DE102021202812A1 - Method for operating a communication network and communication network - Google Patents

Abstract

The invention relates to a method for operating a communication network (100) with a plurality of participants (101-109) which are communicatively connected to one another for the exchange of data, the communication network (100) having one or more buffer storage participants (102-105) as participants which each have a memory (102.2-105.2) and an associated configuration table (102.1-105.1), the configuration table (102.1-105.1) storing which data (203) from the relevant intermediate storage device (102- 105) reaching messages (200) are to be cached by this cache participant (102-105) in the associated memory (102.2-105.2), with a cache participant (102) when a message (200) reaches him, the data (203) that this intermediate storage participant is to intermediately store according to the associated configuration table (102.1), intermediately stores this data (203), wherein, when a request (203') for data (203) from a subscriber (108) reaches one of the buffer subscribers (102) in which the requested data (203) is buffered, this buffer subscriber (102) sends the requested data ( 203) to the requesting subscriber (108), and wherein each configuration table (102.1-105.1) is updated if a trigger criterion is present, in particular by a network controller (130).

Description

The present invention relates to a method for operating a communications network with a number of participants who are connected to one another for exchanging data, such a communications network, methods for operating computing systems as participants in such a communications network, and a computing system and a computer program for carrying them out.

Background of the Invention

In modern, especially connected vehicles, real-time and dynamic data exchange between software applications or software components is required, especially in connection with mobility (e.g. rental services, predictive maintenance, blockchain-based payments, etc.), both internally within the vehicle and from and too external. For this purpose, there are concepts such as service-oriented communication and software-defined networking (so-called "software-defined networking", SDN). Existing E/E architectures of the vehicles are revised accordingly. During e.g. communication intermediate applications (so-called "middleware") such as "Scalable Service-Oriented Middleware over IP" (SOME/IP) and "Data Distribution Service" (DDS, this represents a middleware for data-centric communication in highly dynamic distributed systems). allow a certain degree of dynamics in vehicle communication, but they hinder broad acceptance due to the often insufficiently available real-time properties (response times, latency, jitter, etc.) of such intermediate applications.

Disclosure of Invention

According to the invention, a method for operating a communication network, such a communication network, a method for operating computing systems as participants in such a communication network, and a computing system and a computer program for their implementation are proposed with the features of the independent patent claims. Advantageous configurations are the subject of the dependent claims and the following description.

The invention deals with communication networks with several participants (or communication participants), in particular in vehicles (there then in particular also referred to as on-board network or on-board data network), which are communicatively connected to one another (either directly or indirectly via switches, for example). Ethernet, for example, and communication buses such as CAN and LIN, for example, come into consideration as communication media, particularly in vehicles. The communication network can therefore in particular also include a plurality of communication media, including different ones, which are used for communication. It goes without saying that different communication media can be connected to certain participants, all of which can be operated by the participant in question. Various data is exchanged between participants, possibly also to or from outside the vehicle, and in particular via switching participants (such as switches or routers, but possibly also control devices), which only send messages with requests and/or (useful) data pass through.

Temporary storage of data in (vehicle-internal) participants or network elements (such as switches) and the processing of subsequent requests or requests for data access using e.g. the "Information Centric Networking" paradigm (ICN) allows, for example, time-consuming requests from the enquirer to to the source of the requested data and its transmission from the source back to the enquirer. Response times for data requests can thus be shortened, since the data can be provided, for example, by a switch or other participant in the network through which the request runs. It is not necessary for the request to be forwarded to the source of the data. However, in view of the complexity of modern vehicles with sometimes hundreds of electronic components with several thousand signals (such as sensor values, actuator inputs, etc.), random caching of data on participants such as switches is rather inefficient and impractical.

Against this background, a particularly software-defined approach is proposed to manage or configure network-internal storage (“caches” or so-called “key value stores”, i.e. storage for (only) specific data or values) in order to enable communication between applications to support to achieve a faster data distribution and at the same time to use the memory in the network elements or participants efficiently. For this purpose, in particular, programmable network technologies such as the already mentioned software-defined network and, for example, the programming language P4 can be used in order to create such network-internal key value memories. P4 is a programming language for controlling packet forwarding layers in network devices such as routers and switches. Unlike a general-purpose language like C or Python, P4 is a domain-specific language with a set of constructs optimized for processing and forwarding network data. Each of these memories (or key value memories) can be configured via a standardized control interface (e.g. the gRPC protocol for P4), for example, in order to be able to determine which data (e.g. a set of measurement data from specific sensors) must be cached. This is to be explained below in particular using an exemplary system.

Various use cases ranging from emissions testing to diagnostics, for example, require extracting vehicle data from the vehicle. The so-called on-board diagnosis (OBD) already offers a standardized mechanism for reading out a predefined standardized set of vehicle-internal data (e.g. certain sensor values, parameters, etc.) via the so-called OBD connection in the vehicle. The extraction of any data that may not even have been defined during the design of the vehicle or its E/E architecture is not possible via OBD.

The so-called service-oriented architecture for a communication network is a communication paradigm in which software components (in participants) offer "services" or services that can be subscribed to by the other components (in then usually other participants) that use these services require. For example, with the help of service or feature recognition mechanisms (which can be provided in a centralized or distributed manner in the communication network), the software components, which are usually loosely coupled, can collect information about the available or offered, and required and sought-after services. For example, a participant informs all other participants (or the other software components there) about his offer to provide data from a sensor, for example. The subscribers who need this data can then subscribe to the corresponding service. The data transport from the source (i.e. the offering participant with the sensor or its data) to the recipients (i.e. the participants who have subscribed to the service) can, for example, function like a publish-subscribe communication model. The data is usually transported in a message. One (or more) transfers from one communication medium to another may also be necessary.

However, a subscriber (or software application thereon) can also (actively) request the service using a request-response communication model. In response to this request (which is usually also contained in a message), the source then sends the requested data to the requesting party. While the participants who have subscribed to the service then regularly receive the latest data from the source (e.g. the most recent measurement data from the sensor), the (actively) requesting participant only receives the data (the measurement data from the sensor) once when required, i.e. if he makes a corresponding request.

This architecture leads to a relatively loose coupling between the software applications and participants and thus to a flexible architecture that, for example, allows easy reuse of software. The SOME/IP and DDS mentioned at the beginning are popular examples of such service-oriented communication frameworks in the automotive sector.

There are various challenges when introducing such an architecture, especially in a vehicle. Modern vehicles are highly complex and have numerous components, in particular sensors and actuators, that supply and/or require data. The large number of car models and their countless variants further increase the complexity. There is no standardized mechanism for naming the required data or signals (e.g. a particle sensor value) and the corresponding services. In addition, the location of the data (e.g. the electronic control unit, ECU connected to the sensor providing the data) and the means of extraction (e.g. via CAN or Ethernet) differ for each vehicle model or every vehicle variant. To this end, standardization organizations (e.g. Genivi, W3C) are working on creating standardized data models for naming vehicle data. For example, the VSS (Genivi "Vehicle Signal Specification") data model describes in-vehicle data as a hierarchical tree that provides default paths to specific data points within the tree. A provision of QoS guarantees (QoS stands for “Quality of Service”) for data distribution via services is also desirable. Such QoS mechanisms can be integrated into the network layer, e.g. via "Time-Sensitive Networking" (TSN) in order to equip Ethernet with real-time extensions. These extensions can be configured based on the QoS levels for the services as needed and requested by the software applications.

Irrespective of this, the present invention can be used not only in a data-oriented but also in a service-oriented architecture of a communication network, as will be explained in more detail below.

Modern switches and network interface cards (so-called NICs) can largely implement static forwarding processes without software intervention (e.g. in the form of ASICs or FPGAs). The P4 programming language already mentioned aims, for example, at programming the packet forwarding process (so-called data plane) of network elements. P4 can be used to define the various fields that are of interest from a packet processing point of view and perform the processing steps based on entries in a configuration table or look-up table. In addition, these configuration tables can be programmed or configured at runtime by a network controller (remote), for example. A switch programmed in this way (with P4) processes an input based on the loaded P4 program and the contents of the configuration table entries inserted via the control interface.

In particular, P4 programs enable smaller calculations without additional latency in this context, i.e. network elements or participants could process the packet fields (e.g. the headers in received messages) and insert the results with the required data rate into the outgoing packets or messages. This programming language also makes it possible, for example, to use the memory in the network elements or participants to save a limited status. This can e.g. be used to create a key-value store in network elements, as mentioned above.

Although efforts are being made to provide QoS guarantees for service-oriented communication, the increasing number of services offered and subscribed to, particularly in communication networks in vehicles, is ultimately likely to overwhelm the entire network.

The present invention therefore proposes the use of network-internal, in particular software-defined, memories (or caches) by participants in the communication network for in particular vehicle-internal data in order to make the data available to other participants or their software applications that request it. For this purpose, programmable network technologies, such as those based in particular on the programming language P4 mentioned, can be used to create a distributed network-internal memory (as already mentioned, e.g. key value memory) at least in certain participants or network elements, in particular in switching units such as switches, gateways, etc. Valid requests or inquiries for (vehicle-internal) data can then be resolved in the network, which leads to shorter response times and at the same time reduces the computing load for other participants such as vehicle-related computing units.

In summary, one or more of the participants in the communication network each have a memory and an associated configuration table, which stores which data from the messages reaching the relevant participant should be buffered by this participant in the associated memory. Each of these participants with a memory (also referred to as a buffer participant) stores this data in its memory when it receives a message that contains data that this participant is to buffer according to the associated configuration table. If a request for data from a (different) participant reaches one of the intermediate storage participants in which the requested data is temporarily stored, then this intermediate storage participant transmits the requested data to the requesting participant. The cache participants then no longer forward this request to other participants (including other cache participants).

A special aspect is that the memories or key value memories are dynamically configured in such a way that they store the “correct” information in their (limited) memory for participants such as vehicle components. In other words, each configuration table is updated when a triggering criterion is present (and thus in particular when required), specifically during runtime in particular.

A particularly important finding is that (in-vehicle) data traverses the network from the appropriate service provider (i.e., an offering party, the source of the data) to a specific service consumer (another party, e.g., a requesting party or subscriber), which in certain participants , especially in forwarding network elements (ie Subscribers who only forward messages or data, for example the switching units mentioned) can be temporarily stored in order to serve subsequent data requests, ie requests for data. Instead of introducing dedicated hardware to store data (such as in data centers) which would increase the complexity of in-vehicle systems, and instead of trying to randomly cache each packet of data and eventually remove potentially important data, one can dictate what data is required from which participant or network element must be cached. This can be done via a (logically central) network controller that configures the configuration tables of the participants accordingly. This network controller is also a participant in the communication network and is preferably provided by a control device or a computing system with high computing power, such as a central vehicle computer (this can be achieved, for example, by a specific application or software on it).

With the P4 programming language, it is possible to create memory areas within participants or network elements for storing data, e.g. directly in the switching ASIC. It should be noted that storing data is only an additional task and the forwarding of data to the other subscribers or consumers continues uninterrupted.

With several hundred signals available in each subnet (e.g. CAN domain, Ethernet connection etc.) it is usually not possible to store all data in one device such as a switch. This is where the network controller comes in, telling the relevant participants what exact data items to store and how. For this purpose, the network controller programs or configures the configuration tables of the relevant participants, which are then used to extract the required data fields from certain messages or packets (e.g. Ethernet packets, CAN frames or another communication protocol). For this purpose, the network controller can, for example, send configuration data to the intermediate storage device in question, which receives it. The extracted data or values are stored in memory, e.g. indexed by keys that are also available in the configuration tables. For example, instead of using arbitrarily long strings (like naming the data) and using memory to store the keys themselves, hash values can be used as fixed-length keys.

When applications on participants request or request specific data, the cache participants can intercept these requests or requests and check in their memories whether the requested data is already cached. For example, a hash value of the character string representing the name of the data is then calculated for this purpose and the hash value is used as a key for looking up in memory. If a stored value is available, the participant in question can respond with the data directly to the requesting participant without involving the source of the data (i.e. the participant who originally provided the data) (this is also the sense in which the term "intercept ' to be understood, since the request does not have to be forwarded at all). For example, if the data is not available, the request is forwarded to the data source as usual. However, it is conceivable that another intermediate storage participant has also temporarily stored the requested data, which can answer the request when it receives it.

For example, the participants can be supplemented by mechanisms that allow data in memory to be marked as obsolete in order to avoid the transmission of obsolete information or data in response to a request. In general, it is expedient if the subscriber only transmits the requested data to the requesting subscriber if a condition for the relevant data being up-to-date is met. Such a timeliness condition may require, for example, that the data is not older than a certain period of time. This is relevant, for example, for time-critical sensor data. For other data, however, this may be different. In general, such up-to-dateness conditions can therefore be or can be specified individually for the relevant data.

In particular, the network controller uses its global knowledge of the network architecture (e.g. topology, semantics of CAN signals, etc.) together with e.g. traffic statistics for data traffic in the network to configure or program the configuration tables in the participants concerned. For this purpose, the network controller analyzes in particular the data traffic in the communication network, in particular also with regard to changes. If a change occurs, it can be determined, for example, that said triggering criterion is met, i.e. the configuration tables can be updated.

By gathering information about the data that is frequently requested or requested, the network controller then makes decisions about two aspects of the system in particular. On the one hand, which data should or must be cached, and on the other hand, which participants should or must cache them. For this purpose, the network controller adapts the configuration tables accordingly and transmits them to the participants concerned. Thus, the memory can be optimally configured to the requirements of the applications and the network or the participants with the applications. In a simple embodiment, the network controller can, for example (when required or continuously) determine the most frequently required data and select this for temporary storage. If the order or the frequency of this data changes, this can be a trigger criterion for a new configuration.

Overall, a hardware-accelerated memory in a subscriber or network element reduces the load on the data source by decoupling the generation and provision of the data. Depending on the communication requirements, the data can also be made available to other requesting participants or other downstream participants with memory. By reducing the number of hops in the communication network from participant to participant to answer data requests (as far as possible only the participants with memory, the actual sources of the data then no longer) such a communication network can provide improved response times and thus better real-time properties. In particular, this improvement can be achieved independently of the actual E/E architecture (e.g. zone architecture, domain architecture, etc.) since the memories are an essential part of the overall communication network.

The invention further relates to such a communication network, a method for operating a computing system as a buffer participant in such a communication network and a method for operating a computing system as a network controller in such a communication network. With regard to further preferred configurations and advantages, in order to avoid repetition, reference is made to the above statements, which apply accordingly here.

A computing system according to the invention, e.g. a control unit or switch of a motor vehicle, is set up, in particular in terms of programming, to carry out a method according to the invention.

The implementation of a method according to the invention in the form of a computer program or computer program product with program code for carrying out all method steps is advantageous because this causes particularly low costs, especially if an executing control unit is also used for other tasks and is therefore available anyway. Suitable data carriers for providing the computer program are, in particular, magnetic, optical and electrical memories, such as hard drives, flash memories, EEPROMs, DVDs, etc. It is also possible to download a program via computer networks (Internet, intranet, etc.).

Further advantages and refinements of the invention result from the description and the attached drawing.

The invention is shown schematically in the drawing using an exemplary embodiment and is described below with reference to the drawing.

character list

• 1 shows schematically a communication network according to the invention in a preferred embodiment.
• 2 shows part of a sequence of a method according to the invention in a preferred embodiment.

embodiment(s) of the invention

In 1 a communication network 100 according to the invention is shown schematically in a preferred embodiment. This is, for example, an on-board network or on-board data network of a vehicle. The communication network includes, for example, a number of participants 101 to 109 which are communicatively connected to one another for the purpose of transmitting data.

The subscriber 101 is, for example, what is known as a central vehicle computer, a powerful computing system or control device in a vehicle on which all relevant (software) applications in particular run. The participants 106, 107 and 108 are, for example, control devices to which sensors 120, 121 or actuators 125, 126, 127 are connected and which are read out or controlled by the relevant control device. For example, sensors 120, 121 or actuators 125, 126, 127 are connected via CAN as communication medium 111 to the relevant control device. LIN or an analog connection would also be conceivable here as a communication medium, for example, as well as various communication media or types of connection of the sensors or actuators.

The subscribers 102, 103, 104, 105 are, for example, switches or switching units that are used to connect the subscribers or control devices 106, 107, 108 to the subscriber or central vehicle computer 101 via Ethernet, for example, as the communication medium 110.

The subscriber 109 is, for example, a communication module for wireless communication with, for example, a computing unit external to the vehicle, e.g. via mobile radio or WLAN. This participant or this communication module 109 is also integrated into the communication network 100 via Ethernet 110, as a result of which access to the communication network from outside or communication to the outside is made possible.

By way of example, the four subscribers 102 to 105 used as switches are now buffer subscribers within the meaning of the invention, each of which has a memory 102.2, 103.2, 104.2 or 105.2 and a configuration table 102.1, 103.1, 104.1 or 105.1. The participant or central vehicle computer 101 should serve as a network controller 130 within the meaning of the invention, i.e., for example, include a corresponding application or software that carries out the necessary actions.

A sequence of a method according to the invention in a preferred embodiment is to be explained in more detail below on the basis of this communication network 100 .

As already mentioned, it is possible with the programming language P4, for example, to create memory areas within participants or network elements for storing data, e.g. directly in the switching ASIC. In this way, the memory areas 102.2, 103.2, 104.2 and 105.2 can be created and configured accordingly.

As mentioned, with a large number of available signals or types of messages in each subnet (e.g. CAN 111 etc., also in each CAN or Ethernet connection(s) 110 shown), it is usually not possible to transfer all data in one or each of the participants 102 to 105 to store. This is where the network controller 130 comes into play by informing the relevant participants 102 to 105 which exact data or data elements are to be stored and how. For this purpose, the network controller 130 programs or configures the configuration tables 102.1, 103.1, 104.1 or 105.1 of the relevant participants 102 to 105, which are then used to extract the required data fields from certain messages or packets (e.g. Ethernet packets , CAN frames or another communication protocol for e.g. the automotive industry).

The extracted data or values are stored in the memory indexed by keys, for example, which are also available in the configuration tables 102.1, 103.1, 104.1 or 105.1. An example is in 2 For this purpose, a message 200 shown in the communication network from one of the participants, z. B. the controller 106 is sent. This message 200 includes, for example, five components 201 to 205, which relate, for example, to a header with information about the sender, recipient and type of user data. In the configuration table 102.1 it can now be specified, for example, that the subscriber 102, when he receives the message 200 with a specific header, should extract the component 203 as data from it and store it temporarily in his memory 102.2. Irrespective of this, the subscriber 102 also forwards the message 200 to its recipient, eg control unit 108 . In contrast to subscriber or switch 102, switches 103 and 105 do not store any data of this message, since this is not configured by 130 in configuration table 103.1 and 105.1. Such a configuration table is shown below as an example. Match Fields (MAC Addr., Ethertype, etc.) DataType/Length Bytes/Bits representing Signal of Interest key Src MAC: 12:34:56:78:90 Dst MAC: 09:87:65:43:21 VLAN.ID: 4... Uint32 8-11 payload bytes 0x45a23

The first entry indicates or implies that the node for Ethernet messages or Ethernet packets with fields that match the first column (e.g. the MAC address is specified there), the payload bytes 8 to 11 (as the data to be extracted, third column) should be interpreted as a uint32 value (second column for data type) and should be cached in memory under the key (fourth column, key) 0x45a23.

When applications on participants request or request specific data, the cache participants can intercept these requests or requests and check their memories to see whether the requested data is already cached. For example, a hash value of the string representing the name of the data is computed and the hash value is used as a key for lookup in memory.

If a stored value is available, that participant can reply with the data directly to the requesting participant without involving the source of the data (i.e. the participant who originally provided the data). For example, if the data is not available, the request is forwarded to the data source as usual.

The network controller 130 uses e.g. its global knowledge about the communication network 100 (e.g. topology, semantics of CAN signals etc.) together with e.g.

With the 1 this should be explained with an example. There should be a service 250 (on the subscriber 106) that sends data 203 (e.g. within a message 200, cf. 2 ) distributed by sensor 120 (connected to subscriber 106). The applications 251 and 252 (on the subscriber 107 and 101 respectively) have subscribed to this data 203 and receive regular updates from the service 250.

The network controller 130 can now configure or program the memory 102.2 on the subscriber 102 via the configuration table 102.1 in such a way that data 203 from the sensor 120 is temporarily stored there (indicated by a circle 203), as in relation to FIG 2 explained. The network controller 130 uses its global knowledge to make a prediction as to whether or that data 203 will be required by a control unit 108 in the future. Then requests or requirements for this data 203, such as the request 203 'from the application 253 (on the participant 108) can be served directly from the participant 102 with the data 203 without this application 253 having to turn to the participant 106, the the service 250 provides. In addition, queries or requests for data 203 from sensor 120 from external applications (eg a cloud or other non-vehicle computing systems) that arrive via communication module 109 can also be served directly by subscriber 102 with data 203 .

Overall, such a memory in a participant reduces the load on the data source (participant 106 is this source in the example) by decoupling the generation (by participant 106) and consumption (by participant 108) of the data. Temporal decoupling is also achieved, with a temporary storage participant decoupling the frequency or frequency of the transmission of the participant to be sent from the frequency of the requests from participants, thereby reducing the overall load on the communication network. Depending on the communication requirements, the data can also be made available to other requesting participants or other downstream participants with memory.

Claims (18)

Method for operating a communication network (100) with a plurality of participants (101-109) which are communicatively connected to one another for the exchange of data, the communication network (100) comprising one or more buffer participants (102-105) as participants, each have a memory (102.2-105.2) and an associated configuration table (102.1-105.1), the configuration table (102.1-105.1) storing which data (203) from messages reaching the relevant intermediate storage device (102-105). (200) are to be buffered by this buffering participant (102-105) in the associated memory (102.2-105.2), each buffering participant (102) when a message (200) containing data (203) reaches it , Which this cache participant according to the associated configuration table (102.1) between is intended to cache, this data (203) caches, wherein when a request (203') for data (203) from a subscriber (108) reaches one of the cache subscribers (102) in which the requested data (203) are cached , this intermediate storage participant (102) transmits the requested data (203) to the requesting participant (108), and wherein each configuration table (102.1-105.1) is updated when a triggering criterion is present. procedure after claim 1 , wherein the intermediate storage participant (102) transmits the requested data (203) to the requesting participant (108) only if a topicality condition is met for the relevant data. procedure after claim 1 or 2 , wherein the communication network (100) comprises a network controller (130) as a participant through which each configuration table (102.1-105.1) is updated when the triggering criterion is present. procedure after claim 3 , wherein data traffic in the communication network (100) is analyzed by the network controller (130), and based on the analysis it is determined whether the triggering criterion is present, at which the configuration tables (102.1-105.1) are configured and/or updated. Method according to one of the preceding claims, in which the one or at least one of the plurality of buffering parties (102-105) is a switching unit in the communication network. Method according to one of the preceding claims, wherein each cache participant (102-105) caches data using an indexed key, in particular a hash value, in its memory (102.2-105.2) and/or reads it out of its memory in order to convert the data (203) to the requesting participant. Method according to one of the preceding claims, wherein a data-oriented and/or a service-oriented communication paradigm is used. Method according to one of the preceding claims, wherein the request (203') is forwarded by the caching party concerned if the requested data is not cached at the caching party (102-105). A method according to any one of the preceding claims, wherein each caching agent (102) is operable to send messages (200) containing data (203) to be cached by that caching agent and requests (203') for data ( 203) which this intermediate storage participant (102) is to transmit, at different frequencies. Method according to one of the preceding claims, wherein an on-board data network of a vehicle is used as the communication network (100), in which the participants have control units (106-108) connected to the sensors and/or actuators (120, 121, 125, 126, 127). are, switching units (102-105) and in particular a central vehicle computer (101). Communications network (100) with a plurality of participants (101-109) which are communicatively connected to one another for exchanging data, the communications network (100) comprising one or more buffer storage participants (102-105) as participants, each of which has a memory (102.2 -105.2) and an associated configuration table (102.1-105.1), the configuration table (102.1-105.1) storing which data (203) from messages (200) reaching the relevant intermediate storage device (102-105) through this intermediate storage device (102-105) is to be temporarily stored in the associated memory (102.2-105.2), each intermediate storage device being set up to do so when a message (200) containing data (203) containing this intermediate memory reaches it -Participants according to the associated configuration table (102.1) is to cache this data (203), with each cache-participant wide rhin is set up to do so when a request (203') for data (203) from a participant (108) reaches one of the intermediate storage participants (102) in which the requested data (203) are temporarily stored, this intermediate storage participant ( 102) to transmit the requested data (203) to the requesting participant (108), and wherein the communications network (100) is set up to update each configuration table (102.1-105.1) when a trigger criterion is present. communication network (100) according to claim 11 , which further comprises a network controller (130) as a participant, which is set up to update each configuration table (configuration table (102.1-105.1) when the triggering criterion is present. Communication network (100) according to one of claim 11 or 12 , which is used to carry out a method according to one of Claims 1 until 10 is set up. Method for operating a computing system (101) as a participant in a communication network (100) with a plurality of participants (101-109) who are communicatively connected to one another for the exchange of data, wherein the communication network (100) comprises one or more intermediate storage participants (102-105) as participants, each having a memory (102.2-105.2) and an associated configuration table (102.1-105.1), wherein the configuration table (102.1-105.1) it is stored in each case which data (203) from messages (200) arriving from the relevant intermediate storage participant (102-105) are to be temporarily stored by this intermediate storage participant (102-105) in the associated memory (102.2-105.2), data traffic in the communication network (100) being analyzed by the computing system (101), in particular by means of a network controller (130) running in the computing system (101), it being determined on the basis of the analysis whether there is a triggering criterion in which one or several of the configuration tables (102.1-105.1) are to be configured and/or updated and configuration data for updating the one or more configuration tables are output. Method for operating a computing system (102) as a participant in a communication network (100) with a plurality of participants (102-109) who are communicatively connected to one another for the exchange of data, wherein the computing system (102) has a memory (102.2) and an associated configuration table (102.1), the configuration table storing which data (203) from messages reaching the computing system (102) are temporarily stored by the computing system in the associated memory must, wherein by the computer system (102), when it receives a message (200) containing data that this computer system is to cache according to the configuration table (102.1), caches this data, wherein by the computer system (102), when it receives a request (203') for data (203) from a participant in which the requested data are temporarily stored, the requested data is transmitted to the requesting participant, and wherein configuration data is received from the computing system (102) based on which the configuration table (102.1) is updated. Computing system (101, 102), which is set up to follow all the steps of a method Claim 14 or 15 to perform. Computer program that causes a computing system (101, 102) to follow all the steps of a method Claim 14 or 15 to be performed when executed on the computing system (101, 102). Machine-readable storage medium with a computer program stored on it Claim 17 .

Images