CN108234356B - Optimized application resource distribution strategy based on application relation network - Google Patents
Optimized application resource distribution strategy based on application relation network Download PDFInfo
- Publication number
- CN108234356B CN108234356B CN201711231345.8A CN201711231345A CN108234356B CN 108234356 B CN108234356 B CN 108234356B CN 201711231345 A CN201711231345 A CN 201711231345A CN 108234356 B CN108234356 B CN 108234356B
- Authority
- CN
- China
- Prior art keywords
- application
- resource
- service
- distribution strategy
- network
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/76—Admission control; Resource allocation using dynamic resource allocation, e.g. in-call renegotiation requested by the user or requested by the network in response to changing network conditions
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
- H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1034—Reaction to server failures by a load balancer
Abstract
The invention provides an optimized application resource distribution strategy based on an application relation network, which comprises the following steps: calculating an application importance index based on the application service dependency relationship; calculating an application resource distribution strategy under the influence of minimum resource operation and maintenance according to the application importance index; and dynamically adjusting the resource distribution condition of the application according to the application resource distribution strategy. Based on the importance of the application, the invention researches the most reasonable resource supply mode for each application under the condition of limited resources, thereby achieving the purpose of minimizing the influence of the resources on the operation of the whole service system during operation and maintenance. The method gets rid of the existing blind way of allocating resources to the application according to the time sequence and manual operation, and calculates the resources required by the application with different importance based on the target automation of operation and maintenance optimization and dynamically adjusts the resource supply of the application according to the change of the importance of the application.
Description
Technical Field
The invention relates to a resource allocation strategy in the technical field of cloud computing, in particular to an optimized application resource distribution strategy based on an application relation network.
Background
In today's Web and mobile application development process, developers tend to build applications on a service basis, rather than from a wheel of manufacture. In general, these services are referred to as microservices-single use, and API-accessible applications have become the cornerstone for building large applications. The micro-service architecture is a very hot concept in the field of recent software application, and can greatly improve some typical problems encountered by traditional application development. For example, with traditional Monolithic Architecture (Monolithic Architecture) application development systems, such as CRM, ERP, and other large applications, it becomes increasingly difficult for enterprises to update and repair large Monolithic applications as new demands continue to increase. With the development of mobile internet, enterprises are forced to migrate their applications to modern UI interface architectures so as to be compatible with mobile devices, which requires that the enterprises be able to quickly bring on-line application functions.
Based on this demand, the construction mode of complex systems of more and more enterprises and industries gradually changes from the traditional single-body application to the micro-service architecture. The direct consequence of this change is that the business will be transformed from being composed of several relatively independent large monolithic applications to being composed of a large number of interrelated micro-applications. The mode can powerfully improve the updating development mode of the industry business system and promote the explosive growth of the industry business application in an iterative rolling mode in an internet mode. This will also bring endless vitality to the business. A series of technologies related to DevOps, which are gradually becoming hot spots, and the great development of capabilities of automatic deployment and release of applications based on cloud platforms, will all powerfully promote the process.
However, the way such an industry business system is composed of a large number of micro applications also brings new challenges to the operation and maintenance of the system. The increase in applications brings complexity to operation and maintenance. Especially in the cloud services era, these applications are typically deployed on cloud platforms to enable optimized provisioning of resources. But now the main problem is how to allocate resources for these applications reasonably efficiently. In the operation and maintenance process, if we need to perform maintenance operation on a certain resource or perform troubleshooting on a certain resource, how to ensure that the operation has the minimum influence on the whole service system, thereby reducing the operation and maintenance cost and realizing the optimal application resource allocation becomes a problem to be solved urgently in the field.
Through search, the following results are found:
201410189197.8, discloses a method for evaluating resource allocation strategies in cloud computing, which comprises the following steps: dividing a project deployed on a cloud computing platform into a plurality of functional modules, and modeling according to the relationship among the functional modules to generate a workflow model; according to the resource allocation strategy, cloud computing resources are allocated to each functional module, and the operating parameters of each functional module are determined to realize the configuration of the workflow model; mapping the workflow model to a temporal automaton model; mapping the running time error and running time error distribution of the functional module into the time error of the sub-module in the time automaton model; setting a test standard, and calculating a probability value of the time automaton model meeting the test standard as an evaluation result of the resource allocation strategy. The method can automatically analyze whether the user requirements under the current distribution strategy can be met, and quantitatively analyze the reliability of the current distribution strategy, thereby reducing the probability of violating the service level agreement.
The above method still has the following problems:
the method is formed based on functional module flow of projects, and the object-oriented method has the problems that the functional modules are not subjected to weight setting, are equally distributed with resources, are subjected to resource distribution rationality evaluation according to the fault rate of each module, and cannot detect the optimal distribution scheme of the functional modules and the resources;
the method is used for modeling according to a workflow mode, and the problem that the distribution of application resources cannot be dynamically adjusted according to the upgrading change of the application, so that the minimum cost of operation and maintenance resources is achieved.
At present, no explanation or report of the similar technology of the invention is found, and similar data at home and abroad are not collected.
Disclosure of Invention
In view of the above-mentioned shortcomings in the prior art, the present invention provides an optimized application resource distribution strategy based on an application relationship network. The strategy is based on a strategy formed by applying all micro service relations, and a deployment strategy is defined based on an application importance index formed by an application micro service relation network and an application resource relation network.
The invention is realized by the following technical scheme.
An optimized application resource distribution strategy based on an application relation network comprises the following steps:
step S1, calculating the application importance index based on the application service dependency relationship;
step S2, calculating an application resource distribution strategy under the influence of minimum resource operation and maintenance according to the application importance index;
and step S3, dynamically adjusting the resource distribution situation of the application according to the application resource distribution strategy.
Preferably, step S1 includes the following sub-steps:
step S1.1, applying weight analysis: obtaining the weight of the related application through service calling between the applications;
step S1.2, evolution of application importance: the importance indicators for all applications in the system are periodically recalculated after each new application is deployed or after the old application is taken offline.
Preferably, step S1.1 adopts a multiple-link application network node weight calculation method, including the following processes:
G(E,V)
in the formula, E represents a node relation set, and V represents a node set;
Ef(u)＝{v|v∈Follower(u)∩Response(u)＞ε}
in the formula, epsilon is a non-negative constant threshold value and represents a degree threshold value fed back by a reference service node V of a node u to the node u, and an application node which exceeds the threshold value and belongs to the node u is effective application;
define 3 node weight IRL (U) generated by Link relationshipi) The calculation method is shown as the following formula:
IRL(Ui)＝δN+(1-δ)∑Uj∈Follower(ui)IRL(ui)L(ui)
in the formula, IRL (U)i) Represents a node UiNode weight, Follower, generated by the link relationi) Is a node UiSet of all associated services, L (u)i) Is a node UiThe number of associated services, δ is the damping coefficient between 0 and 1, and N is the total number of nodes in the network map.
Preferably, in step S2, specifically, the method includes:
for a specified resource R, analyzing applications which are directly influenced by R in operation and maintenance, and setting the applications as DE (R); finding all affected applications according to the application service dependency relationship graph, wherein the range of all affected applications is a subgraph in the whole application service dependency relationship graph, and the weight of all applications in the subgraph is defined as an influence factor F (R) of a resource R;
assuming that the total resource amount that can be operated and maintained in the system is R1 to Rn, the total influence factor of the operation and maintenance on the whole industry business is
The optimized application resource distribution strategy should be such that
The minimization is achieved; to do this, we need to develop an innovative calculation method a in the subject to enable
A minimized application resource distribution strategy is achieved;
the innovative calculation method A specifically comprises the following steps:
step SA.1, grading labels according to the conditions of the resources, such as the conditions of a memory, a CPU, a hard disk, a network card and the like;
step SA.2, distributing the resources in a label mode according to the importance of the application and the quality of the resources;
step SA.3, total influence factor is caused by elastic expansion and transverse expansion and resource allocation according to the labeling
A minimization is achieved.
Preferably, in step S3, specifically, the method includes:
after the application importance changes each time, the system needs to calculate an optimal distribution model of the application resources, and then automatically adjusts the resource supply condition of the corresponding application according to the conclusion of the optimal distribution model; the optimization distribution model is obtained by adopting a calculation method B, and the method comprises the following steps:
SB.1, recalculating the application importance weight according to the application service dependency relationship diagram;
step SB.2, according to the calculation method A, minimizing the influence factor of the total resource operation and maintenance service;
and step SB.3, performing optimal distribution model calculation of the application resources in an elastic expansion and/or clustering mode.
Preferably, in steps S1 to S3, a system operation and maintenance system is constructed according to the infrastructure provided by the cloud product, the application automation deployment and the service automation delivery, and the system operation and maintenance system includes an application service dependency relationship management module, an application resource distribution policy management module and an application resource dynamic scheduling module; wherein:
the application service dependency relationship management module is used for generating an application service dependency network and calculating the importance and weight of the application;
the application resource management module is used for generating an application resource dependent network;
the application resource distribution strategy management module is used for calculating and generating an optimal distribution strategy of the application resources;
and the application resource dynamic scheduling module is used for rescheduling according to the optimal distribution strategy of the application resources and distributing the application resources.
Preferably, the method further comprises the following steps:
step S4, including any one or more of the following processes:
-upgrading the application;
migrating and deploying the application, and distributing important applications into different physical nodes by calculating the weight and the dependence of the application to realize high availability of the application;
-applying operation and maintenance automation, including any one or any plurality of:
fault self-isolation, when the application in operation has a fault, isolating the fault application and timely informing the related service to stop or carry out fault transfer;
the high availability self-balancing method has the advantages that high availability self-balancing is achieved, when important applications are deployed in different physical nodes and the physical nodes break down, the applications inform the cloud platform to switch the high availability backup of the applications to the available physical nodes, and the high availability self-balancing is achieved.
Preferably, the application is upgraded by adopting a replacement mode, that is, different service instances are deployed in advance, and then the application is upgraded by an application relationship dependent replacement mode.
Preferably, before upgrading the application, the functional and non-functional tests are performed according to specific conditions and then replaced, and after the upgrade fails, the rollback is performed for rapid degradation.
Preferably, the application relationship dependency replacement mode is specifically: and (3) replacing the access mode between the associated services by applying, namely switching the access address of the old service to be accessed to the new service instance through the modes of service registration and service discovery.
Preferably, the calculating of the applied weight and dependency specifically includes the following steps:
step a, constructing an application service dependence network of an industry business system according to an application service incidence relation provided by a cloud product;
b, for a simple industry business system application service dependent network, weight calculation can be carried out through the calling times of application, and for a complex industry business system application service dependent network, importance applied in a business system can be added for weight analysis;
and c, performing pre-allocation calculation on the application resources by using the optimized distribution model obtained by the calculation method of the optimized distribution model.
Preferably, in the process of migrating and deploying the application, an application weight self-balancing mechanism is set, and the mechanism specifically includes: every time when the application of the business system is newly added or deleted, the system should be able to perform a weight adjustment mechanism according to the application service dependency relationship network.
Compared with the prior art, the invention has the following beneficial effects:
1. based on the importance of the application, the invention researches the most reasonable resource supply mode for each application under the condition of limited resources, thereby achieving the purpose of minimizing the influence of the resources on the operation of the whole service system during operation and maintenance.
2. The method gets rid of the existing blind way of allocating resources to the application according to the time sequence and manual operation, and calculates the resources required by the application with different importance based on the target automation of operation and maintenance optimization and dynamically adjusts the resource supply of the application according to the change of the importance of the application.
Drawings
Other features, objects and advantages of the invention will become more apparent upon reading of the detailed description of non-limiting embodiments with reference to the following drawings:
FIG. 1 is a schematic diagram of an application service dependent network;
FIG. 2 is a schematic diagram of an application resource dependent network;
FIG. 3 is a system operation and maintenance architecture diagram;
FIG. 4 is a diagram of application weight assignment physical nodes;
fig. 5 is an operation and maintenance center system architecture diagram.
Detailed Description
The following examples illustrate the invention in detail: the embodiment is implemented on the premise of the technical scheme of the invention, and a detailed implementation mode and a specific operation process are given. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the inventive concept, which falls within the scope of the present invention.
Examples
The embodiment provides an optimized application resource distribution strategy based on an application relationship network, which comprises the following steps:
step S1, calculating the application importance index based on the application service dependency relationship
According to the existing application service dependency relationship diagram, each application in an industry business system can simply mark an application importance index through the called times of the application in the diagram, and the method is similar to the method for marking the value degree of a paper through the quoted times of the paper in scientific research work.
Furthermore, importance of the application in different levels on business can be considered, or importance of different degrees of service invocation can be considered, and therefore application weight analysis can be added when the importance index of the application in the graph is calculated.
1) Applying weight analysis
Service calling between applications has different calling degrees, so that the importance degrees of different applications in the system are not consistent, and the weight of the related applications can be obtained according to the calling relationship between the applications.
The calculation mode of the application weight can adopt a reference counting mode or a multi-dependent path mode for calculation. In this embodiment, the calculation method may be performed by referring to a calculation method of the weight of the application network node with multiple links.
The basic principle of the method is as follows.
G(E,V)
in the formula, E represents a node relationship set, and V represents a node set.
Defining 2 a set of valid service references ef (u) as shown in the following equation:
Ef(u)＝{v|v∈Follower(u)∩Response(u)＞ε}
in the formula, epsilon is a non-negative constant threshold value, which represents a degree threshold value fed back to the node u by the reference service node v of the node u, and an application node which exceeds the threshold value and belongs to the node u can be calculated as an effective application.
Define 3 node weights IRL (U) generated by the link relationi) The calculation method is shown as the following formula:
IRL(Ui)＝δN+(1-δ)∑Uj∈Follower(ui)IRL(ui)L(ui)
in the formula, IRL (U)i) Represents a node UiNode weight, Follower, generated by the link relationi) Is node uiSet of all associated services, L (u)i) Is node uiThe number of associated services, δ is the damping coefficient between 0 and 1, and N is the total number of nodes in the network map.
2) Evolution of application importance
The importance of an application is not a constant in a business system. As industry business develops, more and more applications are being developed. In this process, each application may have a lifecycle from heavy usage to gradual obsolescence or renewal. The number of times each existing application is referenced, and thus the weight, may be changed after a new application is deployed and used for a period of time. To simplify the model, it can be preliminarily assumed that each application can clarify its business importance from the beginning of deployment. It is then necessary to recalculate the importance indicators for all applications in the system after each new application is deployed or after the old application is taken offline.
Further, the service importance of the application is determined according to the access condition of the user after a period of use. In this case, the importance indicators for all applications need to be recalculated periodically.
Step S2, according to the application importance index, calculating the application resource distribution strategy under the influence of the minimized resource operation and maintenance
From the aspect of operation and maintenance, the operation and maintenance can be simply understood as upgrading maintenance or troubleshooting work on resources. In this case, a resource will stop working, possibly causing the working of the application associated with this resource to be affected. According to the application resource dependency network constructed by the aforementioned work, for a given resource R, applications that R will directly affect during operation and maintenance can be analyzed, and these applications are set to be de (R). And finding all the affected applications from the application service dependency graph, wherein all the affected application ranges are finally a subgraph in the whole application service dependency graph. The weight of all applications in this subgraph will be defined as the impact factor f (R) of resource R.
Assuming that the total resource amount that can be operated and maintained in the system is R1 to Rn, the total influence factor of the operation and maintenance on the whole industry business is
. The optimized application resource distribution strategy should be such that
A minimization is achieved. To do this, we need to develop a calculation method A in the subject to make
A minimized application resource distribution strategy is achieved.
The innovative calculation method A comprises the following steps:
1) grading labels according to the conditions of the resources, such as the conditions of a memory, a CPU, a hard disk, a network card and the like;
2) distributing in a label mode according to the importance of the application and the quality of the resource;
3) the total influence factor is realized by elastic expansion and transverse expansion and resource allocation according to the labeling
A minimization is achieved.
Step S3, dynamically adjusting the resource distribution situation of the application according to the application resource distribution strategy
On the basis of the calculation in step S1 and step S2, the present embodiment needs to implement dynamic adjustment of application resources in the cloud platform. That is, after each change of the application importance occurs, the system needs to calculate an optimal distribution model of the application resources, and then automatically adjust the resource supply condition of the corresponding application according to the conclusion of the model, so that the actual system achieves the best effect.
The method for calculating the optimal distribution model of the application resources comprises the following steps:
1) recalculating the application importance weight according to the application service dependency relationship diagram;
2) minimizing the influence factor of the total resource operation and maintenance service according to the calculation method A;
3) and performing optimal distribution model calculation of application resources in modes of elastic expansion, clustering and the like.
For this reason, the system needs to implement the architecture in the system operation and maintenance architecture diagram, and the basic idea of the architecture is to build the system operation and maintenance architecture on the basis of the infrastructure, application automation deployment and service automation delivery provided by the existing cloud product. The system operation and maintenance system is composed of four modules of application service dependency relationship management, application resource distribution strategy management and application resource dynamic scheduling.
In this step, the system operation and maintenance system functions as: and generating an application service dependency relationship network and an application resource dependency network through calculation, and generating an optimal distribution strategy of application resources through calculation, so that self-balancing and rescheduling of the application resources are performed when the operation and maintenance of the system are upgraded.
Further, the application service dependency relationship management module is used for generating an application service dependency network and calculating the importance and weight of the application;
the application resource management module is used for generating an application resource dependent network;
the application resource distribution strategy management module is used for calculating and generating an optimal distribution strategy of the application resources;
and the application resource dynamic scheduling module is used for rescheduling according to the optimal distribution strategy of the application resources and distributing the application resources.
Step S4, operation and maintenance extension change
With the research of the core innovation point of the embodiment, a reasonable matching mode of the application and the resource can be constructed. On the basis, more complex operation and maintenance behaviors can be further considered, such as upgrading and updating of applications, high availability of applications, automation of operation and maintenance of applications and the like.
1) Application upgrade and replacement
The application upgrading process and the corresponding upgrading operation and the traditional operation and maintenance process change under the micro-service architecture are mainly embodied in the following aspects:
the influence of the upgrade of the application on the whole system can be controlled and isolated in a relatively controllable interval, and the upgrade of a single application does not influence the use of the whole system;
ii, more application upgrading is carried out in a replacement mode, different service instances can be deployed firstly, and then upgrading is carried out in a mode that application relation depends on replacement; the application relationship dependency replacement mode specifically comprises the following steps: and switching the access address of the old service to be accessed to the new service instance by means of service registration and service discovery by applying access mode replacement between the associated services.
And iii, before application upgrading, functional and non-functional tests can be carried out and then replaced according to specific conditions, and the rollback can be rapidly degraded after upgrading failure, so that the influence on the application after upgrading failure is controllable.
2) Application of high availability
In a traditional architecture, an application is deployed in a virtual machine or a physical machine, and scheduling of bottom layer resources is difficult to realize. The bottom layer resources occupied by each application are limited, and to realize high availability of the applications, high availability among the resources must be realized. Under the new framework, resources used by different applications can be allocated, so that the migration and corresponding deployment of the applications can be realized.
By means of the applied weight and the dependency calculation, important applications can be dispersed into different physical nodes, and reliable and high availability of the applications is achieved, and reference is made to fig. 4. The calculation method of the applied weight and the dependency calculation specifically comprises the following steps:
1) constructing an application service dependent network of an industry business system according to the association relation of the application services provided by the product;
2) a simple industry business system application service dependent network can perform weight calculation through the calling times of the application,
the complicated industry business system application service dependence network can be added with the importance applied in the business system to carry out weight analysis;
3) the application resource dependent calculation is the pre-allocation calculation of the application resources by the optimized distribution model obtained by the calculation method of the optimized distribution model.
As shown in the architecture diagram of fig. 5: different applications are deployed on the cloud platform, external physical resources are managed by the cloud platform, and resources including computing, network and storage are distributed to the applications by the cloud platform.
In a large complex environment, there are often a large number of application add/delete operations, and different applications enter or leave the system, which will cause a change in the relationship between each application and service in the system, so the system needs to have an application weight self-balancing mechanism to gradually adjust the weight and topology relationship between each application and service. The application weight self-balancing mechanism is specifically a mechanism for adjusting the weight according to an application service dependency relationship network when the application of the service system is newly added or deleted and modified.
3) Application operation and maintenance automation
The number of applications under the micro-service architecture is increased, and the upgrading and maintenance of the applications need to be carried out automatically depending on the capability provided by the cloud platform. Such automated operation and maintenance can be embodied in two capabilities:
firstly, fault self-isolation:
because of the large number of applications in the environment, once some running applications fail, the applications need to be isolated and related services need to be notified to stop or perform failover in time.
Two, high availability self-balancing
The high availability of the application needs to deploy important applications in different physical nodes, and once a physical node fails, the application needs to inform the cloud platform to switch the high availability backup of the application to the available node, so as to achieve the high availability self-balancing capability.
The embodiment provides an optimized application resource allocation strategy scheme adaptive to resource operation and maintenance. The principle is as follows:
if the resources of the whole system are enough, each application should be deployed on a plurality of unrelated resources, so that when any one resource is in an operation and maintenance state (upgrade update or fault maintenance), the operation of the whole industry business system is not affected at all. This situation is referred to as an ideal application resource allocation policy. But real-world systems tend to be resource inefficient. In this case, it is important that only one optimized application resource allocation policy is set, and applications should be deployed on multiple unrelated resources as much as possible, so that when any resource is in an operation and maintenance state, the whole industry service system is affected as little as possible.
From the above, it can be seen that the importance of the application is the key to deducing this optimization strategy. In the industrial scale micro application system, the applications can be set to be mostly related to each other through services. The industry business system is an application service network constructed by application service calling relations. Without loss of generality, the more a service of an application is referenced by other applications in this network, the more important the application can be considered. Based on the foregoing settings, it is believed that such applications should be allocated for deployment on multiple unrelated resources to ensure reliability and serviceability of the system.
According to this setting, the present embodiment is mainly divided into three links:
1. how the importance of the application is identified;
2. how to define resource distribution strategies to achieve optimization in terms of application importance
3. How to implement the floor of this optimization strategy through automated deployment of applications.
Meanwhile, with the development of an industrial business system, there may be a constantly changing process for the importance of each micro application. Therefore, the above three links of work may need to be repeatedly re-executed so that the whole industry business system is always in an optimized state.
The innovation points of the embodiment mainly include the following three aspects:
1. calculating an application importance index;
2. calculating an application resource distribution strategy under the influence of minimum resource operation and maintenance according to the application importance index;
3. and dynamically adjusting the resource distribution condition of the application according to the application resource distribution strategy.
The foregoing description of specific embodiments of the present invention has been presented. It is to be understood that the present invention is not limited to the specific embodiments described above, and that various changes and modifications may be made by one skilled in the art within the scope of the appended claims without departing from the spirit of the invention.
Claims (9)
1. An optimized application resource distribution strategy based on an application relation network is characterized by comprising the following steps:
step S1, calculating the application importance index based on the application service dependency relationship;
step S2, calculating an application resource distribution strategy under the influence of minimum resource operation and maintenance according to the application importance index;
step S3, dynamically adjusting the resource distribution condition of the application according to the application resource distribution strategy;
step S2 specifically includes: for the designated resource R, the application which is directly influenced by R in operation and maintenance is analyzed, and the applications are set
With the formula DE (R); finding all affected applications according to the application service dependency relationship graph, wherein the range of all affected applications is a subgraph in the whole application service dependency relationship graph, and the weight of all applications in the subgraph is defined as an influence factor F (R) of a resource R;
assuming that the total resource amount that can be operated and maintained in the system is R1 to Rn, the total influence factor of the operation and maintenance on the whole industry business is
The optimized application resource distribution strategy should be such that
The minimization is achieved; for this purpose, the following calculation method A is employed to enable
The minimized application resource distribution strategy is achieved: step SA.1, grading labels according to the quality of resources; step SA.2, according to the importance of the application andthe quality condition of the resource is distributed to the resource in the form of a label;
step SA.3, total influence factor is generated by elastic expansion, transverse expansion and/or resource allocation according to label
A minimization is achieved.
2. The optimized application resource distribution strategy based on application relationship network of claim 1, wherein step S1 comprises the following sub-steps:
step S1.1, applying weight analysis: obtaining the weight of the related application through service calling between the applications;
step S1.2, evolution of application importance: periodically re-executing after each new application is deployed or after an old application is taken offline
The importance indicators for all applications in the new computing system.
3. The optimized application resource distribution strategy based on application relationship network of claim 2, wherein step S1.1 adopts a multi-link application network node weight calculation method, comprising the following processes:
definition 1, and the application network directed graph is G, as shown in the following formula:
g ═ E, V, where E represents a set of node relationships and V represents a set of nodes;
definition 2, set of valid service references ef (u), as shown in the following equation: in the formula of ef (u) { V | V ∈ follower (u) # response (u) > epsilon }, epsilon is a non-negative constant threshold value representing a degree threshold of feedback of the reference service node V of the node u to the node u, and an application node which exceeds the threshold value and belongs to the node u is a valid application;
defining 3, a node weight irl (ui) generated by the link relationship, which is calculated as follows: in the formula irl (Ui) ═ δ N + (1- δ) ∑ Uj ∈ follower (Ui) irl (Ui) l (Ui), irl (Ui) represents a node weight value generated by a link relationship of the node Ui, follower (Ui) is a set of all associated services of the node Ui, l (Ui) is the number of associated services of the node Ui, δ is a damping coefficient between 0 and 1, and N is the total number of nodes in the network graph.
4. The optimized application resource distribution strategy based on application relationship network as claimed in claim 1, wherein the step S3 specifically comprises: after the application importance changes each time, the system needs to calculate an optimal distribution model of the application resources, and then automatically adjusts the resource supply condition of the corresponding application according to the conclusion of the optimal distribution model; the optimization distribution model is obtained by adopting a calculation method B, and the method comprises the following steps:
SB.1, recalculating the application importance weight according to the application service dependency relationship diagram; step SB.2, according to the calculation method A, minimizing the influence factor of the total resource operation and maintenance service; and step SB.3, performing optimal distribution model calculation of the application resources in a flexible extension and/or clustering mode.
5. The optimized application resource distribution strategy based on application relationship network of claim 4, wherein in steps S1 to S3, a system operation and maintenance system is constructed according to the infrastructure, application automation deployment and service automation delivery provided by the cloud product, the system operation and maintenance system comprises an application service dependency relationship management module, an application resource distribution strategy management module and an application resource dynamic scheduling module; wherein:
the application service dependency relationship management module is used for generating an application service dependency network and calculating the importance and weight of the application;
the application resource management module is used for generating an application resource dependent network; the application resource distribution strategy management module is used for calculating and generating an optimal distribution strategy of the application resources; and the application resource dynamic scheduling module is used for rescheduling according to the optimal distribution strategy of the application resources and distributing the application resources.
6. The optimized application resource distribution strategy based on application relation network according to any one of claims 1 to 5, characterized by further comprising the following steps:
step S4, including any one or more of the following processes:
-upgrading the application;
migrating and deploying the application, and distributing important applications into different physical nodes by calculating the weight and the dependence of the application to realize high availability of the application;
-applying operation and maintenance automation, including any one or any plurality of: fault self-isolation, when the application in operation has a fault, isolating the fault application and timely informing the related service to stop or carry out fault transfer; the high availability self-balancing method has the advantages that high availability self-balancing is achieved, when important applications are deployed in different physical nodes and the physical nodes break down, the applications inform the cloud platform to switch the high availability backup of the applications to the available physical nodes, and the high availability self-balancing is achieved.
7. The optimized application resource distribution strategy based on application relation network as claimed in claim 6, wherein the application is upgraded and further comprises any one or more of the following features: upgrading the application by adopting a replacement mode, namely deploying different service instances in advance, and upgrading the application in a mode of application relationship dependent replacement; before upgrading the application, firstly, performing functional and non-functional tests according to specific conditions and then replacing, and backing to perform rapid degradation after upgrading fails; the application relationship dependency replacement mode is specifically replacement of an access mode between services associated with an application, that is, an access address of an old service to be accessed is switched to a new service instance through a mode of service registration and service discovery.
8. The optimized application resource distribution strategy based on application relationship network as claimed in claim 6, wherein the calculating of the application weight and dependence comprises the following steps:
step a, constructing an application service dependence network of an industry business system according to an application service incidence relation provided by a cloud product;
b, carrying out weight calculation through the calling times of the application, or carrying out weight analysis by adding the importance of the application in a service system;
and c, performing pre-allocation calculation on the application resources through the optimized distribution model obtained by the calculation method B.
9. The optimized application resource distribution strategy based on application relationship network as claimed in claim 6, wherein in the process of migrating and deploying the application, an application weight self-balancing mechanism is set, and the mechanism specifically comprises: whenever there is a new or deleted modification of the application of the business system, the system should be able to perform a weight adjustment mechanism according to the application service dependency network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711231345.8A CN108234356B (en) | 2017-11-29 | 2017-11-29 | Optimized application resource distribution strategy based on application relation network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201711231345.8A CN108234356B (en) | 2017-11-29 | 2017-11-29 | Optimized application resource distribution strategy based on application relation network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108234356A CN108234356A (en) | 2018-06-29 |
| CN108234356B true CN108234356B (en) | 2021-07-06 |
Family
ID=62652968
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201711231345.8A Active CN108234356B (en) | 2017-11-29 | 2017-11-29 | Optimized application resource distribution strategy based on application relation network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108234356B (en) |
Families Citing this family (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109586952B (en) * | 2018-11-07 | 2022-04-12 | 广州虎牙信息科技有限公司 | Server capacity expansion method and device |
| EP3722944A1 (en) * | 2019-04-10 | 2020-10-14 | Juniper Networks, Inc. | Intent-based, network-aware network device software-upgrade scheduling |
| CN110738431B (en) * | 2019-10-28 | 2022-06-17 | 北京明略软件系统有限公司 | Method and device for allocating monitoring resources |
| CN110943867B (en) * | 2019-12-05 | 2022-08-16 | 上交所技术有限责任公司 | System and method for deducing application architecture information through network relationship |
| CN111324471B (en) * | 2020-01-22 | 2023-07-21 | 远景智能国际私人投资有限公司 | Service adjustment method, device, equipment and storage medium |
| WO2024032239A1 (en) * | 2022-08-12 | 2024-02-15 | 华为云计算技术有限公司 | Application scheduling method, cloud service platform, and related device |
| CN115484159B (en) * | 2022-09-16 | 2023-06-20 | 建信金融科技有限责任公司 | Network demand and resource supply determining system and method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102214212A (en) * | 2011-05-20 | 2011-10-12 | 西北工业大学 | Method for ordering microblog network node weights based on multi-link |
| CN102413186A (en) * | 2011-12-02 | 2012-04-11 | 北京星网锐捷网络技术有限公司 | Resource scheduling method and device based on private cloud computing, and cloud management server |
| CN105183610A (en) * | 2015-09-22 | 2015-12-23 | 浪潮集团有限公司 | Cloud data center service monitoring system and method based on resource dependency relationship |
| CN106020927A (en) * | 2016-05-05 | 2016-10-12 | 中国人民解放军国防科学技术大学 | Universal method for task scheduling and resource configuration in cloud computing system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9184982B2 (en) * | 2013-01-15 | 2015-11-10 | Hewlett-Packard Development Company, L.P. | Balancing the allocation of virtual machines in cloud systems |
| KR102012259B1 (en) * | 2013-08-21 | 2019-08-21 | 한국전자통신연구원 | Method and apparatus for controlling resource of cloud virtual base station |
-
2017
- 2017-11-29 CN CN201711231345.8A patent/CN108234356B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102214212A (en) * | 2011-05-20 | 2011-10-12 | 西北工业大学 | Method for ordering microblog network node weights based on multi-link |
| CN102413186A (en) * | 2011-12-02 | 2012-04-11 | 北京星网锐捷网络技术有限公司 | Resource scheduling method and device based on private cloud computing, and cloud management server |
| CN105183610A (en) * | 2015-09-22 | 2015-12-23 | 浪潮集团有限公司 | Cloud data center service monitoring system and method based on resource dependency relationship |
| CN106020927A (en) * | 2016-05-05 | 2016-10-12 | 中国人民解放军国防科学技术大学 | Universal method for task scheduling and resource configuration in cloud computing system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108234356A (en) | 2018-06-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108234356B (en) | Optimized application resource distribution strategy based on application relation network | |
| US11567795B2 (en) | Minimizing impact of migrating virtual services | |
| Ismaeel et al. | Proactive dynamic virtual-machine consolidation for energy conservation in cloud data centres | |
| CN109643090B (en) | Method, system and apparatus for dynamically facilitating management of M: N work configuration systems | |
| US10810050B2 (en) | Virtual systems management | |
| Ardagna et al. | SLA based resource allocation policies in autonomic environments | |
| Sharif et al. | Fault‐tolerant with load balancing scheduling in a fog‐based IoT application | |
| US20070143767A1 (en) | Method, system and computer program for dynamic resources allocation | |
| US11765031B2 (en) | System and method of strategy-driven optimization of computer resource configurations in a cloud environment | |
| CN105577475A (en) | Automatic performance test system and method | |
| TWI725744B (en) | Method for establishing system resource prediction and resource management model through multi-layer correlations | |
| KR102607808B1 (en) | Dynamic reallocating resources for optimized job performance in distributed heterogeneous computer system | |
| US20190066016A1 (en) | Benchmarking for automated task management | |
| CN111343219A (en) | Computing service cloud platform | |
| US11212173B2 (en) | Model-driven technique for virtual network function rehoming for service chains | |
| Meng et al. | Service-oriented reliability modeling and autonomous optimization of reliability for public cloud computing systems | |
| Alyas et al. | Resource Based Automatic Calibration System (RBACS) Using Kubernetes Framework. | |
| WO2020206699A1 (en) | Predicting virtual machine allocation failures on server node clusters | |
| Santos et al. | Diktyo: Network-aware scheduling in container-based clouds | |
| Sharma et al. | Virtual machine migration for green cloud computing | |
| CN112580816A (en) | Machine learning training resource management | |
| Daradkeh et al. | Modeling and optimizing micro-service based cloud elastic management system | |
| Lin | Scheduling efficiency on correlated parallel machine scheduling problems | |
| Yousaf et al. | RAVA—Resource aware VNF agnostic NFV orchestration method for virtualized networks | |
| WO2023154051A1 (en) | Determining root causes of anomalies in services |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |