is transferable to a gateway based on the computing resources of the one or more local devices.

2. The system according to claim 1, wherein: The one or more local devices (130) and the cloud computing platform (120) are connected by a gateway (170).

3. The system according to any one of claims 1 to 2, wherein: The sub-algorithms (SA1, SA2, ... SAn) are executed in serial and / or parallel order.

4. The system according to any one of claims 1 to 3, wherein: The workflow (200) regarding the order and location of execution of the sub-algorithms (SA1, SA2, ... SAn) is controlled by a software application.

5. The system according to claim 4, wherein: The software application is stored in the analysis runtime module ARM (300) of the one or more local devices (130) and / or one or more gateways (170) and / or the cloud computing platform (120).

6. The system according to claim 1, wherein: The definition of the liking / preference index is for the runtime of the sub-algorithms (SA1, SA2) and is based on computer performance and resources and / or availability of the cloud computing platform (120).

7. The system according to claim 1, wherein: A mobility indicator is defined for the mobility of the sub-algorithms (SA1, SA2, . . . SAn) and / or the analysis algorithm (A).

8. The system according to any one of claims 1 to 7, wherein: The analysis algorithm A is configured as clustering and / or neural network and / or support vector machine and / or blockchain, and is subdivided into sub-algorithms (SA1, SA2, . . . San).

9. The system according to any one of claims 1 to 8, wherein: Processed data is collected from vibration sensors and / or acoustic sensors and / or optical sensors and / or temperature sensors and / or pressure sensors and / or chemical and / or piezoelectric sensors.

10. The system according to any one of claims 1 to 9, wherein: A plurality of local devices (130) are connected in the network.

11. The system according to any one of claims 1 to 10, wherein: The one or more local devices (130) are configured as industrial pumps, medical devices, imaging devices, mobile devices, automotive devices, and / or analytical scientific instruments.

12. A method for performing data analysis in a network between one or more local devices (130) and a cloud computing platform (120), wherein data collected and / or stored on the one or more local devices (130) and / or stored on the cloud computing platform (120) are processed by an analysis algorithm (A), the analysis algorithm being subdivided into at least two sub-algorithms (SA1, SA2), wherein: Based on a liking / preference indicator, at least one sub-algorithm (SA1) is executed on the one or more local devices (130), and at least one other sub-algorithm (SA2) is executed on the cloud computing platform (120), wherein the liking / preference indicator is defined for an execution location and a runtime of each sub-algorithm based on a data structure of the processed data, wherein the execution location is about allocating the sub-algorithm to its execution location according to a deployment, wherein the at least one sub-algorithm (SA1) designated for execution on the one or more local devices is transferable to a gateway based on the computing resources of the one or more local devices.

13. A local device (130) configured for use in a system according to any one of claims 1 to 11, wherein: One or more of the local devices (130) are configured as industrial pumps, medical devices, imaging devices, mobile devices, automotive devices, and / or analytical scientific instruments.

14. A cloud computing platform (120) configured to be used in the system according to any one of claims 1 to 11.

Images