Topology-Based Management Broker for Cloud Service Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cloud computing systems face challenges in designing, provisioning, and managing cloud services due to the lack of intuitive association of policies with physical topologies, making it difficult to describe infrastructure architectures and application models, leading to inefficient deployment and management processes.
Innovation Solution
The introduction of architecture-descriptive topologies that define the physical architecture of cloud services, enabling the association of policies and lifecycle management actions with nodes, groups, and the entire topology, allowing for automated deployment and management through a topology-based management broker that supports both blueprints and architecture-derived topologies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If manual deployment processes are used for cloud services, then flexibility in deployment is maintained, but administrative time and human errors increase significantly
Solution Approach 1:
The patent uses blueprints as abstract copies or models that represent the desired cloud service topology. These blueprints define nodes, connections, and policies in a standardized format that can be automatically interpreted by the management broker, enabling automated deployment without manual intervention while maintaining the intended architecture through the blueprint model.
Solution Approach 2:
The topology-based management broker acts as an intermediary between the blueprint (design model) and the actual cloud infrastructure deployment. It automatically translates blueprint definitions into deployment actions, managing the provisioning and configuration of cloud services based on the abstract topology model, thereby eliminating manual deployment steps.
2Productivity
If policies are not associated with physical topologies, then infrastructure flexibility is maintained, but policy enforcement and management become inefficient
Solution Approach 1:
The patent segments the cloud infrastructure management into distinct topological elements (nodes, connections, groups) that can be independently defined in blueprints and individually associated with policies. This segmentation allows policies to be applied at specific topological levels rather than requiring management of the entire infrastructure as a monolithic unit, improving both efficiency and manageability.
Solution Approach 2:
The blueprint model serves as a universal representation that can describe various cloud infrastructure topologies (networks, storage systems, computing clusters) using the same nodal structure and policy association mechanisms. This universal model allows consistent policy enforcement across different types of cloud services and infrastructure components.
3Adaptability or versatility
If cloud services are deployed without architecture-descriptive topologies, then deployment speed may be faster initially, but scalability and resource utilization improve significantly with topology-based management
Solution Approach 1:
The patent introduces a topological dimension to cloud service management by representing infrastructure as a graph of nodes and connections. This topological view adds a new layer of abstraction that enables scalability through hierarchical grouping of nodes and systematic policy propagation across multiple levels of the topology, allowing complex architectures to be managed through structured patterns rather than individual component configuration.
Data Source
AI summary
In one implementation, a method for topology based management of second day operations can include identifying a cloud service operating on a second system, discovering, via a first system, an existing realized topology of the second system as an inferred realized topology for the first system, wherein the existing realized topology is provisioned by the second system, defining a management process to be performed within the cloud service, via the first system, upon the instantiation of the inferred realized topology by the first system, and executing the management process utilizing the first system.