Predictive Desktop Workload Migration Across Data Centers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Increased network congestion and bandwidth constraints in cloud-based networks lead to higher end-user latencies and reduced quality of service when accessing virtual desktops, particularly when users travel to different geographical locations.
Innovation Solution
A system that migrates an end-user's desktop workload across multiple data centers based on predicting future location changes using calendar information, selecting the data center with the shortest geographic distance, lowest latency, or lowest cost, and migrating the workload before the user arrives at the new location.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If the virtual desktop is hosted in a fixed data center, then the system complexity is reduced, but the end-user latency increases when users travel to different geographical locations
Solution Approach 1:
The patent implements dynamic data center reassignment based on user location. The system continuously monitors user travel patterns and automatically reassigns virtual desktops to different data centers as users move between geographical locations, transforming the static hosting model into a dynamic one that adapts to user mobility patterns
Solution Approach 2:
The system performs preliminary actions by predicting user travel plans using calendar information and proactively migrating workloads to appropriate data centers before users arrive at new locations. This anticipatory approach prevents latency issues rather than reacting to them after they occur
2Loss of time
If the workload is migrated frequently to follow user location changes, then the end-user latency is reduced, but the network bandwidth consumption increases
Solution Approach 1:
The system applies partial migration by selectively moving only the necessary workload components to new data centers based on predicted user needs, rather than migrating entire virtual desktop environments. This reduces unnecessary network traffic while still achieving the latency reduction benefit
Solution Approach 2:
The system creates predictive copies of workload data in advance at target data centers based on calendar analysis, allowing users to access their virtual desktops from multiple locations without requiring complete real-time synchronization, thereby reducing network bandwidth consumption
3Measurement precision
If the system uses calendar information to predict user location changes, then the migration accuracy is improved, but the processing complexity increases
Solution Approach 1:
The system leverages existing calendar applications that users already maintain for their own scheduling purposes. By integrating with these self-maintained calendars, the system obtains accurate location prediction data without requiring additional user input or complex tracking mechanisms, thus improving accuracy while minimizing processing complexity
Data Source
AI summary
A computer system includes a client device, geographically distributed data centers and a server. The client device remotely accesses a virtual desktop, with the virtual desktop configured to run and store a workload for an end-user of the client device. One of the data centers is assigned to host a virtual desktop for the client device based on a current location of the end-user. The server determines an indication of a future change in location of the end-user from the current location to a target location that is different from the current location. The server further determines which data center is to be reassigned to host the virtual desktop in response to the determined indication, and cooperates with the data centers to migrate the workload to the reassigned data center in response to travel of the end-user to the target location.