Data Center Selection for Conference Calls
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online conferencing systems lack efficient data center selection methods that consider participant interaction patterns, leading to suboptimal network performance and user experience during conference calls.
Innovation Solution
A method and system that analyze previous media interactions between participants to determine weighting factors and select an initial data center based on these patterns, optimizing network performance and user experience by dynamically re-selecting data centers during calls based on participant behavior.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data center selection is based on simple geographic proximity, then implementation complexity is low, but network performance and user experience are suboptimal
Solution Approach 1:
The system performs preliminary analysis of participant interaction patterns and media sharing behaviors before the conference call begins. Historical data from previous communications is processed in advance to pre-determine weighting factors and identify likely media sharers, enabling proactive data center selection rather than reactive adjustment during the call
Solution Approach 2:
The system continuously monitors actual participant behavior during the conference call and compares it against predicted patterns. When deviations are detected (e.g., a participant not expected to share media begins sharing), the system provides feedback to the data center selection mechanism, triggering dynamic re-selection to optimize network performance in real-time
2Ease of operation
If data center selection considers participant interaction patterns, then user experience is improved, but processing time and computational resources increase
Solution Approach 1:
Participant profiles including interaction patterns, media sharing preferences, and communication behaviors are built and stored in advance based on historical data. Before each conference call, the system quickly retrieves these pre-analyzed profiles rather than performing comprehensive analysis in real-time, significantly reducing selection processing time while maintaining personalized optimization
Solution Approach 2:
The system focuses computational resources on analyzing only the most relevant factors for data center selection, such as identifying the single or top few participants most likely to share media, rather than comprehensively analyzing all participant attributes. This partial analysis approach achieves sufficient optimization without exhaustive processing
3Productivity
If dynamic re-selection of data centers is performed during calls, then network efficiency is optimized, but system complexity and computational overhead increase
Solution Approach 1:
The data center selection mechanism transitions from a static, one-time decision to a dynamic process that adapts during the conference call. The system continuously monitors participant behavior, detects deviations from predicted patterns, and triggers re-selection only when necessary, balancing optimization benefits against the complexity of maintaining dynamic adjustment capabilities
Solution Approach 2:
The system changes key parameters such as weighting factors in the geographic centroid calculation based on observed participant behavior. When a participant begins sharing media unexpectedly, the system adjusts the weighting factors to reflect current realities, thereby changing the optimal data center selection without requiring complete system reconfiguration
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
The patterns of usage, participation, and/or item sharing by the participants in a conference call are correlated, recognized, and used to select a data centers. This selection can be done both before and/or during a call. When significant patterns are recognized (e.g., heavy video sharing by one participant at every Friday afternoon meeting) there may be a better (e.g., faster, better quality, more efficient, less use of network resources) data center to host the call than one that is selected based on the geography of the participants alone. Likewise, during a call, other patterns of behavior can be correlated and recognized dynamically such that when these patterns occur, the call is hosted by a different (i.e., better) data center.