Virtual Desktop Session Placement Using Predicted Logoff Times
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing remote desktop virtualization systems inefficiently manage session placement on host servers, leading to underutilization as servers remain powered on due to individual sessions terminating at different times, resulting in excessive energy consumption.
Innovation Solution
Implement a mechanism to group remote desktop sessions based on predicted user logoff times, placing sessions with similar logoff times together on servers, allowing for more efficient server utilization by powering off servers when all sessions are likely to terminate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If sessions are placed on host servers in round-robin or static order, then session distribution is simple and predictable, but servers cannot be powered off until all sessions are terminated leading to increased power consumption
Solution Approach 1:
The system performs preliminary actions by predicting user logoff times before sessions actually terminate. The session placement manager analyzes historical session data and user behavior patterns to forecast when users will log off, then proactively places sessions on servers that will be available and optimal at those predicted times. This allows servers to be powered down in advance when no sessions are expected, reducing power consumption while maintaining simple operational complexity.
2Productivity
If multiple servers remain powered on to service few sessions, then session service availability is maintained, but server utilization efficiency decreases
Solution Approach 1:
The session placement system dynamically adjusts session distribution across servers based on real-time and predicted session termination patterns. Rather than using static round-robin placement, the system continuously monitors session states, predicts future logoff times, and redistributes sessions to optimize server utilization. This dynamic approach maintains service availability by ensuring sessions are always placed on active servers while improving productivity by consolidating sessions and enabling timely server shutdowns.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring actual session termination times and comparing them against predictions. The session placement manager uses this feedback to refine its prediction models and adjust future session placement decisions. This closed-loop control ensures that server utilization is optimized while maintaining reliable session service availability, as the system learns from past performance to improve future decisions.
3Ease of operation
If sessions are spread across multiple servers, then load distribution is balanced, but the number of servers that can be powered off is reduced
Solution Approach 1:
The system performs preliminary consolidation of sessions onto fewer servers based on predicted logoff patterns. By analyzing when sessions are expected to terminate, the system proactively moves sessions to be concentrated on specific servers that will remain active, allowing other servers to be powered down. This maintains ease of operation through automated management while significantly reducing energy consumption by minimizing the number of active servers.
Data Source
AI summary
Systems and methods are provided for grouping remote desktop sessions on servers in remote desktop environments according to predicted user logoff times so that sessions with similar predicted logoff times can be placed together on the servers, allowing for more efficient utilization of servers. A user can request a virtual desktop session. Once the request is received, a predicted logoff time for the session can be determined and the session can be placed on one of available host servers based on the predicted logoff time. Different logoff time ranges can be assigned to different servers and sessions can be grouped on the servers according to predicted logoff time.