Distributed Multicloud Service Placement Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems lack an efficient method for automatically matching clouds to services considering varying properties of available clouds, leading to challenges in optimizing service placement across multiple clouds, especially in hybrid environments with diverse capabilities and requirements.
Innovation Solution
A service-placement engine is introduced, comprising cloud observers and placement units that acquire and aggregate cloud information, including compliance, capability, and resource availability, to select suitable clouds for hosting services, with connectivity schemes for efficient data communication and intercloud coordination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple public cloud providers are used in a hybrid cloud with varying prices and capabilities, then service placement flexibility and capability diversity are improved, but the complexity of comparing and matching clouds to services increases
Solution Approach 1:
The patent transforms the complex cloud comparison problem into a standardized parameter matching process. Each cloud is characterized by specific parameters (capability vectors, compliance vectors, resource availability), and services are matched to clouds based on parameter compatibility. This parameter-based approach converts qualitative cloud differences into quantitative comparisons, resolving the complexity while maintaining flexibility.
Solution Approach 2:
The patent segments the cloud comparison task into distinct components: capability assessment, compliance verification, and resource availability checking. Each aspect is evaluated separately using dedicated vectors and algorithms, then integrated to form the overall placement decision. This segmentation reduces the cognitive and computational complexity of comparing heterogeneous clouds.
2Productivity
If deployment environments are shared among many tenants with constant flux in capability and capacity, then resource utilization efficiency is improved, but the reliability of service placement decisions deteriorates
Solution Approach 1:
The patent performs preliminary characterization of each cloud's capabilities, compliance attributes, and resource capacity before service placement decisions are made. This advance profiling creates a stable reference framework that remains valid even as tenant workloads fluctuate. The system continuously updates resource availability while maintaining the core capability model, ensuring reliable placement decisions in dynamic environments.
Solution Approach 2:
The patent implements continuous monitoring and feedback mechanisms that track actual cloud performance and resource availability. This feedback loop allows the system to adjust placement decisions in real-time based on observed conditions, maintaining reliability despite the dynamic nature of multi-tenant environments. The feedback also enables predictive adjustments to prevent future placement conflicts.
3Adaptability or versatility
If different types of public and private clouds require different placement rules, then customization for specific cloud requirements is improved, but the ease of automated service placement deteriorates
Solution Approach 1:
The patent creates a universal placement framework that can handle multiple cloud types through a common interface. The system uses standardized vector representations for cloud capabilities, compliance, and resources that work across public and private clouds alike. Different cloud-specific requirements are expressed as variations of the same underlying parameters, allowing a single automated algorithm to serve diverse cloud environments without requiring cloud-type-specific logic.
Data Source
AI summary
Several cloud observers monitor a set of clouds to collect cloud information and communicate the information to several service placement units thus forming a distributed service-placement system. The service-placement units communicate with a population of clients to receive service-assignment requests and select at least one compatible cloud for each request. The cloud observers share the cloud-monitoring workload and the service-placement units share the cloud-assignment workload. According to a first connectivity scheme, each cloud observer has a path to each service-placement unit. According to a second connectivity scheme, the cloud observers are interconnected to pool cloud information so that each cloud observer possesses cloud information of all clouds. Each cloud observer communicates with a respective subset of the service-placement units. According to a third connectivity scheme, cloud information is communicated through an intermediate stage of multicast distributors, each coupled to a respective subset of service-placement units.