Server group selection system, server group selection method, and program
AssigneeNTT, Inc.
InventorsKatsumi Fujita, Masashi Kaneko
Life of this patent
Estimating the term…
Abstract
A system (S) includes: a classification unit (21) that classifies a plurality of physical servers into a plurality of clusters based on parameters collected from the physical servers; and a destination selection unit (22) that selects, from the plurality of clusters classified by the classification unit (21), any of clusters capable of ensuring resources required for placing virtual machines or containers to be added as a placement destination. Accordingly, the system (S) can determine a server group as the placement destination based on the failure factors of the physical servers.
Claims20
A server group selection system comprising one or more processors configured to perform operations comprising:
- collecting, over a network and from each of a plurality of physical servers, parameters indicative of failure probability of the physical server, wherein the parameters comprise (i) a cumulative operation time of the physical server and a cumulative storage write count of the physical server;
- performing multi-dimensional clustering to classify the plurality of physical servers into a plurality of clusters based on the parameters collected from the physical servers;
- selecting, from the plurality of clusters classified by the classification unit, any of a subset of clusters capable of ensuring that satisfy resource required requirement for placing virtual machines or containers to be added as a placement destination;
- and
- deploying the virtual machines or containers on one or more of the physical servers in the selected subset of clusters for execution;
- wherein the one or more processors are configured to preferentially set, as the placement destination, a cluster on which no active virtual machine or container is installed, the cluster having a lowest failure probability among the clusters in the plurality of clusters that have no active virtual machine or container installed.
A server group selection method performed by a computer system comprising one or more processors, the method comprising:
- collecting, over a network and from each of a plurality of physical servers, parameters indicative of failure probability of the physical server, wherein the parameters comprise (i) a cumulative operation time of the physical server and (ii) a cumulative storage write count of the physical server;
- performing multi-dimensional clustering to classify the plurality of physical servers into a plurality of clusters based on the parameters collected from the physical servers;
- selecting, from the plurality of clusters, a subset of clusters that satisfy resource requirement for placing virtual machines or containers to be added as a placement destination;
- and
- deploying the virtual machines or containers on one or more of the physical servers in the selected subset of clusters for execution;
- wherein the one or more processors are configured to preferentially set, as the placement destination, a cluster on which no active virtual machine or container is installed, the cluster having a lowest failure probability among the clusters in the plurality of clusters that have no active virtual machine or container installed.
A non-transitory computer-readable storage medium storing a program for causing a computer comprising one or more processors to execute operations comprising:
- collecting, over a network and from each of a plurality of physical servers, parameters indicative of failure probability of the physical server, wherein the parameters comprise (i) a cumulative operation time of the physical server and (ii) a cumulative storage write count of the physical server;
- performing multi-dimensional clustering to classify the plurality of physical servers into a plurality of clusters based on the parameters collected from the physical servers;
- selecting, from the plurality of clusters, a subset of clusters that satisfy resource requirement for placing virtual machines or containers to be added as a placement destination;
- and
- deploying the virtual machines or containers on one or more of the physical servers in the selected subset of clusters for execution;
- wherein the one or more processors are configured to preferentially set, as the placement destination, a cluster on which no active virtual machine or container is installed, the cluster having a lowest failure probability among the clusters in the plurality of clusters that have no active virtual machine or container installed.
Dependencies are inferred from the claim text.