Agile elastic telescoping method in cloud environment

An elastic scaling, cloud environment technology, applied in the direction of climate sustainability, resource allocation, multi-programming devices, etc., can solve problems such as inability to guarantee service quality

Active Publication Date: 2013-12-04
ZHEJIANG UNIV
View PDF5 Cites 54 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The present invention aims at the hysteresis existing in the elastic scaling strategy in the prior art, and the shortcomings that the

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Agile elastic telescoping method in cloud environment
  • Agile elastic telescoping method in cloud environment
  • Agile elastic telescoping method in cloud environment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0090] 1. A first-level prediction algorithm based on the ARIMA model, which uses the periodicity and trend of the application load to predict the load of the next time slice. The algorithm flow chart is as follows figure 1 As shown, the specific operation steps are as follows:

[0091] Step 1.1. Obtain application load data. It can be obtained directly from the load balancer, or obtained by analyzing log files.

[0092] Step 1.2, calculate the sample value. Sort the collected load data of the latest time slice, and use its quantile statistics as a new sample and put it into the historical load.

[0093] Step 1.3, sample stabilization. The load samples of the latest N time slices are taken out, and the periodic difference operation is performed according to the formula (10). Then follow the formula (5) to carry out the differential operation until W t is a stationary sequence.

[0094] Step 1.4, put W t As training data, ARMA is ordered according to the AIC criterion. ...

experiment example 1 2

[0109] Experimental Example 1 Validation of the effect of the second-level prediction algorithm

[0110] The original application load data records the number of requests per second, and we perform average and quantile statistics for different time periods. Such as image 3 As shown, we performed average, 90%, 95%, and 99% quantile statistics for each hour. It can be seen from the statistical chart that the quantile statistical value is much higher than the average value most of the time, which can better reflect the actual resource demand of the application in a certain period of time.

[0111] Figure 4 It is the data of a certain three hours in the above application load (the mean and quantile statistics are performed in units of 5 minutes). It can be seen that the load applied within 1 hour also fluctuates greatly. If the predicted load value is predicted in units of hours, the maximum required amount of resources will be allocated to this time period, resulting in was...

experiment example 2

[0115] Experimental example 2 Verification effect of agile elastic scaling method

[0116] Using the first-level forecast (95% quantile) with a cycle T=1h, combined with the load / resource model, the resource allocation results are as follows Figure 9 , it can be seen that the forecast resource fits the load curve well. However, in order to reduce SLA errors caused by underestimation of resources, it is usually necessary to allocate some additional resources according to the forecast results. Figure 10 When adding additional resources in different proportions, the corresponding SLA error. It can be seen that when no additional resources are used, only 97.374% of the service quality can be guaranteed, and 99.5% of the service quality can be achieved by adding 50% of the additional resources.

[0117] However, for the coarse-grained resource allocation with a cycle T=1h, in order to ensure the quality of service, resources will be allocated with the maximum load in this time ...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention relates to the field of elastic computing of cloud computing, and discloses an agile elastic telescoping method in a cloud environment. The agile elastic telescoping method includes the specific steps: forecasting the load of a next time slice according to historical load data of a data center through an ARIMA (autoregressive integrated moving average) model and an ARMA (autoregressive moving average) model by taking the time slice as a cycle; performing saving operation and restoring operation on a virtual machine, saving the memory state of the virtual machine by the saving operation to hang up the virtual machine, and then restoring the memory state of the virtual machine by the restoring operation to restore use of the virtual machine; hanging up one or a plurality of application-ready virtual machines or rapidly placing the virtual machines into service through the forecasted load of the data center obtained by the load forecasting step and by the aid of the rapid supply step of the virtual machines to dynamically adjust resources of application clusters of the data center. The agile elastic telescoping method has the advantages that the sizes of the clusters are adjusted in real time according to current conditions of the application clusters, and energy consumption of the data center is reduced.

Description

technical field [0001] The invention relates to the elastic computing field of cloud computing, in particular to an agile elastic scaling method in a cloud environment. Background technique [0002] With the advent of the cloud computing era, the high energy consumption of data centers has gradually become a prominent problem. Since 2011, the energy consumption of global data centers has approached 100,000,000,00KWh, and produced more than 40,568,000 tons of CO2. However, the resource utilization rate of most data centers is only 20-30%. Cloud computing leases data center resources to users in a pay as you go way, allowing users to customize the right amount of resources according to current needs. The elastic scaling technology enables the scale of the application cluster to be dynamically adjusted according to the implementation load of the application, thereby saving resources to the greatest extent while ensuring the quality of service. Combined with server integration...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): G06F9/50G06F9/455H04L29/08
CPCY02B60/142Y02B60/167Y02B60/146Y02D10/00
Inventor 尹建伟杨勇吴朝晖邓水光李莹吴健
Owner ZHEJIANG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products