Server Selection via Historical RTT Sorting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for selecting destination servers do not effectively account for geographical distance, leading to lessened user experience due to inefficient connection choices, and often rely on bandwidth-consuming active probing and additional server data.
Innovation Solution
A method that sorts destination servers based on historical round-trip times (RTTs) to identify the most efficient server for a client, using existing server-related information like IP addresses and RTTs without active probing, thereby reducing bandwidth consumption and overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If active probing and additional server data are used to select destination servers, then selection accuracy is improved, but bandwidth consumption and system overhead increase
Solution Approach 1:
The system performs preliminary actions by collecting RTT data during normal server operations and pre-processing this data to create sortable lists. Historical RTT measurements are accumulated over time and stored for later use, so that when server selection is needed, the sorting can be done immediately using pre-collected data without requiring additional probing at the moment of selection
Solution Approach 2:
The patent creates a copy of the server list that is sorted based on historical RTT data. Instead of modifying the original server configuration or performing active probes on the actual servers, the system works with a copied and sorted representation of the server list, allowing efficient selection without impacting the actual servers or consuming additional bandwidth
2Adaptability or versatility
If destination servers are geographically dispersed to serve multiple users, then service coverage is improved, but connection efficiency decreases due to distance
Solution Approach 1:
The system applies local quality by sorting servers based on their geographical proximity to each client. Each client receives a customized sorted list where servers are ordered according to their local RTT characteristics. This means that while servers remain geographically dispersed to provide broad coverage, each individual client experiences optimized connection speed by receiving a list prioritized for their specific location
Solution Approach 2:
The system changes the parameter used for server selection from static geographic location to dynamic RTT measurements. By collecting actual RTT data between clients and servers and using this measured parameter for sorting, the system adapts to varying network conditions and provides optimal server selection that accounts for both geographic dispersion and actual connection performance
3Speed
If historical RTT data is collected and sorted for server selection, then connection speed is improved, but data processing complexity increases
Solution Approach 1:
The system implements self-service by having the client device perform the sorting operation locally using pre-collected RTT data. Instead of requiring a complex centralized system to manage and distribute sorted server lists, each client independently sorts the server list based on their own historical RTT measurements, reducing the complexity of the overall system while maintaining fast connection selection
Data Source
AI summary
One or more techniques and/or systems are disclosed for identifying a list of destination servers ordered by round-trip time (RTT) for a requesting machine. When a client interacts with an address, such as an IP address, when accessing a remote service, an RTT is collected for the address, or the RTT may be derived for a range of addresses, and saved in a database. When the client subsequently requests access to a remote service, a list of destination servers are returned, respectively identified by corresponding addresses. The destination servers can be sorted in order of efficiency by determining an expected RTT associated with a matching address, or range, in the database, and sorting the destination servers by their respective expected RTTs.