	US 12,273,279 B2
	Cloud software service resource allocation method based on QoS model self-correction
Xing Chen, Fujian (CN); Haijiang Wang, Fujian (CN); and Fangning Zhu, Fujian (CN)
Assigned to FU ZHOU UNIVERSITY, Fujian (CN)
Appl. No. 17/603,269
Filed by FU ZHOU UNIVERSITY, Minhou County (CN)
PCT Filed Jul. 30, 2019, PCT No. PCT/CN2019/098255 § 371(c)(1), (2) Date Oct. 12, 2021, PCT Pub. No. WO2020/228143, PCT Pub. Date Nov. 19, 2020.
Claims priority of application No. 201910403638.2 (CN), filed on May 15, 2019.
Prior Publication US 2022/0210028 A1, Jun. 30, 2022
Int. Cl. H04L 41/14 (2022.01); H04L 41/5009 (2022.01); H04L 43/55 (2022.01); H04L 47/70 (2022.01); H04L 47/83 (2022.01)

CPC H04L 47/83 (2022.05) [H04L 41/145 (2013.01); H04L 41/5009 (2013.01); H04L 43/55 (2022.05); H04L 47/82 (2013.01)]

5 Claims

1. A cloud software service resource allocation method based on QoS model self-correction, characterized in that the method comprises the following steps:

step S1: performing online self-learning to obtain a QoS prediction model;

step S2: collecting runtime data under a certain workload, and improving the accuracy of the QoS prediction model under a current workload through self-correction control;

step S3: constructing a fitness function in combination with a quality of service (QOS) and a cloud resource cost (Cost), and searching for a target resource allocation scheme by using the following steps of:

step S31: assuming that there are m types of virtual machines, a quantity of each type of virtual machines is expressed as vm_l, wherein 1≤l≤m, then defining the particle coding of a certain resource allocation scheme as the following formula:

VM={vm₁,vm₂,vm₃, . . . ,vm_m} (10);

initializing a values of related parameters, including but not limited to a particle swarm size, a maximum number of iterations and an initial particle swarm size;

step S32: calculating a fitness value of each particle according to the following formula, setting each particle as its own personal best particle, and setting the particle with a minimum fitness value as a global best particle:

wherein r₁denotes trade-off of the quality of service (QOS), and r₂denotes trade-off of the cloud resource cost;

step S33: updating the particles according to mutation and crossover calculations, and recalculating a fitness value of each particle;

step S34: if the fitness value of the updated particle is better than its personal best particle, replacing the personal best particle with the updated particle;

step S35: if the fitness value of the updated particle is better than a global best particle, replacing the global best particle with the updated particle;

step S36: returning to step S33 until a stop condition is met;

step S4 specifically comprises the following steps of:

step S4: comparing a current resource allocation situation with the searched target resource allocation scheme to obtain a difference there between, and then adjusting resources according to a proportion;

the step S4 comprises the steps of:

step S41: using formula (12) to denote the allocated resources situation VM_allocated, while using formula (13) to represent the target allocation scheme VM_Objectiveobtained in each circulation, and denoting the difference VM_Differencebetween the allocated resource situation VM_allocatedand the target allocation scheme VM_objectiveby formula (14):

VM_allocated={vm₁^A,vm₂^A,vm₃^A, . . . vm_m^A} (12);

VM_Objective={vm₁^O,vm₂^O,vm₃^O, . . . ,vm_m^O} (13);

VM_Difference={vm₁^Λ−vm₁^O,vm₂^Λ−vm₂^O,vm₃^Λ−vm₃^O, . . . ,vm_m^Λ−vm_m^O} (14);

step S42: calculating the overall change trend between the allocated resource situation and the target allocation scheme in each circulation by using the following formula:

when VM_sum>0, it denotes that virtual machine resources of the cloud software service VM sum as a whole need to be increased, and the quantity of virtual machines are not reduced at this time; and when VM_sum<0, it denotes that the virtual machine resources of the cloud software sum service have a remaining computing power, and the quantity of the virtual machines needs to be appropriately reduced instead of increasing the quantity of the virtual machines at this time;

step S43: assuming that the resource adjustment proportion is Proportion, and 0<Proportion<1, in order to ensure that the quantity adjusted each time for each type of the virtual machines is an integer, performing a rounding operation, as shown in formula (16), it is the quantity of the lth type of the virtual machines that needs to be adjusted, and, when VMAdjustement>0, it denotes adding VMAdjustement virtual machines of this type: on the contrary, VMAdjustement virtual machines of this type are reduced; wherein Intpart( ) is a rounding function;

VM_adjustement(l)=Intpart(VM_Difference(l)*Proportion) (16);

step S5: repeating steps S2 to S4 until the current resource allocation situation is the same as the target resource allocation scheme which means that resource adjustment is completed.