Method and system for supporting cloud storage sudden performance

Abstract

The invention relates to the technical field of cloud storage, in particular to a method and system for supporting the sudden performance of cloud storage, and the method comprises the following steps: constructing a double leaky bucket model, namely a time window control leaky bucket and a performance peak control leaky bucket, which are used for intelligently adjusting the sudden performance of storage access; the method and the system for supporting the cloud storage sudden performance have the beneficial effects that by introducing a sudden performance upper limit mechanism, short-term and high-strength storage access requirements, such as scenes of cloud server quick starting, data intensive task processing and the like, can be effectively met, the response time is remarkably shortened, and the efficiency is improved. And the satisfaction degree of the user to the cloud storage service is improved. Especially in a business peak period, the user does not need to upgrade to a storage instance with a higher specification due to temporary performance demand improvement, so that the cost burden is reduced, and the business flexibility and the market response speed are improved.

Description

Technical Field

[0001] The present invention relates to the field of cloud storage technology, and in particular to a method and system for supporting cloud storage burst performance. Background Art

[0002] The instance specification parameters of cloud storage include maximum bandwidth, maximum IOPS, etc. Cloud storage services provide different performance parameters according to different storage instance specifications. For example, "maximum IOPS = 10000" means that the storage can process a maximum of 10,000 storage read and write disk requests per second, and "maximum bandwidth = 1536Mbps" means that the storage can process a maximum of 1536 megabits of data requests per second.

[0003] Current storage throttling controls usually only set a fixed performance cap based on the storage instance specifications, and do not allow storage access capabilities that exceed the performance cap of the current instance specifications. Specifically, storage resources are logically divided through software or hardware means, and independent IOPS and throughput quotas are allocated to each user or instance. This prevents the high I / O demands of a certain user or application from affecting other users, and ensures fairness and stability in a multi-tenant environment. When a user's I / O request exceeds its allocated IOPS or throughput limit, the excess requests are placed in a queue or buffer for processing. By adjusting the queue length and processing strategy, cloud storage can control the response time of requests and avoid performance degradation caused by excessive competition for resources.

[0004] However, this performance restriction method is not flexible enough. User business access to cloud storage does not always maintain the performance upper limit of the specification. There is a certain period of time when the performance is lower than the upper limit or even no traffic reaches the cloud storage. Therefore, when the business needs to access a scenario that exceeds the performance upper limit within a certain period of time, the existing flow control solution cannot support this demand. Summary of the invention

[0005] The object of the present invention is to provide a method and system for supporting cloud storage burst performance, so as to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for supporting burst performance of cloud storage, the method comprising the following steps:

[0007] Construct a dual leaky bucket model, namely the "time window control leaky bucket" and the "performance peak control leaky bucket", to intelligently adjust the burst performance of storage access;

[0008] Specifically, it includes: a) dynamically configuring leaky bucket parameters, defining the basic performance rate and maximum performance rate, as well as the duration of burst performance according to the storage instance specifications;

[0009] b) Monitor client I / O requests in real time and dynamically adjust the processing speed through the leaky bucket algorithm model, allowing the performance within the controllable range to exceed the basic performance limit;

[0010] c) Scheduled tasks are combined with the dual-bucket structure to jointly process requests that exceed burst performance based on the number and size of client I / Os.

[0011] Preferably, the "time window control leaky bucket" is designed to limit the effective duration of burst performance, and the capacity is composed of the maximum performance rate multiplied by the preset burst duration, ensuring that the performance improvement is effective within the specified time; the "performance peak control leaky bucket" is designed to constrain single or short-term performance peaks, and its capacity is set to a preset percentage of the maximum performance rate to balance performance bursts and system stability. When the client I / O request exceeds the burst duration / burst performance, the burst performance is reduced by extending its execution time through scheduled tasks.

[0012] Preferably, real-time monitoring of client I / O requests also includes:

[0013] The automated dynamic adjustment mechanism adaptively adjusts leaky bucket parameters based on real-time I / O activities and historical data, optimizes resource allocation strategies, adapts to changing workloads, introduces multi-level priority scheduling, and configures personalized leaky bucket parameters for I / O requests of different priorities, ensuring that the performance requirements of key businesses are prioritized.

[0014] Preferably, processing requests exceeding the burst performance specifically includes:

[0015] Scheduled tasks are combined with dual buckets to jointly process requests that exceed burst performance. While ensuring the security of client data and preventing loss, I / O requests that exceed the performance limit can be controlled within the set burst performance and burst duration.

[0016] A system for supporting sudden performance of cloud storage, the system comprising the following modules:

[0017] Double leaky bucket model building module: used to build and maintain a double leaky bucket model including a "time window control leaky bucket" and a "performance peak control leaky bucket" to intelligently adjust the burst performance of storage access;

[0018] Dynamic configuration and adjustment module: dynamically configures leaky bucket parameters according to storage instance specifications and real-time I / O activities, and makes adaptive adjustments based on historical data to optimize resource allocation strategies;

[0019] Multi-level priority scheduling module: configures personalized leaky bucket parameters for I / O requests of different priorities to ensure that the performance requirements of key businesses are prioritized;

[0020] Scheduled task processing module: Combined with the double leaky bucket structure, it regularly processes requests that exceed burst performance while ensuring the security and integrity of client data.

[0021] Preferably, the double leaky bucket model building module includes:

[0022] Time window control leaky bucket submodule: defines the effective duration of burst performance. The capacity is determined by the maximum performance rate and the preset burst duration, which is used to limit the effective time range of burst performance.

[0023] Performance peak control leaky bucket submodule: sets the upper limit of the performance peak for a single time or a short period of time. The capacity is a preset percentage of the maximum performance rate. It is used to constrain the performance peak and balance performance burst with system stability.

[0024] Preferably, the dynamic configuration and adjustment module includes:

[0025] The leaky bucket parameter configuration submodule is used to set the basic performance rate, maximum performance rate and burst duration parameters;

[0026] The adaptive adjustment submodule is used to monitor I / O activities in real time and dynamically adjust the leaky bucket parameters based on real-time and historical data to adapt to changing workloads.

[0027] Preferably, the multi-level priority scheduling module includes:

[0028] A priority determination submodule, used to determine the priority of an I / O request according to the requested business type or user-defined rules;

[0029] The personalized parameter configuration submodule is used to configure corresponding leaky bucket parameters for requests of different priorities to ensure that key businesses receive priority processing.

[0030] Preferably, the scheduled task processing module includes:

[0031] The over-performance processing submodule is used to reduce the burst performance by extending the execution time when the I / O request exceeds the burst performance;

[0032] The data security protection submodule is used to ensure that the security and integrity of client data are not affected during the processing.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] The method and system for supporting burst performance of cloud storage proposed in the present invention can effectively respond to short-term, high-intensity storage access requirements, such as cloud server rapid startup, data-intensive task processing, etc., by introducing a "burst performance cap" mechanism, significantly shortening the response time and improving user satisfaction with cloud storage services. Especially during business peaks, users do not need to upgrade to higher-specification storage instances due to temporary performance demand increases, thereby reducing cost burdens and improving business flexibility and market response speed.

[0035] Through dynamic adjustment mechanism and intelligent prediction model, the present invention can automatically optimize leaky bucket parameters according to historical data and real-time monitoring information, ensuring that the allocation of burst performance can meet the burst demand without compromising system stability and the user experience of other users. This adaptability helps the cloud storage platform maintain high performance and high reliability under different workloads, enhancing the sustainable operation capability of the system.

[0036] The introduction of the priority control mechanism enables critical businesses or urgent tasks to gain an advantage in resource competition, ensuring that the performance requirements of high-priority applications are met first under sudden access pressure. This is particularly important for enterprise-level users and can ensure the continuity and stability of core business processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0038] In order to make the purpose and technical solution of the present invention clearly and completely described, and the advantages more clearly understood, the embodiments of the present invention are further described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, rather than all of the embodiments, and are only used to explain the embodiments of the present invention, and are not used to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] Embodiment 1, the present invention provides a technical solution: a method for supporting burst performance of cloud storage, the method comprising the following steps:

[0040] A dual leaky bucket model is constructed, namely the "time window control leaky bucket" and the "performance peak control leaky bucket", which are used to intelligently adjust the burst performance of storage access; the "time window control leaky bucket" is designed to limit the effective duration of burst performance, and its capacity is composed of the maximum performance rate multiplied by the preset burst duration to ensure that the performance improvement is effective within the specified time; the "performance peak control leaky bucket" is designed to constrain the performance peak of a single or short period of time, and its capacity is set to a preset percentage of the maximum performance rate to balance performance bursts and system stability. When the client I / O request exceeds the burst duration / burst performance, the burst performance is reduced by extending its execution time through scheduled tasks.

[0041] Specifically, it includes: a) dynamically configuring leaky bucket parameters, defining the basic performance rate and maximum performance rate, as well as the duration of burst performance according to the storage instance specifications;

[0042] b) Monitor client I / O requests in real time, dynamically adjust processing speed through the leaky bucket algorithm model, and allow performance within the controllable range to exceed the basic performance ceiling; the automated dynamic adjustment mechanism adaptively adjusts leaky bucket parameters based on real-time I / O activities and historical data, optimizes resource allocation strategies, adapts to changing workloads, introduces multi-level priority scheduling, and configures personalized leaky bucket parameters for I / O requests of different priorities, ensuring that the performance requirements of key businesses are prioritized.

[0043] c) Scheduled tasks combine the dual-bucket structure to jointly process requests that exceed burst performance based on the number and size of client I / Os; Scheduled tasks combine the dual-bucket structure to jointly process requests that exceed burst performance. While ensuring that client data is secure and not lost, I / O requests that exceed the performance limit can be controlled within the set burst performance and burst duration.

[0044] Embodiment 2, refer to the attached Figure 1 Based on the first embodiment, the leaky bucket algorithm principle is proposed: a classic algorithm for traffic shaping and rate limiting, whose working principle can be figuratively compared to a bucket with a fixed volume of water, with a small hole at the bottom of the bucket, and water will leak out of the bucket at a constant rate. In this scheme, the client I / O request enters the bucket in the form of adding water. If the water in the bucket does not overflow, the I / O request is allowed to pass, otherwise the amount of water blocked in the bucket is less than the bucket capacity and the client I / O is allowed to be processed.

[0045] Based on the above principles, the specific implementation scheme of the present invention is as follows:

[0046] The present invention adds a leaky bucket data structure in the storage engine and initializes two leaky buckets for each specification instance. The bucket used to control the duration of burst performance is called the "burst time control bucket", referred to as the time bucket, which is mainly used to limit the duration of burst performance; the bucket used to control the upper limit value of burst performance is called the "burst upper limit control bucket", referred to as the burst bucket, which is mainly used to limit the upper limit of burst performance. The two buckets cooperate with each other to achieve the ability of burst performance.

[0047] First, define the relevant concepts: time bucket capacity: time_bkt_capacity; actual water volume of time bucket: time_bkt_size; burst bucket capacity: burst_bkt_capacity; actual water volume of burst bucket: burst_bkt_size. The burst performance limit provided by each instance specification is max / s, and the supported burst duration is burst_length seconds.

[0048] The burst time bucket capacity is the burst upper limit value (max) * burst time (burst_length). The leakage rate of the time bucket is the basic performance of the instance specification, that is, avg / s; when the storage access pressure is greater than or equal to the burst upper limit, the leakage rate is the basic performance upper limit / s, that is, max / s. The purpose of the burst bucket is to limit the maximum storage access pressure to the basic performance upper limit (max). The capacity is 1 / 10*burst upper limit value (max), and the leakage rate is the burst performance upper limit / s, that is, max / s.

[0049] 1. When the client I / O arrives, the flow control device of the specification instance calculates the time difference delta_leak_time between the current time and the last client I / O arrival, in nanoseconds, and then updates the last leak time to the current time.

[0050] 2. Calculate the reduction and remaining amount of water in the two buckets respectively through delta_leak_time.

[0051] a) The amount of water reduced in the time bucket time_leak = avg*delta_leak_time,

[0052] b) The amount of water reduction in the burst bucket burst_leak = max*delta_leak_time.

[0053] c) Time bucket remaining amount time_bkt_size = max(time_bkt_capacity -

[0054] time_leak,0)

[0055] d) Burst bucket remaining amount burst_bkt_size = max(burst_bkt_size -

[0056] burst_leak,0)

[0057] 3. Calculate the capacity of the time bucket and burst bucket respectively.

[0058] a) time_bkt_capacity=max*burst_length

[0059] b)burst_bkt_capacity=max*1 / 10

[0060] The burst bucket capacity limit is relatively small, only 1 / 10*max, in order to limit the client access to the storage backend to not exceed the burst upper limit. This will be introduced later.

[0061] 4. Determine whether the current client can use the burst upper limit max capability

[0062] a) Calculate the amount of water in the time bucket minus the bucket capacity extra = time_bkt_size -

[0063] time_bkt_capacity. If it is less than 0, it means that the water has overflowed, that is, the current I / O is overflowing, then the calculation waiting time of this part of I / O = extra / avg (leakage speed = avg).

[0064] b) If the extra calculated in (a) is greater than 0, that is, I / O can continue to be processed, then calculate the overflow water volume of the burst bucket extra = the current water volume of the burst bucket burst_level - the capacity of the burst bucket burst_size. If the extra of the burst bucket is less than 0, it means that the water (I / O) of the burst bucket is overflowing, and calculate the waiting time (burst bucket leakage speed = max). The burst bucket water volume minus the bucket capacity extra = burst_bkt_size-burst_bkt_capacity. If it is less than 0, it is necessary to wait, and the waiting time = extra / max until the water volume in the bucket is less than the capacity of the burst bucket.

[0065] 5. If the calculation result in 4 does not need to wait, the client I / O request is placed in the bucket, which means that the request volume this time meets the burst upper limit performance requirements, and cloud storage will process this I / O request.

[0066] 6. In step 4, if the calculation shows that it is necessary to wait for execution, there are two possibilities. (1) The water volume in the time bucket has exceeded the capacity of the time bucket, and the corresponding burst performance time has reached the burst_length duration, and the performance begins to recover to the basic performance; (2) The water volume in the burst bucket exceeds the capacity of the burst bucket, and the corresponding client I / O pressure exceeds the burst performance max / s. In these two cases, a timer will be enabled. This timer will be triggered after the waiting time calculated in step 4, and the process of steps 2-4 will be executed repeatedly until step 4 calculates the conditions that do not allow direct execution of client I / O.

[0067] 7. After step 5 or step 6 is completed, it means that the last I / O request has been completed, and there is enough space in the time bucket and burst bucket to process the new I / O request. When the client I / O request arrives this time, the size of the I / O request is added to the time_bkt_size of the time bucket and the burst_bkt_size of the burst bucket, and the I / O request in the bucket is processed in the next I / O scheduling.

[0068] 8. Dynamic adjustment mechanism: In order to further optimize resource allocation and adapt to changes in different workloads, the present invention introduces a dynamic adjustment mechanism. Based on historical data analysis and real-time monitoring, the capacity and leakage rate of the time bucket and burst bucket are automatically adjusted. For example, during off-peak hours, the capacity of the burst bucket can be appropriately increased to allow for more burst performance requirements, and in anticipation of upcoming high-load periods, the burst bucket capacity can be reduced in advance to reserve sufficient resources for key businesses or prevent potential resource shortages.

[0069] 9. Priority control: In a multi-tenant environment, different users may have different cloud storage requirements and priorities. The present invention adds priority tags based on the leaky bucket algorithm to allow high-priority I / O requests to be processed first when resources are tight. In specific implementation, different leaky bucket parameters can be set for each priority, such as a shorter burst duration limit, a larger burst performance ceiling, or a faster basic leak rate, so as to achieve more sophisticated resource scheduling and priority management at the system level.

[0070] 10. Intelligent prediction and reservation: Introduce machine learning models to analyze and predict user I / O behavior patterns and identify potential burst access patterns in advance. Based on the prediction results, the system dynamically adjusts the leaky bucket configuration of each instance and reserves resources in advance for the expected high-traffic period to avoid the impact of burst traffic on the system and improve resource utilization.

[0071] 11. Feedback and adaptive learning: The system collects performance data and user feedback during the execution of the leaky bucket, and continuously optimizes the algorithm parameters through the feedback mechanism to achieve self-learning and adjustment. For example, if I / O requests are frequently rejected due to long waiting times, the system can automatically adjust the burst bucket capacity or leaking speed to better match the actual workload requirements.

[0072] Visual monitoring and management interface: Provides a user-friendly graphical interface that displays information such as the leaky bucket status of each instance, current I / O traffic, historical trends, and performance threshold settings. This allows administrators to intuitively understand resource usage and adjust burst performance parameters in real time according to business needs, enhancing the manageability and transparency of the system.

[0073] The following effects are achieved:

[0074] 1. Improve user experience and business agility: This invention can effectively respond to short-term, high-intensity storage access needs, such as cloud server rapid startup, data-intensive task processing, etc., by introducing a "burst performance cap" mechanism, significantly shortening response time and improving user satisfaction with cloud storage services. Especially during business peaks, users do not need to upgrade to higher-specification storage instances due to temporary performance demand increases, thereby reducing cost burdens and improving business flexibility and market response speed.

[0075] 2. Efficient resource utilization and cost optimization: This solution allows the cloud storage system to accumulate underutilized performance capacity during periods of low access pressure, and then release it in a centralized manner when needed, forming "burst performance". This mechanism effectively avoids the idle waste of resources during most non-peak periods, allowing users to enjoy storage services that exceed the basic performance limit without incurring additional costs, greatly improving resource utilization efficiency and cost-effectiveness.

[0076] 3. Dynamic adaptability and self-adjustment capability: Through dynamic adjustment mechanism and intelligent prediction model, the present invention can automatically optimize leaky bucket parameters according to historical data and real-time monitoring information, ensuring that the allocation of burst performance can meet burst demand without compromising system stability and other users' experience. This adaptability helps the cloud storage platform maintain high performance and high reliability under different workloads, enhancing the sustainable operation capability of the system.

[0077] 4. Fine-grained management and priority guarantee: The introduction of the priority control mechanism enables key businesses or urgent tasks to gain an advantage in resource competition, ensuring that the performance requirements of high-priority applications are met first under sudden access pressure. This is especially important for enterprise-level users and can ensure the continuity and stability of core business processes.

[0078] 5. Visual monitoring and simplified management: The visual monitoring interface provided enables administrators to intuitively monitor storage performance status, adjust policies in a timely manner, and effectively plan resources, reducing management complexity and improving operation and maintenance efficiency. At the same time, this transparent management also helps users better understand resource usage and promote trust and cooperation with cloud service providers.

[0079] Embodiment 3, based on embodiment 2, proposes a system for supporting cloud storage burst performance, the system comprising the following modules:

[0080] Double leaky bucket model building module: used to build and maintain a double leaky bucket model including "time window control leaky bucket" and "performance peak control leaky bucket" to intelligently adjust the burst performance of storage access; time window control leaky bucket submodule: defines the effective duration of burst performance. The capacity is determined by the maximum performance rate and the preset burst duration, which is used to limit the effective time range of burst performance;

[0081] Performance peak control leaky bucket submodule: sets the upper limit of the performance peak for a single time or a short period of time. The capacity is a preset percentage of the maximum performance rate. It is used to constrain the performance peak and balance performance burst with system stability.

[0082] Dynamic configuration and adjustment module: dynamically configures leaky bucket parameters according to storage instance specifications and real-time I / O activities, and makes adaptive adjustments based on historical data to optimize resource allocation strategies; leaky bucket parameter configuration submodule is used to set basic performance rate, maximum performance rate and burst duration parameters;

[0083] The adaptive adjustment submodule is used to monitor I / O activities in real time and dynamically adjust the leaky bucket parameters based on real-time and historical data to adapt to changing workloads.

[0084] Multi-level priority scheduling module: configures personalized leaky bucket parameters for I / O requests of different priorities to ensure that the performance requirements of key businesses are prioritized; priority determination submodule is used to determine the priority of I / O requests based on the requested business type or user-defined rules;

[0085] The personalized parameter configuration submodule is used to configure corresponding leaky bucket parameters for requests of different priorities to ensure that key businesses receive priority processing.

[0086] Scheduled task processing module: Combined with the double leaky bucket structure, it can process requests that exceed the burst performance on a scheduled basis while ensuring the security and integrity of client data;

[0087] The over-performance processing submodule is used to reduce the burst performance by extending the execution time when the I / O request exceeds the burst performance;

[0088] The data security protection submodule is used to ensure that the security and integrity of client data are not affected during the processing.

[0089] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for supporting burst performance of cloud storage, Features: The method comprises the following steps: Construct a dual leaky bucket model, namely the "time window control leaky bucket" and the "performance peak control leaky bucket", to intelligently adjust the burst performance of storage access; Specifically, it includes: a) dynamically configuring leaky bucket parameters, defining the basic performance rate and maximum performance rate, as well as the duration of burst performance according to the storage instance specifications; b) Monitor client I / O requests in real time and dynamically adjust the processing speed through the leaky bucket algorithm model, allowing the performance within the controllable range to exceed the basic performance limit; c) Scheduled tasks are combined with the dual-bucket structure to jointly process requests that exceed burst performance based on the number and size of client I / Os.

2. A method for supporting burst performance of cloud storage according to claim 1, Features: The "time window control leaky bucket" is designed to limit the effective duration of burst performance. The capacity is composed of the maximum performance rate multiplied by the preset burst duration to ensure that the performance improvement is effective within the specified time; the "performance peak control leaky bucket" is designed to constrain single or short-term performance peaks. Its capacity is set to a preset percentage of the maximum performance rate to balance performance bursts and system stability. When the client I / O request exceeds the burst duration / burst performance, the burst performance is reduced by extending its execution time through scheduled tasks.

3. A method for supporting burst performance of cloud storage according to claim 1, Features: Real-time monitoring of client I / O requests also includes: The automated dynamic adjustment mechanism adaptively adjusts leaky bucket parameters based on real-time I / O activities and historical data, optimizes resource allocation strategies, adapts to changing workloads, introduces multi-level priority scheduling, and configures personalized leaky bucket parameters for I / O requests of different priorities, ensuring that the performance requirements of key businesses are prioritized.

4. A method for supporting burst performance of cloud storage according to claim 1, Features: Processing requests that exceed burst performance specifically includes: Scheduled tasks are combined with dual buckets to jointly process requests that exceed burst performance. While ensuring the security of client data and preventing loss, I / O requests that exceed the performance limit can be controlled within the set burst performance and burst duration.

5. A system for supporting cloud storage burst performance according to the method for supporting cloud storage burst performance according to any one of claims 1 to 4, Features: The system includes the following modules: Double leaky bucket model building module: used to build and maintain a double leaky bucket model including "time window control leaky bucket" and "performance peak control leaky bucket" to intelligently adjust the burst performance of storage access; Dynamic configuration and adjustment module: dynamically configures leaky bucket parameters according to storage instance specifications and real-time I / O activities, and makes adaptive adjustments based on historical data to optimize resource allocation strategies; Multi-level priority scheduling module: configures personalized leaky bucket parameters for I / O requests of different priorities to ensure that the performance requirements of key businesses are prioritized; Scheduled task processing module: Combined with the double leaky bucket structure, it regularly processes requests that exceed burst performance while ensuring the security and integrity of client data.

6. A system for supporting sudden performance of cloud storage according to claim 5, Features: The double leaky bucket model building module includes: Time window control leaky bucket submodule: defines the effective duration of burst performance. The capacity is determined by the maximum performance rate and the preset burst duration, which is used to limit the effective time range of burst performance. Performance peak control leaky bucket submodule: sets the upper limit of the performance peak for a single time or a short period of time. The capacity is a preset percentage of the maximum performance rate. It is used to constrain the performance peak and balance performance burst with system stability.

7. A system for supporting sudden performance of cloud storage according to claim 5, Features: The dynamic configuration and adjustment module includes: The leaky bucket parameter configuration submodule is used to set the basic performance rate, maximum performance rate and burst duration parameters; The adaptive adjustment submodule is used to monitor I / O activities in real time and dynamically adjust the leaky bucket parameters based on real-time and historical data to adapt to changing workloads.

8. A system for supporting sudden performance of cloud storage according to claim 5, Features: The multi-level priority scheduling module includes: A priority determination submodule, used to determine the priority of an I / O request according to the requested business type or user-defined rules; The personalized parameter configuration submodule is used to configure corresponding leaky bucket parameters for requests of different priorities to ensure that key businesses receive priority processing.

9. A system for supporting sudden performance of cloud storage according to claim 5, Features: The scheduled task processing module includes: The over-performance processing submodule is used to reduce the burst performance by extending the execution time when the I / O request exceeds the burst performance; The data security protection submodule is used to ensure that the security and integrity of client data are not affected during the processing.

Images