DE102011006490A1 - Computer-based system e.g. mechatronics system, for coordinating access to e.g. printers in distributed network system, has generator for determining address while reservation queue is updated, where queue is transferred with data - Google Patents

Abstract

The system has an address generator for determining an address while a reservation queue (Q) is updated, where a sequence of transaction nodes (T) is addressed for performing a transaction. The sequence of transaction nodes is defined by entries in the reservation queue. The reservation queue is dynamic and transferred with data (D) to be processed within the sequence of transaction nodes. The address generator is designed as a central root node or distributed hash table or operated in accordance with broadcast basis. Independent claims are also included for the following: (1) a method for coordinating access to limited resources in a distributed system (2) a computer program product comprising a set of instructions to perform a method for coordinating access to limited resources in a distributed system.

Description

The present invention is in the fields of information technology and network technology, and more particularly relates to a computer-based system for access coordination to limited resources in a distributed network system.

Modern software systems are typically designed as so-called distributed systems and have a computer architecture consisting of a plurality of computer-aided nodes on which data processing operations are carried out. In particular, in systems based on a service-based architecture (so-called service-oriented architecture - SOA), a central data repository is frequently used, which communicates with a large number of transaction nodes over the network.

Disadvantages of such systems, which are based on a central data repository, on the one hand, is that there is a performance bottleneck that results from the necessary data exchange with the central repository. On the other hand, this can represent a security vulnerability. Another disadvantage is the fact that in a large network with (geographically) very distributed network nodes local stability properties of the network can not be exploited. Local stability characteristics include scenarios in which one local subnetwork is functional, while another subnetwork is no longer working, or interconnect structures between the network nodes are compromised and only work on the local subnetwork. Disadvantageously, in a network based on a central data repository, it is no longer possible to use the locally functioning data processing in the aforementioned error scenarios. Thus, the locality properties can not be meaningfully exploited.

Basically, the coordination of access in the database technology is a known problem. Depending on the design of the database system, the data is either stored centrally in a database or distributed over several databases and managed and addressed via a central coordinator. In order to be able to guarantee a conflict-free database transaction, the following features should be assured: atomicity, consistency, serializability and persistence of the data. To ensure this, central lock mechanisms are usually provided to lock accesses exclusively. The data is transferred to the processing client, edited there and sent back to the central database repository after successful processing or modification. Afterwards (after successful data processing), the lock can be released again, so that the data can also be accessed (and edited) by other clients. If the change extends over several distributed computer-based instances or stations, a so-called two-phase commit protocol for the coordination of distributed transactions is frequently used in the prior art. With the 2-phase commit protocol, however, it is necessary to provide a central coordinator, who must be known to all transaction nodes and who takes over the coordination in a central function. However, this is a serious drawback when decentralized coordination brings significant advantages, in particular performance advantages, for distributed data and distributed applications.

Moreover, it is also known in the art to use decentralized database access coordination. In this case, several schedulers (or coordinators) are used, which are each arranged between a requesting node (client) and the data and control the accesses and coordinate. The data are stationary and fixed on the database and, if necessary, processed by one or more stations and thus changed and then stored back in an altered state back into the database. In the case of the known methods, the data are thus stationary in each case at the same storage location, namely in the database. However, this is a significant disadvantage and sometimes not applicable if the data is to be flexibly and decentrally coordinated.

Based on this prior art, the present invention has the object of demonstrating a possibility with which the access coordination in a distributed network can be improved without compromising the consistency properties of the transactions. In addition, the access coordination should be decentralized.

This object is solved by the appended claims, in particular by a computer-based system, a computer-based method and a computer program product. It should be noted that the method is computer implemented and can be implemented as software and / or hardware.

In the following, the solution of the problem will be described according to one aspect of the invention with regard to the method according to the solution. Here mentioned advantages, features and alternative embodiments are also to be transferred to the other patent claims or categories, ie in particular to the computer-based System and the computer program product. It should be mentioned that corresponding functional features of the method are implemented by corresponding components of a microprocessor or as other hardware units that take over the functionality of the corresponding software feature and vice versa. In one aspect, the present invention thus relates to a computer-based method for access coordination to limited resources in a distributed computer system. In doing so, computer-based instances in the computer network interact with each other and exchange data. It is intended to provide a plurality of transaction nodes and a reservation queue and an address provider.

The transaction nodes are used to perform the data access, which will usually be carried out as part of a data processing and thereby requires an exclusive or parts-exclusive sovereignty over the data to be processed.

The queue is used to manage all access requests for exclusive or part-exclusive reservation of resources and to coordinate access. According to one aspect of the invention, provision is made for the reservation queue to be kept available in each case at the transaction node which currently carries out the data processing and / or which currently has the sovereignty over the data to be processed. An essential aspect of the invention relates to the fact that the reservation queue is transferred with the data. Thus, the reservation queue is transmitted along with the data under access in the network.

The address generator is used to identify and manage an address via which the reservation queue can be reached and addressed. The address is maintained dynamically and automatically updated upon change, so that it is ensured that a current address of the reservation queue can always be made available.

As part of a transaction or data processing, a specific sequence of operations is now to be executed on a dataset. This can be done on a transaction node of the network or on several transaction nodes, in the latter case, the transaction nodes are addressed or addressed sequentially. The sequence of transaction nodes for performing the access is defined by entries in the reservation queue. It is usually provided that a node enters a request for a resource directly in the reservation queue. Alternatively, an indirect description of the reservation queue may be provided here. The individual transaction nodes of the reservation queue are then processed or addressed in the intended sequence. According to one essential aspect, the data is transmitted together with the reservation queue from transaction nodes to transaction nodes. The current address of the reservation queue or of that transaction node is updated with data sovereignty.

In the following, the terms used in the context of this application are explained in more detail.

The computer-based system is designed as a network system having, for example, a SOA architecture and providing certain services that can be executed on the transaction nodes. In this case, the association between service and transaction node is free, so that it is possible that several - even different - services on a transaction node or distributed to multiple transaction nodes can be executed. It is also possible that a service is executed on several transaction nodes and thus distributed. The transaction nodes are in data exchange via a network that can be operated in different network types and / or with different topologies, for example as a Local Area Network (LAN) or as a TCP / IP-based network.

The limited resources can be data or data stocks, which can sometimes be kept available distributed on different transaction nodes. Likewise, this term are subject to other resources, such. B. those that are operated exclusively or partially-exclusive and therefore must be reserved. These include, for example, printers, memory systems, central processing units (CPU's), computation units, compression modules, parts of a mechatronic system, parts of software or hardware systems, control units, control units and other processing modules.

The transaction nodes are computer-based instances of different categories. Thus, a transaction node is subject to microprocessor chips as well as computers, computer systems or computer networks with different computer architecture. At the respective transaction nodes, a data access is executed which requires an exclusive or partial exclusive sovereignty over the data to be processed. This can be any type of access, ie a write (WRITE), a read (READ) and / or an execute (EXECUTE) access.

The reservation queue will usually be implemented in software and includes a sequence of entries for exclusive or part-exclusive resource reservation. According to the invention, it is provided that the accesses are coordinated and managed via the reservation queue. In addition, the transaction nodes are addressed according to the reservation queue. Depending on the entries in the reservation queue, the data with the reservation queue is transmitted sequentially from node to node. It is important that all participating computer-based instances and thus also all transaction nodes, the current address of the reservation queue (and thus the transaction node with data sovereignty) is known. According to the reservation queue is transmitted along with the data to be processed in the network. In accordance with one aspect of the invention, provision is made for the reservation queue and the data to be transmitted in an encapsulated packet and in particular over the same data channel. Alternatively, however, different data transmission paths can be provided, which, however, coincide with respect to sender and receiver. It is likewise possible for the reservation queue and the data to be sent to the same addressee separately from one another or independently of one another via the same communication channel (eg in separate electronic messages).

By the "address provider" is meant an instance that is intended to determine the current reservation queue address. In other words, the participating transaction nodes are enabled to know the current address of the reservation queue via the address provider. The address provider can be implemented in different ways. So it is possible in a first variant to use a broadcast principle here. The broadcast principle stipulates that the respective transaction node with data sovereignty sends its address to all other nodes. In a second variant, a central address node can be provided which manages the current address of the reservation queue. In this case, it is provided that the respective transaction node with data sovereignty sends its address to the central address node and that the requesting transaction nodes can find out the current address of the reservation queue via the address node. Usually, the central address node will be formed as a root node, from which the transformation of the data is initiated. According to a third variant, a distributed hash table can also be provided, via which the respectively current address of the reservation queue can be determined.

In the invention, two variants are preferably provided, as the transaction nodes transmit their access request or reservation request. On the one hand, the transaction node can, if necessary, transmit its reservation request to the address provider. This will preferably be the case when a central node is provided as an address node. On the other hand, it is possible that the respective transaction node makes its reservation request directly to the transaction node in use and thus to the transaction node, which currently has the data sovereignty and on which the reservation queue is kept. Of course, according to alternative embodiments, other computer instances can also be involved, so that the reservation request is forwarded indirectly and thus indirectly to the respective reservation queue node.

Particularly in large network systems that are arranged over a large geographical area (eg transnationally), it is preferably provided that a structure of transaction nodes is provided. It is provided that a certain amount or selection of transaction nodes is assigned to a subnetwork. The individual sub-networks are usually managed separately or independently of each other. As an example, corporate networks (usually organized as LANs) or country-specific subnetworks are mentioned here. An essential advantage of the solution according to the invention can now be seen in the fact that even in the event of a failure of a first sub-network, error-free access coordination and data processing can still be performed on other sub-networks. It is also possible that local data processing can still be performed even if external data connections (that is, outside the internal network) are corrupted.

According to a variant of the solution according to the invention, a hierarchy of transaction nodes can also be provided. For example, so-called root nodes can be provided, from which the data processing with the respective accesses is initiated and which are addressed, so to speak, as first transaction nodes in the sequence of transaction nodes. In this case, a hierarchy of root nodes can also be implemented (for example, a global root node and a country-specific root node).

According to one aspect of the invention, it is provided that the data is used for the configuration of technical devices which are addressed in succession, such as a decompression unit for decompressing received data and forwarding it to a presentation unit, which then "renders" the data. , In one Next, this data will be forwarded to the printer to create a printout. In addition to the use of data for the configuration of technical devices, other data uses are conceivable, for. B. for pure data processing. At this point, however, it should again be noted that it is not absolutely necessary for the execution of the invention that the data be processed at all transaction nodes. Thus, it is also possible that the data is only read in to the transaction node or used for subordinate processing steps.

According to the invention, it is ensured that the current processor (in the form of a transaction node) also has exclusive access to the data, the data (for access) being available exclusively locally at the transaction node (and at no other transaction node of the network).

In contrast to known pipeline architectures, it is provided in the solution according to the invention not to determine the order of processing or the order of the transaction nodes to be used a priori. According to the invention, the order of the transaction nodes to be used can be determined dynamically on a case by case basis. In addition, according to the invention, it is also not intended to make it absolutely necessary to provide a central administrator or coordinator who is used to manage the accesses. According to the invention, a decentralized access coordination is obtained, in which the reservation queue is used as an access coordinator and is sent together with the data from transaction nodes to transaction nodes.

According to a preferred embodiment, different optimizations of the access method are provided.

Thus, on the one hand, the transactions can be rolled back if an error is detected (in the network or at certain transaction nodes). For this purpose, it may be provided that a copy of the processing data remains on the previously involved transaction nodes. It is preferably configurable to which and / or on how many transaction nodes the data is stored redundantly before the current transaction node. The following example is intended to explain this in more detail: If, for example, the following nodes A, B, C, D, E and F were involved in the transactions within the context of an access coordination and node F is the current node, a triple redundancy can be provided. This means that the nodes E, D and C have a copy of the record, the copy corresponding to the processing by the respective node, so that on the node C is a copy of the record after processing by the node C and on the Node D is a record that corresponds to the processing by the node D etc.

Advantageously, with this embodiment, the security of access can be increased by resorting to earlier editing versions.

In a further optimization, it may be provided to design the access coordinator with further control and / or regulation options. For example, it is possible to automatically detect which other entries exist in the reservation queue. These further entries of transaction nodes are then evaluated with regard to certain criteria. For example, it would be conceivable to analyze which of the future transaction nodes are located within the same local subnetwork and would nevertheless be executable in the event of a fault of another subnetwork. Likewise, a prioritization of transaction nodes could be made.

According to the solution according to the invention, it is provided that the data arrive in a specific order at the selected transaction nodes. The transaction nodes that are higher in the reservation queue receive the data earlier than those transaction nodes that are registered at a subordinate position in the reservation queue. This also defines a sequence of the processing. It should be emphasized that according to the invention no central administration is necessary and the accesses to the respective transaction nodes are managed decentrally. The access coordinator is transmitted along with the data to be processed on the network. This ensures that both the data and the access coordinator are in the same physical context or on the same transaction node, so that even with limited connectivity, the data can still be processed.

Another solution to the above object is a computer-based system designed to carry out the above-mentioned method. The system includes a plurality of transaction nodes and a reservation queue, as well as at least one address provider. The access coordinator in the form of the reservation queue is thereby transferred together with the data from node to node.

Another task solution is a reservation queue, which is provided as a data structure. The reservation queue may be implemented as a list of resource reservation entries. According to the reservation queue is transferred as a data structure together with the data to be processed. This can preferably take the form of a common data packet or within a common message. Alternatively, different transformation mechanisms are also conceivable here, so that the reservation queue is sent in a separate message to the same recipient as the data D).

Another task solution provides a computer program, which is intended for carrying out the method described above. It is possible that the computer program is stored on a disk. The storage can be provided as executable or as loaded code. Likewise, the computer program may be available as a download from a server. Alternatively, the computer program can also be stored on another storage medium or provided as a distributed system, so that individual modules of the method described above come on a client and other modules on a server for execution.

An alternative solution of the tasks is also in a computer-implemented method according to the attached method claim, which can be distributed in the form of a computer program.

Further solutions to the problem and alternative embodiments can be found in the accompanying claims.

In the following detailed description of the figures, non-limiting exemplary embodiments with their features and further advantages will be described with reference to the drawing. In this show:

1 an overview representation of transaction nodes in a specific system state according to a preferred embodiment of the invention,

2 an overview of the in 1 represented system in another state,

3 an overview and schematic representation of the in the 1 and 2 shown system in another state,

4 a schematic representation of transaction nodes and transmitted messages according to a preferred embodiment of the invention and

5 an overview representation of participating transaction nodes and their communication with other computer-based entities according to a preferred embodiment of the invention.

The following is - with reference to the 1 to 3 - The invention explained in more detail according to a preferred embodiment.

As in 1 In the computer-based system, a plurality of transaction nodes T are provided. The highest authority in the 1 to 3 labeled "Data Home", shows the source of the data. In a preferred embodiment, the data source also represents a root node for reservation requests. The other transaction nodes T are labeled "System A", "System O", "System R", "System X" and "System Y". As in 1 As can be seen, the transaction nodes T with the systems A to O had exclusively reserved the data at the time in the past. In the system state in 1 is shown, the system R currently has the sovereignty over the data and thus also has the (in itself configurable) right over the resources and resources. According to the invention, it is provided that a reservation queue Q is also kept at the current transaction node T, which has the data sovereignty. The reservation queue Q is in the in 1 empty example. By definition, that system or the transaction node T has the data sovereignty on which data D is located. In this case, this is the node T named "System R".

In 2 Now another system state is shown. The following transaction nodes T "System X" and "System Y" are added. Both "System X" and "System Y" make reservation requests (here: at or via the root node). The root node points to the current data ruler - in this case this is "System R".

According to the invention, it is thus provided that requesting computer-based entities receive a message requesting a reservation with an address of the transaction node T which currently has the data sovereignty and on which the data D is currently located. Then it is the requesting systems "System X" and "System Y" possible to make each a reservation request directly to the transaction node T with data sovereignty (here: "System R").

The messages sent by the requesting systems ("System X", "System Y") to the root node ("Data Home") are in 2 marked with the reference symbols X1 and Y1. X1 is to identify the request of the system X and Y1 the request of the system Y. The messages to be sent by the requesting systems ("System X", "System Y") are sent to the data sovereign node ("System R") are in 2 marked with the reference symbols X2 and Y2. The message X2 indicates the reservation request of the system X to the transaction node T and the reference Y2 indicates the reservation request of the system Y to the current data host node ("System R").

After transmission of the reservation requests of the requesting systems to the transaction node T with the reservation queue Q, the reservation entries are automatically entered in the reservation queue Q. This is in 2 characterized in that the reservation queue Q is no longer empty (as originally in 1 was shown), but now includes the entries Y and X.

After completion of the processing of the data on the transaction node T "System R", the thus processed data D are transmitted together with the transaction queue Q to that transaction node T, which is in the first place in the reservation queue Q. In the in the 2 and 3 As shown, this is the node "System X". According to one aspect of the invention, the data D and the reservation queue Q are thus transmitted in a message packet Q, D.

• – Eine erste Ausführungsform sieht es vor, dass die so bearbeiteten Daten D auf dem letzten Transaktionsknoten T verbleiben.
• – Gemäß einer zweiten Ausführungsform ist es vorgesehen, dass die so bearbeiteten Daten von dem letzten Transaktionsknoten T an den Wurzelknoten ”Data Home” zurück übermittelt werden.

The method described above is carried out iteratively on all the transaction nodes T registered in the reservation queue Q until all pending transactions are executed or until the data processing is finished. Once the reservation queue has been processed, there are basically two possibilities:

• A first embodiment provides that the data D thus processed remain on the last transaction node T.
• According to a second embodiment, it is provided that the data thus processed are transmitted back from the last transaction node T to the root node "Data Home".

Of course, there are also other further processing or forwarding to other computer-based entities within the scope of the invention.

In connection with 4 Let us now show an example of a system architecture of the solution according to the invention. In this case, a multiplicity of transaction nodes T ₁ , T ₂ ,... T _n communicate with one another via a network N. The network N is preferably a TCP / IP network. Alternative versions, however, provide other protocols and network systems.

Preferably, the reservation queue Q and the data are transmitted in a common message packet Q, D. This message packet may be encapsulated by other messages. Usually, the message packet is transmitted in a message. Alternatively, however, different messages can be used here. Moreover, it is possible to send the reservation queue Q and the data D in separate messages and / or via separate communication channels (different protocols, different network frequencies, different data lines, etc.). This makes the system more flexible, but also less secure because inconsistencies must be checked in the event of a channel failure.

In 5 the communication of a transaction node T with other computer-based instances is shown. According to a preferred embodiment, a transaction node T comprises a central processing unit CPU or a microprocessor. In addition, a memory (memory), interfaces (user interface - UI), an operating system (operating system), interfaces of other types (interface) and communication modules (communicator) is provided. Usually, different types of memory are provided (Random Access Memory, Read Only Memory, etc.). In addition, a clock and usually also graphical user interfaces, for example in the form of a touchscreen, are provided. Of course, any other processing units can be integrated into the transaction node T or connected to the latter via corresponding interfaces. In other embodiments, the transaction nodes T are made slimmer and include only a selection of the aforementioned electro-technical or electronic modules. The transaction node T is in communication with an address generator A via the network N.

The address generator A serves to provide the current address of the reservation queue Q or of the transaction node T, which currently holds the data sovereignty. For the realization of the address generator A different variants are provided. On the one hand, it is possible that the address generator A is formed by the central root node, which in the 1 to 3 labeled "Data Home". In these figures, this variant is shown, therefore the requesting systems X and Y also ask the root node to know the current address of the reservation queue Q.

Alternatively, the address generator A can also be designed as a distributed hash table.

In a third variant, it is possible to realize the function of the address generator A by a broadcast principle. According to this variant, it is provided that the respectively current data controller node sends its state (data sovereignty) to all other transaction nodes T which are connected to the system via the network N. Although this third variant relieves the load on the root node, it disadvantageously leads to increased network traffic since, in the event of a transition from one data node to the next, a large number of messages must be exchanged.

Of course, it is also within the scope of the invention to implement the address generator A as a separate module on one or more selected transaction nodes T of the system.

Usually, it is provided that the data is not fragmented and thus is not distributed to a plurality of transaction nodes T. The access should be executed safely on this dataset. Alternatively, however, it may also be provided that the data stock is fragmented and thus distributed to different transaction nodes T. Then the method described above is applied to the individual data fragments.

According to one aspect of the invention, the operations to be performed on the dataset are to be performed transactionally. In this case, according to the invention, the data consistency can be assured, since it is ensured that in each case only one data host node T has the data sovereignty over the data stock at a time.

Another advantage is that a relatively high degree of serializability is possible.

An advantageous development of the invention provides that the individual transactions that are executed on the transaction node T can be rolled back to a certain preconfigurable state (rollback). For this purpose, it is provided that in each case a copy of the data remains on a transaction node T, while the message packet from reservation queue Q and data D is forwarded to the next transaction node T. If an error then arises in the further processing on the next transaction node T and it therefore becomes necessary to rewind the transaction so that the previous state of the data stock must be restored, the copy of at least one preceding transaction node T can be used. Advantageously, it is adjustable on how many previous transaction node T before each current transaction node T a copy of the data D should remain. If, for example, the number "3" is defined as rollback, a copy of the data processed there remains on the three preceding transaction nodes T in each case. If the number is exceeded, the data on the respective node T can be deleted.

In another advantageous development of the invention, the reservation queue Q is designed with additional functions. For this purpose, it comprises further modules which can be implemented either as a microprocessor component and / or as a software module.

Thus, a prioritization module is designed to prioritize the reservation requests entered in the reservation queue Q. The parameters for prioritization are preferably configurable. Depending on the implementation of the queue model, different parameters can be used here, for example: operating time, number of operating stations, priority in the application process, etc. This makes it possible, optionally, to re-sort certain entries in the reservation queue Q according to their priority in the sequence sort the entries. The priority criteria can also be modified dynamically, ie also at runtime. This advantageously makes it possible to adapt the access coordination very flexibly to the application case.

In another aspect, the queue may be operated with different scheduling algorithms. These include priority-based scheduling with static or dynamic priorities, and service-time-based scheduling (for example, Shortest Job First - SJF scheduling). With this strategy, it is envisaged that the shortest job is calculated first.

According to a further variant of the invention, it is provided that the message packet comprising the reservation queue Q and the data stock D comprises further data, for example specifications of the procedure or methods which are to be executed on the dataset. Likewise, meta information for sending and / or receiving messages can be transmitted.

As already mentioned above, there are different possibilities for the implementation of the queue model. To process the queue, it may also be provided to consider a load-balancing module. The load-balancing module serves to optimize the load distribution between the individual transaction nodes T and to design the data processing so that as far as possible Transaction node T are activated, which do not overload at the current time. In this variant, it is provided that additional load information messages are exchanged between the transaction nodes T. These load information messages indicate the load with which the respective transaction node T is currently being operated at the respective time. Depending on the node load then a reorganization of the reservation queue Q can be made.

According to a further advantageous embodiment of the invention, it is provided that all or selected states of the reservation queue Q are kept. This provides a persistent reservation queue Q, so to speak. This has the advantage that, in the event of a system crash, earlier states of the reservation queue Q can also be provided in order to be able to trace back the "path of the data stock" in the network. At this point, it should be noted that the embodiment shown in the 1 to 3 is just one possibility of a network topology. However, the invention is not limited to this network topology and can be applied to other network types and architectures.

In principle, a transaction node T can be operated in different roles: on the one hand as a data controller node and on the other hand as a requesting transaction node. Depending on the embodiment, further rollers may be provided (eg as root nodes).

• – Er greift auf den Adressgeber A zu, um den aktuellen Datenherrscherknoten in Erfahrung zu bringen. Dies kann beispielsweise durch Versenden einer Nachricht an den Wurzelknoten ausgeführt werden.
• – Der anfragende Knoten sendet eine Reservierungsanfrage an den Datenherrscherknoten. Dabei übermittelt er einen Identifikator des anfragenden Knotens bzw. des wartenden Prozesses und möglicherweise noch Zusatzinformationen, wie Lastnachrichten, Metainformationen zur Bedienzeit des anfragenden Systems etc.

According to a preferred embodiment of the invention, it is provided that a network node T carries out the following steps in its role as requesting accounts:

• - It accesses the address generator A in order to find out the current data controller node. This can be done, for example, by sending a message to the root node.
• The requesting node sends a reservation request to the data master node. He transmits an identifier of the requesting node or the waiting process and possibly additional information, such as load messages, meta information on the service time of the requesting system, etc.

• – Er bearbeitet die Daten durch Ausführen von Transkationen.
• – Er empfängt eine Reservierungsanfrage eines anfragenden Transaktionsknotens und trägt die Reservierungsanfrage gegebenenfalls nach Prüfung auf Zulässigkeit in die Reservierungsschlange Q ein, die lokal vorgehalten wird.
• – Nach Abarbeitung der Transaktion erfolgt ein Zugriff auf die Reservierungsschlange Q, um den nächsten Transaktionsknoten als Empfänger des Datenpaktes Q, D zu bestimmen.
• – Das Datenpaket Q, D wird automatisch generiert und an den bestimmten Empfänger (den nächsten Transaktionsknoten T gemäß der Reservierungsschlange Q) weitergeleitet. Optional kann das Datenpaket noch Zusatzinformationen umfassen, wie Lastnachrichten, Metainformationen für bereits erfolgte und/oder noch auszuführende Transaktionen etc.

According to one aspect of the invention, it is provided that a network node T carries out the following steps in its role as processing nodes or data controller accounts (used synonymously for the term "nodes with data sovereignty") (in this case, the order of the steps is not necessarily predetermined but can can also be varied):

• - He processes the data by executing transactions.
• - It receives a reservation request from a requesting transaction node and enters the reservation request, if necessary after checking for admissibility in the reservation queue Q, which is held locally.
• After the transaction has been processed, access is made to the reservation queue Q in order to determine the next transaction node as the recipient of the data packet Q, D.
• The data packet Q, D is automatically generated and forwarded to the specific receiver (the next transaction node T according to the reservation queue Q). Optionally, the data packet may also include additional information, such as load messages, meta information for already performed and / or yet to be executed transactions, etc.

The steps in the preferred embodiment of the invention are carried out automatically and without user interaction.

Finally, it should be pointed out that the different embodiments mentioned in the above description can also be used in combination. Thus, some of the mentioned aspects can also be combined. In addition, it is possible to implement parts of the method for coordinating accesses on a first transaction node T ₁ and other parts of the method on another transaction node T ₂ . Thus, the inventive method can also be used as a distributed system, for. B. according to the client-server principle operated. The main application of the invention is in the field of the technical resources of a computer system which require access coordination, e.g. As printers, printer drivers, ports, other interfaces, connected technical equipment or instruments. Basically, computer-internal and / or computer-external resources can be affected. For example, in the medical field of application computer-external resources could be in connected devices, such as in image acquisition devices (X-ray, CT, MRI, etc.) or in laboratory equipment or in diagnostic stations.

In summary, the invention can be described in abbreviated form as follows:
For access coordination on resources which have to be reserved exclusively or partially exclusively by participating transaction nodes T, a reservation queue Q is provided which is transferred together with the data stock D in the network N. The reservation queue Q acts as a dynamic and decentralized access coordinator.

LIST OF REFERENCE NUMBERS

• T

  transaction node

  Q

  reservation queue

  D

  Data or data stock

  Q, D

  News package

  A

  address donor

  X1

  Message: Search for address of Q

  X2

  Reservation request from X → Q

  Y1

  Search for address Q / R

  Y2

  Reservation request from Y → Q / R

  N

  TCP / IP network

Claims (10)

Computer-based access coordination system for limited resources in a distributed system where computer-based entities communicate with each other over a network (N), comprising: A plurality of transaction nodes (T), wherein a transaction node (T) carries out a transaction in each case and has the exclusive or partial-exclusive sovereignty over the data (D) to be processed, A reservation queue (Q), in which all access requests for exclusive or partial-exclusive reservation of resources are entered to coordinate the access, and which is respectively kept at the transaction node (T), which is currently executing a transaction and / or the Sovereignty over the data to be processed (D) and which is in-transferred with the data (D) At least one address generator (A), which is intended to determine an address under which the reservation queue (Q) is currently kept, wherein a sequence of transaction nodes (T) is addressed for executing a transaction, the sequence of transaction nodes (T) being defined by entries in the reservation queue (Q), and wherein the reservation queue (Q) is dynamic and is transferred with the data (D) to be processed within the sequence of transaction nodes (T). The computer-based system of claim 1, wherein the access comprises data access and / or resource access. A computer based system as claimed in any one of the preceding claims, wherein the system comprises a mechatronic system, an embedded system or a software based system. Computer-based system according to one of the preceding claims, wherein the data to be processed (D) with an access coordinator, in particular with the reservation queue (Q), are transferred. Computer-based system according to one of the preceding claims, in which the address generator (A) is a central root node or a distributed hash table or is operated according to a broadcast principle. Computer-based system according to one of the preceding claims, wherein the access comprises a READ, WRITE and / or EXECUTE access and / or other types of access. Computer-based system according to one of the preceding claims, in which the transaction nodes (T), if necessary, make a reservation request to the address provider (A) and / or to a current transaction node (T) on which the reservation queue (Q) is held at that time. A computer-based system according to any one of the preceding claims, wherein a structure or hierarchy of transaction nodes (T) is provided, wherein a selection of transaction nodes (T) are each associated with a sub-network and are preferably managed independently of other sub-networks. A method of access coordination to limited resources in a distributed system where computer-based entities communicate with each other over a network (N), comprising: A plurality of transaction nodes (T), wherein a transaction node (T) is in each case intended for a specific transaction and thereby has the exclusive or partially exclusive sovereignty over the data (D) to be processed, A reservation queue (Q), in which all access requests for exclusive or partially-exclusive reservation of resources are entered and which are respectively kept at the transaction node (T), which currently carries out the data processing and / or the sovereignty over the data to be processed (Q). D) and which is transferred with the data (D), At least one address generator (A), which is intended to determine an address under which the reservation queue (Q) is currently kept, wherein a sequence of transaction nodes (T) is addressed for executing a transaction, the sequence of transaction nodes (T) being defined by entries in the reservation queue (Q), and the reservation queue (Q) being dynamic and containing the data to be processed ( D) within the sequence of transaction nodes (T). Computer program product that can be loaded directly into an internal memory of a digital computer and includes software code sections that perform the steps of the method according to the above method claim when the computer program product is loaded on a computer or is executed.

Images