Web Server Resource Constraint Profiling via Distributed Flash Crowd Simulation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Web server providers face challenges in determining resource constraints to handle unexpected increases in traffic, as traditional benchmarking approaches fail to simulate real-world Internet conditions and cannot assess the impact of access bandwidth on performance.

Innovation Solution

A system where a coordinator instructs multiple clients to transmit synchronized requests to a target server, allowing for the isolation and measurement of performance characteristics such as network access bandwidth, data processing capabilities, and HTTP request processing time, by executing stages that specifically target different server sub-systems.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If traditional benchmarking approaches are used to test web servers, then controlled local area network settings can be achieved, but effects of wide-area conditions and actual Internet connectivity characteristics cannot be revealed

Engineering Contradiction:
Improvetesting accuracyVSAvoidreal-world condition simulation
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent introduces a mediator system that coordinates between multiple geographically distributed clients and the target web server. This mediator manages the flash crowd simulation by distributing requests across different locations, thereby revealing wide-area network effects and Internet connectivity characteristics that cannot be observed in controlled LAN environments.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent transitions testing from a single-dimensional controlled LAN environment to a multi-dimensional real-world Internet environment by distributing clients across different geographic locations, network providers, and connection types. This dimensional expansion enables observation of wide-area network effects, varying connectivity characteristics, and real-world performance scenarios.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Measurement precision

If comprehensive stress testing is performed to observe server performance under heavy load, then performance data can be obtained, but service disruption occurs

Engineering Contradiction:
Improveperformance observation accuracyVSAvoidservice continuity
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent performs preliminary characterization of the web server's normal performance baseline before executing the flash crowd stress test. This preliminary action establishes reference metrics for response time, throughput, and resource utilization, enabling accurate measurement of performance degradation while maintaining service continuity during the test.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The flash crowd stress testing is conducted as a periodic, time-bound event rather than continuous stress. The test activates for a predetermined duration to collect performance data under heavy load, then terminates to allow the server to return to normal operation. This periodic approach enables accurate performance observation while limiting service disruption to specific time windows.

Inventive Principle:
Principle #19Periodic action

3Reliability

If server resources are over-provisioned to handle unexpected traffic surges, then service robustness is improved, but infrastructure cost increases

Engineering Contradiction:
Improveservice robustnessVSAvoidinfrastructure resources
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent implements feedback mechanisms that monitor web server performance metrics during flash crowd events and provide data to providers about actual resource utilization patterns. This feedback enables providers to make informed decisions about resource provisioning based on empirical evidence of how the server handles traffic surges, rather than relying on conservative over-provisioning estimates.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent enables dynamic adjustment of server resource parameters based on observed performance characteristics from flash crowd testing. Providers can modify provisioning parameters such as bandwidth allocation, processing capacity, and memory resources based on measured performance thresholds and traffic surge patterns, optimizing the balance between robustness and resource efficiency.

Inventive Principle:
Principle #35Parameter changes

4Speed

If access bandwidth is increased to improve server performance under overload, then network capacity is improved, but constraints on other server-side resources become more significant

Engineering Contradiction:
Improvenetwork access bandwidthVSAvoidresource constraint analysis
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent segments the web server system into distinct sub-systems (network access, storage, data processing, CPU) and independently characterizes the resource constraints of each segment during flash crowd events. By analyzing performance bottlenecks at the sub-system level, the patent identifies which specific resources become constraining under different bandwidth conditions, enabling targeted resource optimization rather than blanket increases across all components.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentUS7933745B2System and method for profiling resource constraints of web servers
Publication Date: 2011.04.26 AT&T INTELLECTUAL PROPERTY I L P
  • US7933745B2 patent drawing
  • US7933745B2 patent drawing
  • US7933745B2 patent drawing

AI summary

Disclosed is a method and system for determining one or more performance characteristics of a target server. A command is transmitted from a coordinator to a plurality of clients. The command instructs the plurality of clients to each transmit a request targeting a sub-system of said target server. A response time is then received from each client and a performance characteristic is determined from the received response times.