Traffic-Based Process Clustering for Balanced Cloud Workloads
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud computing systems face inefficiencies in clustering virtual machines (VMs) due to uneven workload distribution, leading to reduced performance and availability of cloud services, as current methods fail to optimize VM grouping based on traffic patterns.
Innovation Solution
A method involving hierarchical clustering of VMs based on traffic data, using unsupervised machine learning to form optimized clusters by generating a matrix from traffic data, applying Ward linkage and dendrogram analysis to merge similar processes, and implementing these clusters in cloud platforms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If VMs are clustered using traditional methods, then the clustering process is simple, but the workload distribution becomes uneven leading to reduced performance
Solution Approach 1:
The patent applies preliminary action by collecting and analyzing traffic data between VMs before performing clustering. The system gathers network traffic patterns, communication frequencies, and data transfer volumes in advance, then uses this pre-analyzed information to optimize VM placement. This ensures that VMs with similar traffic patterns are grouped together, achieving even workload distribution and improved cloud service performance without requiring complex real-time adjustments.
Solution Approach 2:
The patent implements parameter changes by using multiple traffic-related parameters (communication frequency, data transfer volume, response time) to dynamically adjust VM clustering. Instead of static clustering based on simple criteria, the system continuously monitors traffic parameters and re-clusters VMs based on changing traffic patterns, thereby optimizing workload distribution and maintaining high performance as system conditions evolve.
2Reliability
If VM clustering is optimized based on traffic patterns, then workload distribution improves, but the computational complexity of clustering increases
Solution Approach 1:
The patent applies self-service by implementing an automated clustering system that autonomously collects traffic data, analyzes patterns, and performs VM re-clustering without manual intervention. The system self-adjusts to changing traffic conditions by continuously monitoring network communications and automatically reorganizing VM groups to maintain optimal workload distribution, thereby ensuring reliability while managing complexity through automation rather than manual processes.
Solution Approach 2:
The patent implements feedback mechanisms by continuously monitoring traffic patterns between VMs and using this information to adjust clustering decisions. The system collects feedback on communication frequencies and data transfer volumes, then uses this feedback to refine VM groupings and achieve more consistent workload distribution. This closed-loop approach ensures that clustering optimizations are based on actual system behavior rather than static assumptions.
3Measurement precision
If hierarchical clustering with Ward linkage is used, then clustering accuracy improves, but processing time increases
Solution Approach 1:
The patent applies segmentation by dividing the VM clustering process into distinct phases: traffic data collection, pattern analysis, cluster formation using Ward linkage, and validation. By segmenting the process, the system can apply computationally intensive hierarchical clustering with Ward linkage only to smaller, manageable subsets of VMs or update clusters incrementally rather than re-clustering all VMs simultaneously, thereby maintaining high clustering accuracy while reducing overall processing time.
Solution Approach 2:
The patent implements partial action by performing hierarchical clustering with Ward linkage on only the most critical or frequently communicating VMs, rather than applying the computationally intensive algorithm to all VMs in the system. This selective approach maintains high clustering accuracy for the most important relationships while reducing overall processing time by excluding less critical VMs from the intensive clustering process.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosure is made of methods, apparatus and system for clustering processes for use by a cloud platform. Process clustering may include receiving traffic data transmitted and received between each pair of processes in a set of processes. A matrix may be generated based on the traffic data, the matrix including a row and a column for each process in the set of processes. The matrix may be hierarchically clustered based on the traffic data, the hierarchical clustering outputting a plurality of clusters, each cluster including one or more processes in the set of processes. The plurality of clusters may then be merged into a set of merged clusters of processes.