Active updating method, device and server of physical machine in iaas cloud system

Abstract

The invention discloses an active update method, device and server of a physical machine in an IaaS cloud system, wherein the active update method includes the following steps: acquiring historical fault data of each physical machine in the IaaS cloud system; Analyze and process the fault data to obtain the distribution law of the historical fault data of each physical machine, and predict the next fault data of the corresponding physical machine based on the obtained distribution law; perform an update operation on the corresponding physical machine according to the prediction result, the IaaS cloud system of the present invention The active update method of the physical machine in the system analyzes and models the historical failure data of each physical machine, obtains the occurrence law of the physical machine failure, predicts the next failure point according to the obtained law, and dynamically adjusts the physical machine. Compared with the method of setting a fixed update time or only based on the average value of historical fault data in the prior art, it can better balance the update system overhead and cloud system reliability.

Description

technical field [0001] The invention belongs to the technical field of cloud computing, and relates to resource implementation scheduling and optimization control technology in cloud computing operation, in particular to an active updating method, device and server of a physical machine (Physical Machine, PM) in an IaaS cloud system. Background technique [0002] Cloud computing refers to a computing model that distributes computing on a large number of distributed computers, uses a cloud computing platform, and provides users with information services through the network. Compared with traditional software forms, cloud computing has significant advantages such as loose coupling, on-demand, cost control, resource virtualization, and heterogeneous collaboration, making it more suitable for today's e-commerce, flexible manufacturing, and mobile Internet applications. [0003] Cloud computing includes two meanings: one is the underlying cloud computing platform infrastructure, ...

AI Technical Summary

Problems solved by technology

[0006] The above post-event repair measures have certain shortcomings: there is still a period of delay from the failure of the PM to the detection of the failure by the system, so there may be an error of "the system continues to dispatch tasks to the PM that has failed"; The state is different from the state represented by the saved VM image at the latest archive point, so after the PM is repaired, the involved VM is actually restored to a certain moment before the failure rather than the state at the time of the failure, which actually leads to useless work; The system overhead of repairing after a failure is large, and has a large negative impact on the performance of the IaaS cloud system

However, due to the dynamic variability of system load and reliability, updating at fixed intervals is often difficult to meet the requirements of low update operation overhead and high system reliability;

[0011] (2) Lack of mechanism for quantitative trend prediction

The existing IaaS cloud maintenance technology does not fully analyze and model the historical fault data of the system, but mostly uses historical average or recent historical fault data as the basis for control decision-making;

[0012] (3) The difference between different PMs is not considered

The existing technology treats all PMs in the IaaS cloud system equally, adopts the same control strategy and control time interval, and does not consider the differences in load, execution efficiency, and reliability of different PMs

Images