TWI670607B - High-availability provisioning system in distributed architectures and method thereof - Google Patents

Abstract

The invention relates to a distributed high-availability supply system and a method thereof, and the distributed high-availability supply system adopts a distributed high-availability supply process technology, and supports the supply required for the supply task under the distributed deployment architecture. The process is executed interactively by a plurality of work units under the system. The invention satisfies the high availability of the supply process, the horizontal scalability, the workload balancing and the execution efficiency of the supply process, and can also deploy the work unit according to the actual computing energy required by the service to achieve resource utilization optimization.

Description

Decentralized high availability supply system and method thereof

The present invention relates to service provisioning mechanisms and, more particularly, to a decentralized high availability supply system and method therefor.

In the past, Provisioning Technology used many small module parts included in a large system to belong to a single host. The disadvantage is that when a module fails or is disabled, it will affect the operation of the entire system.

At present, the supply technology is dedicated to micro-modeling and decentralized efforts, that is, by disassembling and dispersing multiple sub-modules in a large system into different hosts to reduce the module's disability to other modules. It may affect. However, when the sub-modules are dispersed, the modules need to transmit data to each other. If there is no proper data transmission mechanism, there may be a problem of low system performance. Specifically, although the current supply system has a decentralized concept, because the data transfer mechanism collects Chinese management and subordinates, the data transfer host must carry information exchange requests from a plurality of host systems, resulting in high load and loss. It can be risky, so it does not meet high availability requirements.

It can be seen from the above that if a service supply mechanism can be found, so that the supply system can have high availability control and good communication technology, especially in a decentralized system, it may be because of the high load of information transmission between the hosts. As a result, the risk of disability is increased and the performance is poor, which will become an urgent problem for those skilled in the art.

The object of the present invention is to provide a distributed service provisioning mechanism, which can be applied to a distributed subordinate architecture to provide an appropriate supply process, and in particular to consider the problem that information transmission between hosts can be highly loaded, so the present invention proposes A variety of event delivery mechanisms to reduce communication problems between hosts due to high load.

The invention provides a distributed high-availability supply system, comprising: a supply work management host group, a supply station management group, and a supply station management and setting host group, wherein the supply work management host group The first category host includes a supply work management unit and a first message management unit, wherein the supply work management unit receives the supply work from the external system for the supply work At the beginning of the life cycle, the first message management unit sends a first message, and the supply station process management host group includes a plurality of second category hosts, each of the second category hosts includes a service station management unit and a a message management unit, wherein the loading station process management unit receives the first message, and the specification parameter of the loading operation is retrieved from the first message, so that the specification parameter is used from the second category host The database obtains the execution sequence of the process, and performs the supply of the supply stations in the supply work according to the execution sequence of the process. The second message management unit sends a second message describing the installation site to be installed, and the installation station management and setting host group includes a plurality of third category hosts, each of the third category hosts includes The station management and setting unit and the third message management unit, wherein the supply station management and setting unit receives the second message, and the second message extracts the installation site to be executed The installation station identification code, based on the supply station identification code, finds the corresponding supply station to execute the corresponding supply station setting, so that the third message management unit sends the reply end. Sending a message to the supply station management unit, wherein the supply station process management unit continues to send the second message to the other supply station after determining that the other supply station has not completed the installation. The supply station management and setting unit is completed until the other supply stations are completed.

In an embodiment, when the supply station management unit determines that all of the other supply stations have completed the installation, the second message management unit returns the execution completion message to the supply work management unit to The supply work management unit is caused to end the life cycle of the supply work.

In another embodiment, the first message management unit, the second message management unit, and/or the third message management unit respectively transmit the first by using one of a random load policy, a priority policy, or a specified work unit policy. The message, the second message, and/or the third message.

In another embodiment, the loading and management unit further records the identification code, parameters, specifications, execution status and time information of the loading work in the database of the first category of hosts.

In addition, the execution sequence of the process records that each of the supply stations performs the loading and unloading The order of precedence.

The present invention is directed to a distributed high availability supply method, which is suitable for the installation system to execute a supply program, the supply system comprising a plurality of first category hosts respectively having a supply work management unit and a first message management unit, Each of the plurality of second-class hosts having the installation station management unit and the second message management unit, and the plurality of third-class hosts respectively having the installation station management and setting unit and the third message management unit, the method includes: Having the supply work management unit receive the supply work from the external system to send the first message through the first message management unit at the beginning of the life cycle of the supply work; and to enable the supply station to receive the process management unit After the first message, the specification parameter of the loading operation is retrieved from the first message, so that the process execution sequence is obtained from the database of the second category host through the specification parameter; and the supply station management unit is configured According to the execution sequence of the process, the supply of each supply station in the loading operation is performed, wherein the second message management unit sends a record indicating that the supply is to be performed. a second message of the station; after the receiving station management and setting unit receives the second message, the second message is used to retrieve the installation station identification code of the installation station to be installed.找出 finding a corresponding supply station according to the installation station identification code; and causing the supply station management and setting unit to execute the corresponding installation setting of the corresponding station to be used by the third message management unit Sending a reply end message to the supply station flow management unit, wherein the supply station flow management unit continuously transmits the other relevant supply station when it is determined that another supply station has not completed the supply Two messages are sent to the supply station management and setting unit until the other supply stations complete the supply.

Compared with the prior art, the distributed high-availability supply system and method thereof proposed by the present invention adopts distributed high-availability supply process technology to support the supply of the required tasks for the distributed task under the distributed deployment architecture. Process, in which the different work units (ie, multiple host groups) under the installation system are executed interactively, thereby satisfying the high availability of the supply process, scale out, workload balancing, and efficiency of the installation process. Optimized, the work unit can also be deployed according to the actual computing energy required by the service to optimize resource utilization.

1‧‧‧Distributed high availability supply system

11‧‧‧Installation work management host group

111‧‧‧First class host

1111‧‧‧Installation work management unit

1112‧‧‧Message Management Unit

12‧‧‧Installation station process management host group

121‧‧‧Second category host

1211‧‧‧Site station management unit

1212‧‧‧Message Management Unit

13‧‧‧Installation station management and setting host group

131‧‧‧Category of the third category

1311‧‧‧Installation station management and setting unit

1312‧‧‧Message Management Unit

301‧‧‧Host A

302‧‧‧Working unit

303‧‧‧Message Management Unit

305‧‧‧Host Group B

306‧‧‧Selected host B1 randomly

307‧‧‧Working unit

308‧‧‧Message Management Unit

309‧‧‧Host B2

310‧‧‧Working unit

311‧‧‧Message Management Unit

401‧‧‧High priority host B1

404‧‧‧Low priority host B2

501‧‧‧Designated host B1

504‧‧‧Not designated host B2

S21~S25‧‧‧Steps

S601~S611‧‧‧Process

1 is a schematic structural diagram of a system of a distributed high availability supply system of the present invention; FIG. 2 is a step diagram of a distributed high availability supply method of the present invention; and 3A-3C is a high dispersion type of the present invention; A schematic diagram of three message delivery modes of the availability supply system; and FIG. 4 is a schematic flow chart of a specific implementation of the distributed high availability supply method of the present invention.

The technical contents of the present invention are described below by way of specific embodiments, and those skilled in the art can easily understand the advantages and effects of the present invention from the contents disclosed in the present specification. The invention may be embodied or applied by other different embodiments.

Please refer to FIG. 1 , which is a schematic structural diagram of a system of a distributed high availability supply system of the present invention. Decentralized high availability supply system as shown The system 1 includes a supply work management host group 11, a supply station management group 12, and a supply station management and setting host group 13, wherein the supply work management host group 11 is a host group of the supply work management unit. It includes a plurality of first category hosts 111 (hosts A1-An), each of the first category hosts 111 includes a provisioning work management unit 1111 and a first message management unit 1112, and the loading station process management host group 12 is provided The host group of the station flow management unit includes a plurality of second category hosts 121 (hosts B1-Bn), and each of the second category hosts 121 includes a loading station flow management unit 1211 and a second message management unit 1212 for loading The station group management and setting host group 13, that is, the host group of the installation station management and setting unit, includes a plurality of third category hosts 131 (hosts C1-Cn), and each of the third category hosts 131 includes installation station management and The setting unit 1311 and the third message management unit 1312.

The loading work management unit 1111 can receive the loading operation from the external system (not shown) to send the first message to the loading station through the first message management unit 1112 at the beginning of the life cycle of the loading operation. Flow management unit 1211.

After receiving the first message, the loading station process management unit 1211 retrieves the specification parameter of the loading operation from the first message, and obtains the specification database from the database of the second category host (not shown). The process execution sequence is executed according to the execution sequence of the process, and the second message management unit 1212 sends a second message describing the installation site to be installed. In addition, the loading and unloading work management unit 1111 records the identification code, parameters, specifications, execution status and time information of the loading work in the database of the first category host 111. The above process execution sequence It is a record of the order in which the supply stations perform the installation.

After receiving the second message, the loading station management and setting unit 1311 retrieves the installation station identification code of the supply station to be executed from the second message, and identifies the installation station according to the supply station. After the code finds the corresponding supply station, the installation setting of the corresponding installation station is executed, so that the third message management unit 1312 sends a reply end message to the supply station process management unit 1211, wherein the supply station The process management unit 1211 continues to send the second message about the other supply stations to the supply station management and setting unit 1311 until it is determined that other supply stations have not completed the installation, until the other supply stations All are ready for installation.

The above-mentioned system loading work includes a plurality of loading stations that need to perform the loading mode. If the loading station process management unit 1211 is installed at the first supply station, if it is judged that all other supply stations are When the loading is completed, the execution completion message is returned to the supply work management unit 1111 through the second message management unit 1212, so that the supply work management unit 1111 ends the life cycle of the supply work.

In addition, in order to provide a message transmission mechanism between hosts, and avoid overloading the host due to information exchange, the present invention proposes three different ways to perform message delivery, namely a random load policy, a priority policy, or a designated work unit policy. Therefore, in the present invention, the first message management unit 1112 transmits the first message to the loading station flow management unit 1211 and the second message management unit 1212 to transmit the second message to the loading station management and setting unit 1311 and the third message. When the management unit 1312 sends the reply end message to the installation station flow management unit 1211, all of the above three strategies may be used. Choose to execute the delivery program. The three strategies will be explained later by way of example.

The overall operation of the distributed high availability supply system of the present invention will be specifically described below.

The decentralized high availability supply system 1 is responsible for performing the loading work. After receiving the supply work request, the decentralized high availability supply system 1 can drive different work units, as shown in the figure, the hosts A1, B1 and C1 interact with each other. The division of labor is performed to complete the installation work. Based on the design characteristics of the distributed high availability supply system 1, all execution units are composed of stateless work units. The foregoing work units are classified into three types according to the functional characteristics, namely, a supply work management unit 1111, a supply station flow management unit 1211, and a supply station management and setting unit 1311.

The supply work management unit 1111 is responsible for receiving the supply work given by the external system or the external module and managing the life cycle of the installation work, and the supply station flow management unit 1211 is responsible for generating and managing the execution order of the supply station. The supply station management and setting unit 1311 is responsible for managing the supply stations of all the different tasks, passively receiving the message call of the supply station flow management unit 1211 to perform the setting work of the designated supply station.

Because the distributed high-availability supply system 1 can perform multiple installations at the same time, in order to cope with a large number of installation work, the work units adopt a decentralized deployment architecture, which are respectively deployed on different hosts, and the work units on different hosts use messages. The first message management unit 1112 for transmitting the message is also distributed, and the first message management unit 1112 on each host of the work management host group 11 records the host information of each work unit. And the same type of work unit host group information, and provide To determine the file, you can define the message exchange policy between hosts and use it for host message exchange.

After the loading and unloading work management unit 1111 receives the supply work given by the external module, the life cycle of the penning work is started, and a message is sent through the first message management unit 1112 to notify the loading station process management unit 1211 to take over. The processing, when the deployment host of the installation station management unit 1211 receives the message, the installation station management unit 1211 obtains the necessary installation and operation parameters in the message from the second message management unit 1212 on the host. The process execution sequence of a corresponding set of supply stations is generated, and then the execution order of the supply stations is managed, and each of the supply stations is executed by the message call installation management and setting unit 1311.

After the receiving station management and setting unit 1311 passively receives the message call from the loading station flow management unit 1211, the specified station loading information contained in the message is intercepted to perform the setting work of the loading station and complete the setting. After that, the reply end message is sent to the loading station flow management unit 1211, and the loading station flow management unit 1211 will continue to drive the next supply station until the end of the process. After the process ends, the reply completion message is sent to The work management unit 1111 is installed, thereby ending the life cycle of the supply work.

Since the work unit (the supply work management unit 1111, the supply station flow management unit 1211, and the supply station management and setting unit 1311) adopts a distributed deployment architecture, the same type of work units can be copied and deployed to each other. On multiple hosts, when one of the work units fails, the same type of host group (for the work management host group 11, the supply station process management) The host group 12 and the installation station management and the remaining work units of the same type in the set host group 13) take over the processing to meet the high availability requirements of the supply work. As the volume of business increases, the operational energy required for the installation work can also meet the business scalability requirements by expanding the number of deployments of the work units or improving the computing power of the host.

When the different types of work units, such as the supply work management unit 1111, the supply station management unit 1211, and the supply station management and setting unit 1311, are to exchange information, they may be pre-defined in the first and second message management units. 1112, 1212 message delivery strategy for system load optimization.

The messaging policy includes three strategies: random load policy, priority policy, or specified work unit policy. When the random load policy is adopted, the first and second message management units 1112 and 1212 randomly select the host group information of the same type of work units recorded in the first and second message management units 1112 and 1212 before performing the message transmission. A work unit performs message transfer to achieve dynamic load balancing between the work units and the work units. If the priority policy is adopted, the first and second message management units 1112 and 1212 also select the highest priority work unit host transmission according to the priority order of the same type of work unit host, and the high priority work unit host can provide higher operation. To meet the necessary computing power requirements for the supply work. If the specified work unit policy is adopted, the first and second message management units 1112 and 1212 record the work information of each work unit, and the target work unit host can be obtained and the message is transmitted. The policy is applicable to the pre-message routing plan. A large number of message routing optimization and supply work diversion.

In addition, as shown in the figure, because the same type of work unit can be replicated and deployed to multiple hosts as a host group, as the system traffic increases, the work unit host can be increased, giving the system scalability. When one host in the host group fails, the loading and unloading work can be handled by the working unit of another host, so that the present invention has high availability.

It can be seen from the above that in the design stage of the supply process, considering the characteristics of the distributed high availability supply system 1, all execution processes are combined by stateless work units and individually deployed on the computing host for installation. In progress, a large supply process can be divided into small work units and delivered to work units on different hosts for high availability, scalability, load balancing, and optimization of execution process efficiency. .

In addition, different hosts communicate with each other through messages. To meet the decentralized deployment architecture of the system as a whole, each deployment host needs an Event Manager Unit for message provisioning. This can also overcome the problem of high load of information between the hosts.

Please refer to FIG. 2, which is a step diagram of the distributed high availability supply method of the present invention. The decentralized high availability supply method is suitable for the installation system to perform the installation procedure, and the supply system may include a plurality of first category hosts respectively having a supply work management unit and a first message management unit, respectively having a supply station A plurality of second-class hosts of the process management unit and the second message management unit, and a plurality of third-class hosts respectively having a supply station management and setting unit and a third message management unit.

As shown in the figure, in step S21, the supply work management unit is configured to receive the supply work from the external system to open the life cycle of the supply work. Initially, the first message is sent through the first message management unit.

In an embodiment, the supply management unit further records the identification code, parameters, specifications, execution status, and time information of the installation work in the database of the first category of hosts.

In step S22, after the loading station process management unit receives the first message, the specification parameter of the loading operation is retrieved from the first message, and the specification parameter is obtained from the database of the second category host. The order in which the processes are executed. Among them, the process execution sequence records the order in which the supply stations perform the installation.

In step S23, the loading station flow management unit executes the loading and unloading of the loading stations in the loading operation according to the execution sequence of the flow, wherein the second message management unit sends the loading station that records the installation to be performed. Another second message.

In step S24, after the loading station management and setting unit receives the second message, the second station message retrieves the installation station identification code of the installation station to be installed, according to the The station identification code is used to find the corresponding supply station.

In step S25, the installation station management and setting unit executes the supply setting of the corresponding supply station to send a reply end message to the supply station flow management unit by the third message management unit, wherein The station maintenance process management unit continues to send the second message about the other supply station to the supply station management and setting unit until it is determined that other supply stations have not completed the supply until the other supply station All are ready for installation.

In addition, when the supply station process management unit determines all other supplies When the station completes the installation, the second message management unit returns the execution completion message to the supply work management unit to cause the supply work management unit to end the life cycle of the supply work.

As mentioned earlier, this paper proposes three messaging strategies, including random load policies, priority policies, and designated work unit policies. According to the above, the first message management unit and the second message management unit can respectively transmit the first message and the second message to the installation station and the process management unit by using one of a random load policy, a priority policy or a designated work unit policy. Supply station management and setting unit.

Please refer to FIG. 3A-3C, which is a schematic diagram of three types of message transmission modes of the distributed high availability supply system of the present invention. When the work units in the decentralized high availability supply system, such as the supply work management unit and the supply station management unit, the station management and setting unit, etc., when the message exchange is performed, the messages on each host can be used. The management unit assists in the transmission. The message management unit can support three policy modes. The policy mode needs to be preset to the configuration file of the message management unit. When the message is transmitted, it is transmitted according to the strategy required for the specification of the installation work.

Figure 3A shows the random load strategy, that is, the message management unit randomly selects a work unit host for message transmission, so that the supply work can be dynamically load balanced between the work units. As shown, host A 301 includes a work unit 302 and a message management unit 303 that, when it wants to deliver a message to a host of another host group B 305, for example, includes a random selection of work unit 307 and message management unit 308. Host B1 306, or host B2 309 including work unit 310 and message management unit 311, because Using a random load policy, the message management unit of host A randomly selects a host, such as randomly selected host B1 306, to receive messages by randomly selected host B1 306 and to process messages by randomly selected host B1 306 work unit 307. To do so, when a host is randomly selected, the other hosts, such as host B2 309, will not receive the message.

The third figure shows that the priority strategy is adopted, that is, the highest-priority work unit host in the same type of work unit host is selected for transmission, and the high-priority work unit host can deploy high computing capability hardware to meet the necessary computing capability of the supply work processing. As shown, the host A 301 includes a work unit 302 and a message management unit 303. When it wants to deliver a message to one of the other host groups B 305, because the priority policy is adopted, the message management unit of the host A is The message is passed to the high priority high priority host B1 401 to receive the message by the high priority host B1 401 and to process the message delivery by the work unit of the high priority host B1 401. In this strategy, the remaining hosts are, for example, low priority hosts. B2 404 will not receive the message.

Figure 3C shows the use of the specified work unit policy, which will specify a work unit host to receive messages. This policy is applicable to pre-message routing planning and supports a large number of message offloading and route optimization. As shown, the host A 301 includes a work unit 302 and a message management unit 303. When it wants to deliver a message to one of the other host groups B 305, the message management unit of the host A is used because of the specified work unit policy. The message is delivered to the designated host B1 501 that has been designated to receive the message by the designated host B1 501 and processed by the working unit of the designated host B1 501. In this strategy, the remaining hosts are, for example, Non-designated host B2 504 I won't receive the message.

Please refer to FIG. 4, which is a schematic flow chart of a specific implementation of the distributed high availability feeding method of the present invention.

As shown in the figure, in the process S601, after the supply work management unit receives the request from the external system, the supply work management unit records its identification code, parameters, specifications, execution status and time information for the pen work. In the database, in order to maintain the statelessness of the work unit, the execution status of the supply work is changed to execution, which represents the beginning of the life cycle of the supply work.

In the process S602, the installation work management unit will continue to process the sending message call installation station flow management unit, and before transmitting the message, refer to the specification parameters of the supply work, and transmit through the foregoing three message transmission strategies in the present case, thus According to different strategies, one of the processes S603, S604, and S605 is advanced.

After the loading station process management unit receives the message call from the supply work management unit, the supply station process management unit intercepts the specification of the installation work from the message, and obtains the query according to the specification parameter. A set of process execution sequences corresponding to the installation station, and then the execution of the installation station is managed in this order, as shown in the figure, the installation can be performed by the first supply station. After one of the supply stations has been executed, the supply station process management unit will obtain an execution end message replied by the supply station, and proceeds to process S606.

In the process S606, it is determined whether the supply station is the last supply station in the execution sequence, and if so, it indicates that all the supply stations are all installed. After the execution is completed, the process proceeds to process S611, that is, the execution process ends, and the supply station management unit sends a message to the installation work management unit to terminate the life cycle of the supply work. If not, the call message needs to be sent continuously. Perform the next supply station. Similarly, as shown in the process S607, the call message is also transmitted by three message policies, and according to different strategies, it proceeds to one of the processes S608, S609, and S610.

The supply station management and setting unit is responsible for managing all types of supply stations. When the supply station management and setting unit receives a message call from the process management unit of the installation station, the supply station is intercepted in the interception message. The identification code is executed by the supply station that is managed by the installation station and managed by the setting unit, and the parameter interception and supply setting in the message are performed.

Finally, returning to process S606, it is determined whether the feeding station is the last feeding station in the execution order, and if yes, proceeding to process S611, that is, the execution flow ends, and if not, the call message needs to be continuously sent. Repeat the previous process by executing the next supply station.

In summary, the distributed high-availability loading system and the method thereof are provided, wherein the working unit can be copied and expanded and deployed in multiple hosts, and when multiple jobs are simultaneously installed, the message management unit can Actively adopt a load balancing strategy to distribute the loading and unloading work, thereby achieving load balancing of the supply work and improving the supply efficiency. In addition, when a single work unit fails, the remaining work units of the same type can be handled, so that the supply height can be satisfied. Availability. The invention can calculate the energy of the actual demand of the supply work, deploy the number of hosts of various work units, and pre-plan the deployment of the installation process. The flow execution strategy directs the provisioning of high computing requirements to the high-performance deployment host to achieve resource utilization optimization and execution process efficiency optimization features.

The above embodiments are merely illustrative of the principles and effects of the invention and are not intended to limit the invention. Modifications and variations of the above-described embodiments can be made by those skilled in the art without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention should be as set forth in the scope of the claims described below.

Claims (10)

A decentralized high availability supply system includes: a supply work management host group including a plurality of first category hosts, each of the first category hosts including a supply work management unit and a first message management unit, wherein the supply The work management unit receives the supply work from the external system to send the first message through the first message management unit at the beginning of the life cycle of the supply work; the maintenance station process management host group includes the second a class host, each of the second class hosts includes a loading station process management unit and a second message management unit, wherein the loading station process management unit receives the first message and then extracts the first message The specification parameter of the loading work, the process execution sequence is obtained from the database of the second category host through the specification parameter, and the supply station of the supply station is performed according to the execution sequence of the process, wherein The second message management unit sends a second message describing the installation site to be installed; and the installation station management and setting host group, which includes the plurality of The category host, each of the third category hosts includes a loading station management and setting unit and a third message management unit, wherein the second station message is received by the loading station management and setting unit after receiving the second message Taking the installation station identification code of the installation station to be installed, and determining the supply setting of the corresponding supply station after finding the corresponding installation station according to the installation station identification code, Sending a reply end message to the supply station flow management unit by the third message management unit, wherein the supply station flow management unit determines that there is another supply The installation station does not complete the installation, and continuously sends the second message about the other supply station to the supply station management and setting unit until the other supply stations complete the supply. The distributed high availability supply system of claim 1, wherein the supply station management unit passes the second message management unit when determining that all of the other supply stations are ready for installation. The execution completion message is returned to the supply work management unit to cause the supply work management unit to end the life cycle of the supply work. The distributed high availability provisioning system of claim 1, wherein the first message management unit, the second message management unit, or the third message management unit utilizes a random load policy, a priority policy, or a designated One of the work unit policies, the first message, the second message, or the third message is delivered. The distributed high availability supply system according to claim 1, wherein the supply work management unit records the identification code, parameters, specifications, execution status and time information of the installation work in the first In the category host's database. The distributed high availability supply system according to claim 1, wherein the execution sequence of the process records the order in which the supply stations perform the installation. A decentralized high-availability supply method is suitable for the installation system to execute a supply program, the supply system comprising a plurality of first-class hosts respectively provided with a supply work management unit and a first message management unit, respectively The second category of the station management unit and the second message management unit And a plurality of third-class hosts respectively having a supply station management and setting unit and a third message management unit, the method comprising: causing the supply work management unit to receive the supply work from the external system, so as to At the beginning of the life cycle of the loading operation, the first message management unit sends the first message; after the receiving station process management unit receives the first message, the loading and receiving work is captured by the first message. a specification parameter, wherein the process execution sequence is obtained from the database of the second category host through the specification parameter; and the supply station management unit performs the supply of the supply stations in the supply work according to the execution sequence of the process The second message management unit sends a second message describing the installation site to be installed; and the receiving station management and setting unit receives the second message, and then the second message is captured by the second message. The installation station identification code of the installation station to be installed, and the corresponding installation station according to the installation station identification code; and the installation station management and setting unit executes the The supply station setting of the supply station is configured to send a reply end message to the supply station flow management unit by the third message management unit, wherein the supply station process management unit determines that there are other supply stations After the installation is not completed, the second message about the other supply station is continuously sent to the supply station management and setting unit until the other supply stations complete the installation. Decentralized high availability supply party as described in claim 6 The method, wherein the supply station management unit returns the execution completion message to the supply work management unit through the second message management unit when determining that all the other supply stations have completed the installation, so that the The supply work management unit ends the life cycle of the supply work. The distributed high availability supply method according to claim 6, wherein the first message management unit, the second message management unit, or the third message management unit respectively utilize a random load policy, a priority policy, or a specified One of the work unit policies, the first message, the second message, or the third message is delivered. The distributed high availability supply method according to claim 6, wherein the supply work management unit records the identification code, parameters, specifications, execution status and time information of the supply work in the first In the category host's database. The distributed high availability supply method according to claim 6, wherein the execution sequence of the process records the order in which the supply stations perform the supply.