WO2018228323A1

WO2018228323A1 - Service level control method and system for on-line service system, and readable storage medium

Info

Publication number: WO2018228323A1
Application number: PCT/CN2018/090613
Authority: WO
Inventors: 刘东辉; 王俊杰; 褚建辉
Original assignee: 广东神马搜索科技有限公司
Priority date: 2017-06-13
Filing date: 2018-06-11
Publication date: 2018-12-20
Also published as: CN107317702A

Abstract

Disclosed in the present invention are a service level control method and system for an on-line service system, and a readable storage medium. The service level control method comprises: a service degradation step, when it is determined that the system state of an on-line service system has been degraded, lowering the quality of service of the on-line service system to a level corresponding to the current system state; and a service recovery step, when it is determined that the system state has been raised, suspending recovering the quality of service to a level corresponding to the current system state. Thus, by taking a fast-degradation and slow-recovery level control policy, the present invention may avoid adverse influence caused by the jitter and repeatability of the system state of the on-line service system, and may also allow more time for the on-line service system to further recover the system state thereof.

Description

Service level control method and system for online service system, readable storage medium

Technical field

The present invention relates to the field of online services, and in particular, to a service level control method and corresponding system for an online service system, and a readable storage medium.

Background technique

A typical feature of an online service system based on Internet-provided services is that traffic is unstable. In addition to the obvious peak-to-valley distribution over time, the online service system's access traffic is also affected by hot events or special events. In order to ensure the stability of the services provided by the online service system, it is often necessary to reserve sufficient service capabilities for the online service system. However, the reserved resources are only designed based on the expected traffic peaks. The system still has a fatal risk in the case of a partial module exception of the service system (for example, the cache module is invalid), an upstream or downstream service exception or a malicious attack. In order to cope with these situations, the usual method is to downgrade the service, that is, to release the resources by sacrificing a certain quality of service, thereby ensuring the basic service capability of the system.

The existing downgrade scheme usually sets a threshold in advance. When the health state of the service system is lower than the threshold, the system switches to the degraded mode and provides only the most basic service. When the health status of the service system rises above the threshold, the system Restore to normal service status. The existing downgrade scheme is simple to implement, but due to the instability of the service system, the system is likely to cause jitter near the threshold, and the system is in a high service pressure state for a long time.

Therefore, there is a need for a new service level control method and system for an online service system to address at least one of the above problems.

Summary of the invention

The main object of the present invention is to provide a service level control method and system and a readable storage medium for a relatively stable online service system.

According to an aspect of the present invention, a service level control method for an online service system is provided, wherein a quality of service provided by an online service system is divided into a plurality of levels according to a system state, the method comprising: a service degradation step, in determining an online service When the system state of the system decreases, the service quality of the online service system is lowered to a level corresponding to the current system state; and the service recovery step, when determining that the system state is rising, suspends the restoration of the service quality to a level corresponding to the current system state. .

Therefore, by adopting a level-lowering and slow recovery level control strategy, it is possible to prevent the adverse effects caused by the jitter and repeatability of the system state of the online service system, and also provide more time for the online service system to further recover. Its system status.

Preferably, the service recovery step may include: when determining that the system status is rising, restoring the quality of service to a level lower than a level corresponding to the current system state. Therefore, while improving the quality of service, a certain service capability can be reserved for the online service system to continue to restore its system state.

Preferably, the system state is characterized by a feature value of the current service capability, the quality of service is divided into a plurality of levels corresponding to the plurality of feature value thresholds, and wherein the service downgrading step may include: determining that the system state is degraded such that the current feature When the value satisfies the first eigenvalue threshold, the service quality of the online service system is immediately lowered to a level corresponding to the first eigenvalue threshold; and the service recovery step may include: determining that the current eigenvalue satisfies the second trait when the system state is raised When the value is thresholded, the quality of service is suspended to a level corresponding to the second eigenvalue threshold. Thus, by characterizing the system state by the feature values, the service degradation step and the service recovery step can be performed simply based on the relationship between the current feature value and the feature value threshold.

Preferably, the service recovery step may include restoring the quality of service to a level corresponding to the second feature value threshold when determining that the system state is rising such that the current feature value satisfies the second feature value threshold for a predetermined length of time.

Therefore, for the phenomenon caused by the up and down jitter of the system state, the current feature value can be made to satisfy the second eigenvalue threshold for a predetermined period of time as an additional condition. This can prevent the online service system from being under high stress, and can also provide additional recovery time for the online service system, which helps the system state to recover quickly.

Preferably, the service recovery step may include: restoring the quality of service to a second eigenvalue threshold when determining that the system state is rising such that the current eigenvalue satisfies a third eigenvalue threshold that is higher than the second eigenvalue threshold quality level Level. Therefore, while improving the quality of service, a certain service capability can be reserved for the online service system to continue to restore its system state.

Preferably, the service recovery step may include: when it is determined that the system state is raised such that the current feature value satisfies the second feature value threshold corresponding to the optimal quality of service level, only the current feature value further satisfies the better service than the second feature value representation Quality of service is returned to the best quality of service level when the best service regression is achieved.

Therefore, it can be determined whether the current eigenvalue satisfies the optimal service regression value, whether the system state of the online service system is stable at a state suitable for providing the best quality of service level, and the system state is determined to be stable in providing quality of service. The quality of service is restored to the best quality of service level when the status of the level is reached.

The feature value of the current service capability may preferably be the queue length currently required to be processed by the online service system.

According to another aspect of the present invention, there is also provided an online quality of service control system comprising a distributed service system for providing an online service and one or more quality control servers, wherein one or more quality control servers are used Obtaining a feature value for characterizing the current service capability of the distributed service system; when the current feature value indicates that the system state of the distributed service system is degraded, immediately lowering the service quality of the distributed service system to a level corresponding to the current system state And suspending the restoration of the quality of service to a level corresponding to the current system state when the current eigenvalue characterizes the state of the system.

Preferably, the one or more quality control servers are further configured to: when the current feature value indicates that the system state rises, restore the quality of service to a level lower than a level corresponding to the current system state.

Preferably, the quality of service is divided into a plurality of levels corresponding to the plurality of feature value thresholds, and wherein the one or more quality control servers are further configured to: determine that the current state value meets the first feature value threshold by determining that the system state is degraded Immediately, the quality of service of the online service system is lowered to a level corresponding to the first feature value threshold; and when it is determined that the system state is increased such that the current feature value satisfies the second feature value threshold, the quality of service is suspended to be restored to the second The level corresponding to the eigenvalue threshold.

Preferably, the one or more quality control servers are further configured to restore the quality of service to a level corresponding to the second feature value threshold when it is determined that the system state is raised such that the current feature value satisfies the second feature value threshold for a predetermined length of time.

Preferably, the one or more quality control servers are further configured to restore the quality of service to the first time when it is determined that the system state rises such that the current feature value satisfies a third eigenvalue threshold that is higher than the second eigenvalue threshold quality level. The level corresponding to the second eigenvalue threshold.

Preferably, the one or more quality control servers are further configured to: when the system state is raised such that the current feature value satisfies the second feature value threshold corresponding to the optimal quality of service level, only the current feature value further satisfies the second feature The quality of service is restored to the best quality of service level when the value represents the best service regression for better service quality.

Preferably, the feature value of the current service capability is the queue length currently required to be processed by the distributed service system.

Preferably, the feature value of the current service capability is a queue length currently required by one or more service modules of the distributed service system, and the one or more quality control servers adjust the quality of service of the one or more service modules according to the feature value. level.

In summary, the service level control method and system and the readable storage medium of the online service system of the present invention mainly include a service degradation step and a service recovery step, and the service degradation step may be performed immediately in response to a system state drop of the online service system. The recovery step is to suspend recovery in response to a rise in the system state of the online service system. Therefore, by adopting a level control strategy of fast-downgrading and slow recovery, it is possible to prevent the adverse effects caused by the jitteriness and repeatability of the system state of the online service system. Moreover, the slow recovery can also provide more time for the online service system to further restore its system state.

DRAWINGS

The above and other objects, features, and advantages of the present invention will become more apparent from the aspects of the embodiments of the invention. The same parts.

FIG. 1 is a schematic flowchart showing a service level control method of an online service system according to an embodiment of the present invention; FIG.

2 is a schematic diagram showing a downgrade trigger and upgrade recovery using a specific embodiment of the present invention;

FIG. 3 is a functional block diagram showing the structure of an online quality of service control system 300 in accordance with an embodiment of the present invention; FIG.

4 is a functional block diagram showing the structure of an online quality of service control system 300 in accordance with another embodiment of the present invention.

detailed description

Preferred embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While the preferred embodiment of the present invention is shown in the drawings, it is understood that the invention may be embodied in various forms and not limited by the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

The online service system described in the present invention mainly refers to an Internet application system for providing online services, in particular, a distributed online service system including a plurality of servers and divided into a plurality of modules, such as an online information recommendation service system, an online search service system, and the like. Wait.

The quality of service provided by the online service system may be divided into multiple (two or more) levels according to the state of the system, wherein the system state of the online service system may indicate the service capability or health status of the online service system. The quality of service at each of the divided levels matches the level of service that the online service system can provide or is adapted to provide in the system state corresponding to that level. In the present invention, an increase in the state of the system indicates an increase in the service capability or health state of the online service system, and a decrease in the state of the system indicates a decline in the service capability or health state of the online service system.

FIG. 1 is a schematic flowchart showing a service level control method of an online service system according to an embodiment of the present invention.

As shown in FIG. 1, the service level control method of the online service system of the present invention mainly includes a service downgrading step (step S110) and a service recovery step (step S120).

In step S110, when it is determined that the system state of the online service system is degraded, the quality of service of the online service system is lowered to a level corresponding to the current system state.

As described above, the system status may indicate the service capability or the health status of the online service system. When the system status rises, it may indicate that the service capability of the online service system is enhanced or the health status is improved. When the system status is decreased, the online service system may be indicated. Service capacity declines or health status deteriorates. Therefore, it is possible to judge the rise or fall of the system state according to changes in the health status or service capability of the online service system.

In a specific implementation, the rise or fall of the system state may be determined according to various manners such as the service capability of the online service system or the change trend and degree of change of the health state. In one embodiment, a feature value (eg, queue length to be processed) that can characterize the state of the system can be simply selected, and the system state can be directly determined by comparing the current feature value with a preset or real-time adjusted threshold. Depending on the specific feature value selected, it may be that the feature value increase indicates that the system state is rising (for example, the feature value takes the CPU idle rate), or the decrease of the feature value indicates that the system state is rising (for example, the foregoing pending queue) The case of length). The present invention does not limit the proportional relationship between the feature value and the system state as long as the change can reflect the change in the state of the system.

In another embodiment, a combination of multiple feature values can be selected to determine the state of the system from more dimensions. For example, the corresponding threshold(s) can be set for the CPU idle rate and the queue length to be processed, respectively, and the rise or fall of the system state is determined as a whole according to the relationship between the current values of the two and their respective thresholds. In other embodiments, more or other dimensions are introduced, which are not limited by the present invention.

When it is determined that the system state of the online service system is declining, it indicates that the current service capability of the online service system is degraded, and the service quality of the online service system can be immediately lowered to a level corresponding to the current system state. "Immediately" as used herein refers to a subsequent degradation without additional conditions.

In step S120, when it is judged that the system state is rising, the service quality is suspended to the level corresponding to the current system state.

As mentioned above, due to the unstable traffic characteristics of the online service system, the system status of the online service system often appears to fluctuate up and down. Therefore, if the system status of the online service system is judged to rise, the service of the online service system will be immediately The quality is restored to the level corresponding to the current system state, which may cause the online service system to continue to be in a high service stress state, which is not conducive to the recovery of the system state of the online service system.

Based on the above considerations, when it is judged that the system state is rising, the suspension recovery operation can be performed to suspend the restoration of the service quality to the level corresponding to the current system state, so that the adverse effects due to the jitter and repeatability of the system state can be effectively avoided. .

The suspension recovery operation described herein may have certain additional conditions, that is, the quality of service may be restored to a level corresponding to the current system state after the system state rises and certain additional conditions are met. For example, the additional condition described herein may be that the predetermined time threshold is exceeded, that is, the quality of service may be restored to a level corresponding to the current system state after the system state has risen and exceeded a predetermined time threshold. In addition, the suspend recovery operation may also have other additional conditions as will be detailed below, and will not be described here.

It should be noted that after the suspension of the recovery operation is completed, the quality of service may be restored to the level corresponding to the current system state of the online service system before the suspension operation is resumed, or the quality of service may be restored to the online after the suspension operation is completed. The level of the current system state of the service system.

In addition, in order to alleviate the service pressure of the online service system, the system state of the online service system can be well recovered. After the system state is judged to rise and the suspension recovery operation is performed, the quality of service can be restored to a level corresponding to the current system state. Other levels lower than one or several levels. In this way, while improving the service quality of the online service system, it can also contribute to the subsequent recovery of the system state of the online service system. The current system state herein may refer to the current system state of the online service system before the suspension of the recovery operation, or the current system state of the online service system after the suspension operation is completed.

Based on the above description in conjunction with FIG. 1, the service level control method of the present invention mainly includes a service downgrading step and a service recovery step, and the service downgrading step may be performed immediately in response to a system state drop of the online service system, and the service recovery step is Suspend recovery in response to a system state rise of the online service system. Therefore, by adopting a level control strategy of fast-downgrading and slow recovery, it is possible to prevent the adverse effects caused by the jitteriness and repeatability of the system state of the online service system. Moreover, the slow recovery can also provide more time for the online service system to further restore its system state.

It should be noted that although the above is the service restoration step S110 described in the service degradation step S110 described earlier, it should be understood that the numbering and description order of the above steps does not limit the order in which the degradation and recovery steps occur. In actual operation, the system status of the online service system may be a variety of changes such as first rising and then descending, first falling, then rising, successively rising several levels, successively descending several levels, and moving up and down, and accordingly, using the service of the present invention. When the level control method implements the control of the service level of the online service system, the corresponding degradation or recovery operation may also be performed according to the actual change of the system state.

So far, the principle and process of the service level control method of the present invention are briefly explained in conjunction with FIG. The service level control method of the present invention will be further described in detail below with reference to specific embodiments.

Embodiment 1

In this embodiment, the system state of the online service system may be characterized by the feature value of the current service capability, and accordingly, the quality of service may be divided into multiple levels corresponding to the plurality of feature value thresholds.

As described above, the online service system can be a plurality of types of Internet application systems, and the feature values for characterizing the service capabilities of the online service system are different according to the online service system. For example, for the information recommendation service online system, the queue length that needs to be processed may be used as the feature value for characterizing the system state, and for other distributed task systems, the current available resources of the system, idle servers, etc. may also be used as the representation system state. Characteristic value.

When the system state is characterized by the feature value, the service degradation step and the service recovery step may be performed according to the relationship between the current feature value and the feature value threshold.

Specifically, for the service degradation step, when the system state is degraded such that the current feature value satisfies the first feature value threshold, the service quality of the online service system may be immediately lowered to a level corresponding to the first feature value threshold.

For the service recovery step, when it is determined that the system state is raised such that the current feature value satisfies the second feature value threshold, the quality of service is suspended to a level corresponding to the second feature value threshold.

In this embodiment, the feature value and the system state may be proportional or inversely proportional, that is, the larger the feature value, the better the system state, or the smaller the feature value, the better the system state. Also, the first feature value threshold and the second feature value threshold may be the same or different.

As mentioned above, the suspend recovery operation in the service recovery step may have certain additional conditions. In this embodiment, the following additional conditions can be set.

1.1. The quality of service is restored to a level corresponding to the second feature value threshold when it is determined that the system state is raised such that the current feature value satisfies the second feature value threshold for a predetermined length of time.

Therefore, for the phenomenon caused by the up and down jitter of the system state, the current feature value can be made to satisfy the second eigenvalue threshold for a predetermined period of time as an additional condition. This avoids the online service system being under high stress and also helps in the rapid recovery of the system state of the online service system.

1.2. Restoring the quality of service to a level corresponding to the second feature value threshold when determining that the system state is rising such that the current feature value satisfies a third feature value threshold that is higher than the second feature value threshold quality level.

The third eigenvalue threshold may be a eigenvalue threshold that is higher than the second eigenvalue threshold level, or may be an eigenvalue threshold that is higher than the second eigenvalue threshold level. In this way, the online service system can have an idle service capability to improve the business status. For example, when the feature value is the queue length currently required by the online service system, the online service system can be made to have the task in the idle resource processing queue length to reduce the queue length.

1.3. When it is judged that the system state rises such that the current feature value satisfies the second eigenvalue threshold corresponding to the optimal QoS level, only the current eigenvalue further satisfies the best service regression value that represents better service quality than the second eigenvalue. When the quality of service is restored to the best quality of service level.

The second feature threshold corresponds to the best quality of service level, and the best service regression value is not used to divide the quality level, which is only a threshold for indicating that the online service system is restored to the best quality. When it is judged that the system state rises until the current feature value satisfies the second feature threshold and further satisfies the optimal service regression value, it may indicate that the system state of the online service system has stabilized at a state suitable for providing the best quality of service level, and Quality of service is restored to the best quality of service level.

Embodiment 2

2 is a schematic diagram showing a downgrade trigger and upgrade recovery using a particular embodiment of the present invention. The abscissa is the time axis and the ordinate is the feature value. As shown in FIG. 2, three eigenvalue thresholds a, b, and c can be set. In this embodiment, the feature value may characterize the length of the request queue that the online service system currently needs to process.

1) Before the time t2, the request queue length is less than the feature value threshold b, and the online service system is in a normal service state, and the best quality service can be provided externally.

2) As the traffic increases, the system cannot support high-quality services, and the request queue length increases continuously. The queue length reaches the eigenvalue threshold b at time t2. At this time, the online service system can enter the first-level degradation state and provide relatively poor outward. Quality service.

3) As the traffic continues to increase, the online service system still cannot support the relatively poor quality service. The request queue length continues to increase. The queue length reaches the eigenvalue threshold c at time t3. At this time, the online service system can enter the secondary degradation state. Only provide the most basic services.

4) From t3 to t5, since the online system only provides the most basic services, a large amount of resources are released, the service capability is strengthened, and the request queue length is improved, but before the queue length drops to the eigenvalue threshold b, the online service system can still Only provide the most basic services. The reason why the queue length is not changed to the eigenvalue threshold c immediately after switching to the first-level degraded state is to avoid repetition and prevent the online service system from staying in the high-pressure state, so that the health status of the online service system can be as Get it back quickly.

5) At time t5, as the request queue length further drops below b, the online service system returns to the first-level degradation state.

6) At time t6, the service capability of the online system is fully restored, exiting the first-level downgrade state, and restoring high-quality services.

In summary, the service level control method in this embodiment uses the request queue length as an indicator for judging the system service capability, and designs a multi-level threshold, and adopts different thresholds for triggering and restoring the degraded state, realizing the rapid triggering and slow recovery of the degraded state. In the single-threshold degradation triggering scheme, the system stays in the vicinity of high service pressure and reduces the system risk. In this embodiment, the thresholds b and c are threshold values for dividing the service level, and the threshold a is not used for the actual system service level division (this is because the system service can already be considered when the queue length drops to b). The ability has been restored, but it can be seen as the best service regression. In other words, only after the system state has returned to a can the system service capability be considered "safely" fully restored.

FIG. 3 is a functional block diagram showing the structure of an online quality of service control system 300 in accordance with an embodiment of the present invention. Wherein, the functional modules of the online service quality control system 300 can be implemented by hardware, software or a combination of hardware and software that implements the principles of the present invention. Those skilled in the art can understand that the functional modules described in FIG. 3 can be combined or divided into sub-modules to implement the principles of the above invention. Accordingly, the description herein may support any possible combination, or division, or further limitation of the functional modules described herein.

The online service quality control system 300 shown in FIG. 3 can be used to implement the service level control method described in FIG. 1 and FIG. 2, and only the functional modules that the online service quality control system 300 can have and the operations that can be performed by each functional module. For a brief description, reference may be made to the above description in conjunction with FIG. 1 and FIG. 2 for details of the details involved therein, and details are not described herein again.

As shown in FIG. 3, the online quality of service control system 300 includes a distributed service system 310 and a quality control server 320 for providing online services. The distributed service system 310 described herein can be equivalent to the online service system described above.

The quality control server 320 can obtain the feature values used to characterize the current service capabilities of the distributed service system 310, and immediately reduce the quality of service of the distributed service system 310 when the current feature values characterize the system state of the distributed service system 310. As for the level corresponding to the current system state, and when the current feature value indicates that the system state is rising, the service quality is suspended to the level corresponding to the current system state. In other words, the online quality of service control system 300 incorporating the quality control server 320 can also be viewed as a distributed online service system with quality control functionality.

As an alternative embodiment of the present invention, one or more quality control servers 320 may also restore the quality of service to a level lower than the level corresponding to the current system state when the current feature value characterizes the system state rise.

As another optional embodiment of the present invention, the quality of service may be divided into a plurality of levels corresponding to a plurality of feature value thresholds, and wherein the one or more quality control servers 320 may be further configured to: determine the system status When the current feature value satisfies the first eigenvalue threshold, the service quality of the online service system is immediately lowered to a level corresponding to the first eigenvalue threshold; and the system state is raised so that the current eigenvalue satisfies the second eigenvalue threshold. At this time, the quality of service is suspended to a level corresponding to the second eigenvalue threshold.

As another optional embodiment of the present invention, the one or more quality control servers 320 may be further configured to: when determining that the system state rises such that the current feature value meets the second feature value threshold for a predetermined length of time, the quality of service is restored to The level corresponding to the second eigenvalue threshold.

As another optional embodiment of the present invention, the one or more quality control servers 320 may be further configured to: determine a third eigenvalue threshold that is higher than the second eigenvalue threshold quality level by determining that the system state is rising such that the current eigenvalue satisfies a second eigenvalue threshold quality level The quality of service is restored to the level corresponding to the feature value threshold.

As another optional embodiment of the present invention, the one or more quality control servers 320 may be further configured to: when determining that the system state rises such that the current feature value satisfies the second eigenvalue threshold corresponding to the optimal quality of service level, only The quality of service is restored to the optimal quality of service level when the current feature value further satisfies the best service regression value that represents a better quality of service than the second feature value.

As another alternative embodiment of the present invention, the feature value of the current service capability may be the queue length that the distributed service system 310 currently needs to process.

4 is a schematic block diagram showing the structure of an online quality of service control system in accordance with another embodiment of the present invention.

As shown in FIG. 4, the distributed service system 310 can be refined to include a plurality of service modules, and the quality control server 320 can be connected to the service modules in the distributed service system 310, wherein FIG. 4 is a diagram showing online quality of service control. The system 300 includes a quality control server 320. It should be appreciated that the online quality of service control system 300 can also include a plurality of quality control servers 320, which can be associated with multiple service modules in the distributed service system 310. In one correspondence, each quality control server 320 can adjust its quality of service level according to the current system state of its corresponding distributed service module.

For example, the feature value of the current service capability may be the queue length currently required by one or more service modules of the distributed service system 310, and each quality control server 320 may adjust the quality of service level of the service module corresponding thereto according to the feature value. . For the process of the quality control server 320 adjusting the service quality level of the service module, refer to the related description above, and details are not described herein again.

The service level control method and system of the online service system according to the present invention has been described in detail above with reference to the accompanying drawings.

It will be apparent to those skilled in the art that the technology in the embodiments of the present invention can be implemented by means of software plus necessary general hardware including general-purpose integrated circuits, general-purpose CPUs, general-purpose memories, general-purpose components, and the like. It can be implemented by dedicated hardware including an application specific integrated circuit, a dedicated CPU, a dedicated memory, a dedicated component, etc., but in many cases the former is a better implementation. Based on such understanding, the technical solution in the embodiments of the present invention may be embodied in the form of a software product in essence or in the form of a software product, which may be stored in a storage medium such as a read-only memory. (ROM, Read-Only Memory), Random Access Memory (RAM), disk, CD, etc., including a number of instructions to make a computer device (can be a personal computer, server, or network device, etc.) The methods described in various embodiments of the invention or in certain portions of the embodiments are performed.

Accordingly, the above technical concept of the present invention can also be embodied as a non-transitory machine-readable storage medium (or computer-readable storage medium) on which executable code (or computer program/computer instruction code) is stored. When the executable code (or computer program/computer instruction code) is executed by the processor, the processor is caused to perform the service level control method described above.

In another aspect, the above technical concept of the present invention can also be embodied as a computing device including a processor and a non-transitory machine readable storage medium (or computer readable storage medium). The non-transitory machine readable storage medium stores executable code (or computer program/computer instruction code). When the executable code (or computer program/computer instruction code) is executed by the processor, the processor is caused to perform the service level control method described above.

Furthermore, the method according to the invention can also be implemented as a computer program comprising computer program code instructions for performing the various steps defined above in the above method of the invention. Alternatively, the method according to the invention may also be embodied as a computer program product comprising a computer readable medium on which is stored a computer for performing the above-described functions defined in the above method of the invention program. The various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the disclosure herein may be implemented as electronic hardware, computer software, or combinations of both.

The flowchart and block diagrams in the Figures illustrate the architecture, functionality and operation of possible implementations of systems and methods in accordance with various embodiments of the present invention. In this regard, each block of the flowchart or block diagram can represent a module, a program segment, or a portion of code that includes one or more of the Executable instructions. It should also be noted that in some alternative implementations, the functions noted in the blocks may also occur in a different order than those illustrated in the drawings. For example, two consecutive blocks may be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending upon the functionality involved. It is also noted that each block of the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts, can be implemented in a dedicated hardware-based system that performs the specified function or operation. Or it can be implemented by a combination of dedicated hardware and computer instructions.

The embodiments of the present invention have been described above, and the foregoing description is illustrative, not limiting, and not limited to the disclosed embodiments. Numerous modifications and changes will be apparent to those skilled in the art without departing from the scope of the invention. The choice of terms used herein is intended to best explain the principles, practical applications, or improvements of the techniques in the various embodiments of the embodiments, or to enable those of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

A service level control method for an online service system, wherein the quality of service provided by the online service system is divided into multiple levels according to a system state, and the method includes:

a service degradation step of lowering a service quality of the online service system to a level corresponding to a current system state when determining that a system state of the online service system is degraded;

The service recovery step suspends the restoration of the quality of service to a level corresponding to the current system state when it is determined that the system status is rising.
The method of claim 1 wherein said service recovery step comprises:

When it is determined that the system state is rising, the quality of service is restored to a level lower than a level corresponding to the current system state.
The method of claim 1 wherein said system state is characterized by a feature value of a current service capability, said quality of service being divided into a plurality of levels corresponding to a plurality of feature value thresholds,

And among them,

The service degradation step includes:

When it is determined that the system state is decreased such that the current feature value satisfies the first feature value threshold, the service quality of the online service system is immediately lowered to a level corresponding to the first feature value threshold;

The service recovery steps include:

When it is determined that the system state rises such that the current feature value satisfies the second feature value threshold, the quality of service is suspended to a level corresponding to the second feature value threshold.
The method of claim 3 wherein said service recovery step comprises:

The quality of service is restored to a level corresponding to the second feature value threshold when it is determined that the system state rises such that the current feature value satisfies the second feature value threshold for a predetermined length of time.
The method of claim 3 wherein said service recovery step comprises:

Restoring the quality of service to a second eigenvalue threshold corresponding to determining that the current state value is such that the current eigenvalue satisfies a third eigenvalue threshold that is higher than the second eigenvalue threshold quality level level.
The method of claim 3 wherein said service recovery step comprises:

When it is determined that the system state rises such that the current feature value satisfies the second eigenvalue threshold corresponding to the optimal QoS level, only the current eigenvalue further satisfies the best service that represents better service quality than the second eigenvalue. The quality of service is restored to the optimal quality of service level upon regression.
The method according to claim 3, wherein the feature value of the current service capability is a queue length currently required to be processed by the online service system; wherein the current required queue length is reduced, and the representation The system state rises; the queue length of the current required processing increases, indicating that the system state is degraded.
An online quality of service control system, comprising: a distributed service system for providing an online service and one or more quality control servers, the one or more quality control servers for:

Obtaining a feature value for characterizing a current service capability of the distributed service system;

As soon as the current feature value characterizes the system state of the distributed service system, the quality of service of the distributed service system is immediately lowered to a level corresponding to the current system state;

When the current feature value characterizes the system state rise, the service quality is suspended to a level corresponding to the current system state.
The control system of claim 8 wherein said one or more quality control servers are further for:

When the current feature value characterizes the system state rise, the quality of service is restored to a level lower than a level corresponding to the current system state.
The control system of claim 8 wherein said quality of service is divided into a plurality of levels corresponding to a plurality of feature value thresholds, and wherein said one or more quality control servers are further configured to:

When it is determined that the system state is decreased such that the current feature value satisfies the first feature value threshold, the service quality of the online service system is immediately lowered to a level corresponding to the first feature value threshold;

When it is determined that the system state rises such that the current feature value satisfies the second feature value threshold, the quality of service is suspended to a level corresponding to the second feature value threshold.
The control system of claim 10 wherein said one or more quality control servers are further for:

The quality of service is restored to a level corresponding to the second feature value threshold when it is determined that the system state rises such that the current feature value satisfies the second feature value threshold for a predetermined length of time.
The control system of claim 10 wherein said one or more quality control servers are further for:

Restoring the quality of service to a second eigenvalue threshold corresponding to determining that the current state value is such that the current eigenvalue satisfies a third eigenvalue threshold that is higher than the second eigenvalue threshold quality level level.
The control system of claim 10 wherein the one or more quality control servers are further configured to:

When it is determined that the system state rises such that the current feature value satisfies the second eigenvalue threshold corresponding to the optimal QoS level, only the current eigenvalue further satisfies the best service that represents better service quality than the second eigenvalue. The quality of service is restored to the optimal quality of service level upon regression.
The control system according to claim 10, wherein said feature value of said current service capability is a queue length currently required to be processed by said distributed service system; wherein said currently required queue length is reduced Representing the rise of the system state; the queue length of the current required processing is increased, indicating that the state of the system is degraded.
The control system according to claim 10, wherein the feature value of the current service capability is a queue length currently required by one or more service modules of the distributed service system; wherein the current location The length of the queue to be processed is decreased, and the state of the system is marked to rise; the queue length of the current required processing is increased, and the state of the system is degraded;

And the one or more quality control servers adjust the quality of service level of the one or more service modules according to the feature value.
A readable storage medium, characterized in that a computer program is stored thereon;

The method of any one of claims 1-7 is implemented when the computer program is executed by a processor.