Multi-Tier Application Mapping via Passive Dependency Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional application dependency mapping methods are labor-intensive, prone to errors, and create inaccurate maps due to frequent network changes, requiring manual intervention and network scanning that burdens the network and collects incomplete data.
Innovation Solution
A method for generating application maps automatically across a large network using linear communication orbits, iteratively gathering information from distributed machines without manual intervention, and utilizing a server system to aggregate and present the maps in near real-time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual application dependency mapping is performed by application owners and architects, then mapping accuracy can be maintained through expert knowledge, but the process becomes labor-intensive and outdated quickly due to frequent network changes
Solution Approach 1:
The system enables automated application mapping where the network infrastructure itself provides the mapping data through passive monitoring of traffic flows. Application owners and architects no longer need to manually create and update maps; instead, the system self-generates accurate dependency maps by observing actual network communications, resolving the contradiction between accuracy and update speed.
Solution Approach 2:
The patent replaces the manual mechanical process of creating application maps with an automated electronic monitoring system. Network traffic data is automatically collected and processed by software agents that generate dependency maps without human intervention, substituting expert manual labor with automated computational analysis of network flows.
2Loss of information
If network scanning is used to collect application dependency data, then comprehensive mapping information can be obtained, but the network performance is degraded and the collected data may not be representative of usual traffic
Solution Approach 1:
Instead of continuous or frequent scanning that degrades network performance, the system uses periodic passive monitoring of network traffic flows. Software agents continuously observe and record traffic patterns without actively probing the network, collecting comprehensive mapping information during normal network operation without causing performance degradation.
Solution Approach 2:
The patent introduces software agents as intermediaries that passively monitor network traffic between applications and infrastructure components. These agents collect dependency information by observing actual traffic flows rather than active scanning, ensuring data representativeness while minimizing impact on network performance since they operate in the background without generating additional traffic.
3Measurement precision
If a library of login credentials is maintained to access connection data of each node, then detailed application mapping can be achieved, but security risks increase and maintenance complexity grows
Solution Approach 1:
The patent extracts authentication requirements from the mapping process entirely. Instead of maintaining credentials to access node information, the system uses passive traffic flow monitoring that does not require logging into or accessing individual nodes. Connection data is collected from network traffic observations rather than through authenticated access to each component, eliminating credential management complexity while maintaining mapping accuracy.
4Object-affected harmful factors
If network scans are scheduled during low usage periods, then network performance is preserved, but the collected mapping data becomes unrepresentative of actual traffic patterns
Solution Approach 1:
The system implements continuous passive monitoring of network traffic flows rather than periodic scanning during specific time windows. Software agents continuously observe and record traffic patterns throughout all periods of network operation, ensuring that mapping data accurately reflects actual traffic patterns at all times while maintaining constant network performance since no active scanning occurs.
Data Source
AI summary
An application mapping procedure obtains and aggregates application mapping information from a plurality of machines in a distributed system. An application dependency map, including first layer of application mapping information, is initialized, and then a first query is sent to one or more of the machines. In response, information identifying entities that have participated in predefined communications with entities identified in an existing layer of application mapping information in the application dependency map are received, and a second layer of application mapping information is added to the application dependency map, based at least in part on the information received in response to the first query. After adding the second layer of application mapping information to the application dependency map, a second query is sent to one or more of the of the endpoint machines, the second query being based at least in part on the application dependency map.