Real-time hierarchical selection method of cloud agent for multi-cycle reserved instances

A multi-period and long-period technology, applied in the field of cloud computing, can solve the problems of not considering the price difference between real-time and reserved resources, the applicability of sudden changes is not strong, and the dynamics of demand cannot be reflected, so as to overcome inaccuracy, Overcoming non-real-time and maximizing benefits

Pending Publication Date: 2020-04-28
CHANGCHUN UNIV OF SCI & TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] 2. There is only one reserved instance by default
[0006] 1. For the first method, the prices of reserved instances and real-time instances are the same by default, without considering the price difference between real-time and reserved resources, and cannot reflect the dynamic nature of demand
[0007] 2. There is only one reserved instance by default, which is inconsistent with the actual business situation of cloud computing
[0008] 3. The above methods are all non-real-time, that is, based on user demand prediction and cumulative consumption, and are not very applicable to sudden changes

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
  • Real-time hierarchical selection method of cloud agent for multi-cycle reserved instances
  • Real-time hierarchical selection method of cloud agent for multi-cycle reserved instances

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] Cloud proxy utilizing a single CSP instance is a base case of choice for cloud networking. Such as figure 1 The service provision mechanism of the cloud proxy is described. The cloud proxy multiplexes the dynamic needs of four users, and then provides real-time services to end users at a lower intermediate price by utilizing appropriate reservations and real-time VM instances.

[0047] Different users apply for real-time VM instances from time 1 to 6 (step 1). Agents try to rent resources from CSPs at the lowest price to satisfy user demands (step 2). CSP provides different instances with different prices and different periods τ. Real-time instances are billed at $1 per hour (billing cycle). For reserved instances, the price is a total price for τ billing cycles, which is generally lower than $τ. According to the quantity and duration of the total demand, the agent can make full use of the price advantage and reuse benefits of long-term reservation (step 3). This e...

Embodiment 2

[0050] The present invention assumes that a cloud agent can estimate the total demand in a rather long period T, for any t ∈ (1, T), the total demand at time t is d t . This assumption is reasonable because each user should have a plan for future demand, and total demand can be estimated from the user's plan or forecasted from historical records. Let L=max t (d t ) is the peak demand, then the demand d t into L layers. Define indicators Represents whether there is demand in layer l at time t, that is, if but Otherwise 0.

[0051] For example, figure 2 depicts a demand curve where {τ 0 ,τ 1 ,τ 2 ,τ 3}={1,3,6,10}, T=9. Because Tier 4 has no demand in billing cycle 3, so This billing cycle is referred to as an empty billing cycle. and It is because layer 5 has demand in billing cycle 5. In fact, every Corresponds to a small rectangular shadow block. If the corresponding rectangle lies below the curve, then Otherwise 0. obviously

[0052] Each layer...

Embodiment 3

[0057] Embodiment three, algorithm of the present invention:

[0058] Algorithm: real-time online algorithm

[0059] Enter: d t : The number of instances required to arrive at time t. τ 1 ,τ 2 ,...,τ J : respectively with incremental price c 1 ,c 2 ,...,c J of reserved instance periods

[0060] Output: A real-time resource allocation scheme

[0061] 1: for l=1,...,d t do

[0062] 2:a lj =0

[0063] 3: if there is already an instance to serve d t l or d t l=0then

[0064] 4:break

[0065] 5: end if

[0066] 6: For a period τ j (τ j ≤ T Rl ) instance, calculate a lj ,j=1,2,...,J. Obviously, a l1 ≤a l2 ≤...≤a lJ

[0067] 7: If j exists, making a lj ≥c j then

[0068] 8: i=max{j|a lj ≥c j}, reserve a layer with period τ in layer l i instance of

[0069] 9: else

[0070] 10: Select a real-time instance at layer l

[0071]11: end if

[0072] 12: end for

[0073] d t is the number of instances required to arrive at time t, so d t It is the maxi...

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 a real-time hierarchical selection method of a cloud agent for multi-cycle reserved instances. The method comprises the following steps: S1, collecting or predicting a long period of a user by the cloud agent, stacking demands of each user in the long period to form a total demand, equally dividing the long period into a plurality of moments according to the demands of theuser, and equally dividing the total demand of the user into a plurality of layers; S2, calculating the charging period length of each layer; S3, applying for a plurality of reserved instance periodsfrom the cloud service provider by the cloud agent, and obtaining corresponding prices; and S4, calculating the total price of the layer when the real-time instance period is adopted; comparing the prices of the obtained reserved instance periods, if the price of one or more reserved instance periods is lower than or equal to the total price of the layer, reserving one or more reserved instance periods, and taking the instance with the largest period number in the reserved instance periods as the instance of the layer; and if not, adopting the real-time instance period as the instance of thelayer. The cost of the cloud agent is reduced to the greatest extent and the profit is improved.

Description

technical field [0001] The invention relates to the field of cloud computing, in particular to a real-time hierarchical selection method for multi-period reserved instances by a cloud agent. Background technique [0002] Cost minimization and profit maximization are the two main concerns of cloud network selection. In order to minimize costs. According to the price and type of reserved instances considered, the dynamics of demand, and the real-time nature of the method, there are roughly two types as follows: [0003] 1. The prices of reserved instances and real-time instances are the same by default. El-ghazali Talbi of the French National Institute of Information and Automation has given a variety of heuristic algorithms. Each algorithm differs in the priority evaluation criteria of VM requests, trying to maximize the agent's efficiency by leasing real-time resources to users. profit. This work was extended by Santiago Iturriaga of Universidad Republica, Uruguay, to pr...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06Q30/02
CPCG06Q30/0201G06Q30/0202G06Q30/0283
Inventor 张博宇方明
Owner CHANGCHUN UNIV OF SCI & TECH
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap