Adaptive Workload Routing via Search Index Queries

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing workload placement and service request routing systems lack efficiency in routing requests while maintaining confidentiality of current system state information and adapting to dynamic system conditions.

Innovation Solution

A method that determines a search index representing the current system state and context, associates service instances with search queries, runs these queries against the index to produce scored results, and routes requests to the service instance with the highest score, leveraging existing search indexes like the ELK stack to ensure adaptivity and confidentiality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If existing workload placement systems expose system state information for routing decisions, then routing efficiency is improved, but system state confidentiality is compromised

Engineering Contradiction:
Improverouting efficiencyVSAvoidsystem state confidentiality
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent introduces search queries as an intermediary mechanism that indirectly accesses system state information through search indexes without exposing the actual system state data. The search queries act as a mediator between the routing system and the system state, allowing efficient routing decisions while preserving confidentiality by never directly exposing the underlying system state information.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If workload placement systems continuously monitor system state for adaptive routing, then adaptability to dynamic conditions is improved, but system complexity increases

Engineering Contradiction:
Improveadaptability to dynamic system conditionsVSAvoidsystem complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent pre-processes system state information into search indexes before routing decisions are needed. This preliminary action transforms raw system state data into a structured, queryable format that enables adaptive routing without requiring complex real-time monitoring and processing during the routing decision itself, thereby reducing system complexity while maintaining adaptability.

Inventive Principle:
Principle #10Preliminary action

3Ease of manufacture

If traditional routing methods are used without search technology, then implementation simplicity is maintained, but routing precision and adaptivity deteriorate

Engineering Contradiction:
Improveimplementation simplicityVSAvoidrouting precision
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent creates search indexes as copies or representations of the system state, which can be queried without modifying the actual system state or requiring complex routing logic. This copying approach maintains implementation simplicity by using familiar search technology while dramatically improving routing precision through accurate, up-to-date system state information.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS20240330308A1System state adaptive workload placement and request routing using search technology
Publication Date: 2024.10.03 INTERNATIONAL BUSINESS MACHINE CORPORATION
  • US20240330308A1 patent drawing
  • US20240330308A1 patent drawing
  • US20240330308A1 patent drawing

AI summary

An adaptive workload placement and service request routing while ensuring that information relating to a current system state remains confidential. A search index representing a current system state and context is built and a set of service instances is received. The set of service instances are then associated with a set of search queries that are run against the search index, with the running of the set of search queries against the search index producing a set of scored search results. Each scored search result is associated with a service instance. The service instance associated with the highest scored search result is selected, and a service request is routed to the selected service instance.