Operating system bus and balancing method supporting online expansion and retraction of components

Abstract

The invention provides an operating system bus and a balancing method supporting online expansion and retraction of components. An operating system bus structure comprises a component management module, a queue management module, a massage router, a node monitoring module, a component processing cluster and a component client side. The balancing method comprises an initialization stage and an operation stage. Compared with the prior art, the operating system bus and the balancing method supporting online expansion and retraction of the components have the advantages of overcoming the defects that the prior art cannot support the components to deploy cluster dynamics and rapid expand and retract, solving the load inclination problem caused by heterogeneous nodes and hot spot data and improving a loose coupling and componentized cloud operating system architecture.

Description

technical field [0001] The invention relates to the technical field of cloud computing, in particular to a cloud operating system bus and an equalization method that supports online scaling of stateful component clusters. Background technique [0002] With the rise of cloud computing, traditional data centers are rapidly transforming into cloud data centers. The cloud operating system plays an important role in the four stages of data center evolution from primary form to advanced form: physical resource base integration, physical and application virtualization integration, automated centralized management, and data center collaborative services. The cloud operating system is responsible for the intermediate functions of the upper interface application and the lower scheduling management hardware. It integrates a large number of heterogeneous devices into a standard and unified logical resource pool, and dynamically schedules them for cloud applications to complete services ...

AI Technical Summary

Problems solved by technology

[0006] Lack of an efficient queue group routing mechanism, the efficiency of the keyword-based tree routing algorithm is affected by the expansion of the keyword scale, and the routing algorithm based on the Hash function will generate a large amount of data movement overhead when nodes join and exit;

[0007] For component processing clusters, there is a lack of initialization data distribution methods and runtime adjustment methods to balance the load of heterogeneous nodes

Images