Tenant Clustering via Knowledge Graph Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multitenant computing environments, existing tenant placement strategies do not account for tenant characteristics, leading to inappropriate placement and security vulnerabilities, such as competition for limited resources and potential data corruption among tenants.
Innovation Solution
An intelligent tenant management system that constructs a tenancy knowledge graph, transforms tenant properties into vectors, and clusters tenants based on similarity, enabling informed placement, runtime tuning, and security management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If tenants are placed on shared host machines without considering tenant characteristics, then resource utilization is improved, but security vulnerabilities and resource competition increase
Solution Approach 1:
The patent applies local quality by creating tenant-specific placement decisions based on individual tenant characteristics. The system analyzes each tenant's properties (workload, security requirements, resource usage patterns) and determines optimal placement on specific host machines or containers, rather than applying a uniform placement strategy. This allows different tenants to have different placement qualities suited to their specific needs, resolving the contradiction between resource utilization and security.
Solution Approach 2:
The patent segments tenants into different groups or categories based on their characteristics and requirements. By dividing the tenant population into segments with similar profiles, the system can apply targeted placement strategies to each segment, placing compatible tenants together while isolating those with conflicting requirements. This segmentation approach enables both high resource utilization through efficient packing and enhanced security through appropriate isolation.
2Device complexity
If tenants are placed without considering tenant characteristics, then placement complexity is reduced, but resource competition and data corruption risks increase
Solution Approach 1:
The patent applies preliminary action by performing comprehensive tenant characteristic analysis and compatibility assessment before actual tenant placement. The system pre-evaluates tenant properties, identifies potential conflicts, and determines optimal placement configurations in advance. This preliminary analysis prevents resource competition and data corruption by ensuring compatible tenants are placed together before they actually share resources, reducing placement complexity while avoiding harmful interactions.
Solution Approach 2:
The patent implements feedback mechanisms that monitor tenant behavior and resource usage patterns after placement. The system continuously gathers data on tenant performance, resource consumption, and interaction patterns, then uses this feedback to refine placement decisions. This feedback loop enables the system to learn from actual tenant characteristics and improve placement strategies over time, preventing resource competition while maintaining manageable complexity.
3Ease of operation
If traditional tenant placement strategies are used, then system simplicity is maintained, but operational efficiency and security monitoring improve
Solution Approach 1:
The patent applies self-service by enabling the system to automatically perform tenant placement decisions based on analyzed tenant characteristics. The placement engine autonomously evaluates tenant properties, determines optimal configurations, and executes placement without requiring manual intervention. This self-service capability improves operational efficiency by rapidly making informed placement decisions while keeping system complexity manageable through automation rather than manual processes.
Data Source
AI summary
A computer-implemented method includes constructing a tenancy knowledge graph having a plurality of tenant nodes representing respective tenants in a multitenant computing environment, a plurality of property nodes representing respective properties of the tenants, and a plurality of edges connecting the plurality of tenant nodes and the plurality of property nodes, transforming the plurality of property nodes to corresponding property vectors, performing random walks starting from the plurality of tenant nodes of the tenancy knowledge graph, feeding sequences of nodes traversed by the random walks into a neural network to generate a plurality of tenant vectors corresponding to the plurality of tenant nodes, and clustering the plurality of tenant nodes into one or more tenant clusters based on similarity of the plurality of tenant vectors.