Multi-Datacenter Web App Deployment Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current monitoring tools are unable to provide insights into the quality of service (QoS) and quality of experience (QoE) for web applications across multiple datacenters, failing to recommend optimal deployment in terms of cost, service level agreements (SLAs), and responsiveness, especially when applications are executed in different datacenters.
Innovation Solution
A system and method that collect performance measurements from probes and measuring units across multiple datacenters, group them by client locations, compute expected SLAs, and determine optimal deployment to ensure guaranteed SLAs, using an advisory unit to recommend and configure the deployment of web applications across datacenters based on cost, QoS, and QoE considerations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If web applications are deployed across multiple datacenters to improve scalability and redundancy, then system reliability and availability are improved, but the complexity of monitoring and determining optimal deployment locations increases
Solution Approach 1:
The patent segments the monitoring system into distributed probes deployed in each datacenter, with each probe independently collecting local performance measurements. This segmentation allows the complex multi-datacenter monitoring problem to be divided into manageable local measurements that are then aggregated centrally, resolving the contradiction between improved reliability through distribution and the resulting monitoring complexity
Solution Approach 2:
The patent introduces an intermediary advisory unit that acts as a mediator between the distributed probes and the deployment decision-making process. This intermediary consolidates performance measurements from multiple datacenters, computes expected SLAs, and provides optimized deployment recommendations, thereby simplifying the overall system complexity while maintaining multi-datacenter reliability
2Measurement precision
If performance measurements are collected from all datacenters to compute accurate expected SLAs, then measurement precision is improved, but the loss of time for data collection and processing increases
Solution Approach 1:
The patent implements preliminary action by having probes continuously collect and store performance measurements in local data repositories before they are needed for SLA computation. This pre-collection approach ensures that when the advisory unit needs to compute expected SLAs for deployment decisions, the data is already available, thereby maintaining measurement precision without incurring additional data collection time delays
Solution Approach 2:
The patent uses copying by creating virtual representations of performance data through the advisory unit, which consolidates measurements from multiple datacenters into a unified view. This allows the system to work with copied/aggregated data for SLA computations rather than requiring real-time access to all original measurement sources, reducing processing time while maintaining accuracy
3Loss of energy
If web applications are deployed in datacenters far from client locations to reduce infrastructure costs, then cost is reduced, but network transaction time increases
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the deployment configuration based on computed expected SLAs. The advisory unit analyzes performance measurements and client location data to determine optimal deployment parameters, allowing the system to balance infrastructure cost and network transaction time by selecting datacenter locations that provide the best compromise between these competing parameters
Solution Approach 2:
The patent implements dynamics by making the deployment configuration adaptable and changeable based on performance measurements and SLA requirements. The system can dynamically adjust which datacenters host applications by evaluating expected SLAs computed from actual performance data, enabling flexible optimization of both cost and responsiveness rather than being locked into static deployment decisions
Data Source
AI summary
A system for computing an optimal deployment of at least one web application in a multi-datacenter system comprising a collector for collecting performance measurements with regard to a web application executed in the multi-datacenter system and grouping the performance measurements according to locations of a plurality of clients accessing the web application; a data repository for maintaining at least a performance table including at least the performance measurements grouped according to the plurality of client locations and a service level agreement (SLA) guaranteed to clients in the plurality of client locations; and an analyzer for processing at least information stored in the performance table for generating a recommendation on an optimal deployment of the web application in at least one combination of datacenters in the multi-datacenter system by computing an expected SLA that can be guaranteed to the clients in each combination of datacenters.