Unified Workload Manager for Cross-Platform Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In workload management environments, existing load balancing techniques struggle to effectively distribute requests across sites with combined application and data tiers residing on different platforms, leading to inefficiencies in resource utilization and potential conflicts due to the lack of unified monitoring and decision-making across disparate systems.
Innovation Solution
A workload manager groups application and data tiers into a single workload, using monitoring agents to collect health, availability, and capacity metrics from both tiers to create a distribution policy that directs workload connections to the most suitable site, enabling load balancing across different operating system environments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If application and data tiers are kept on separate platforms, then customers can preserve investment in mainframe data tiers while migrating application fronts to different environments, but load balancing cannot effectively distribute requests across sites with combined tiers on different platforms
Solution Approach 1:
The patent combines application tier and data tier into a unified workload group that spans multiple platforms. The workload manager aggregates metrics from both tiers (CPU utilization, memory usage, I/O activity from application tier; storage capacity, access patterns from data tier) and makes load balancing decisions based on the combined state, enabling effective load distribution across sites even when tiers reside on different platforms.
Solution Approach 2:
The workload manager acts as an intermediary between the load balancer and the heterogeneous platform environment. It translates diverse platform-specific metrics from application and data tiers into a unified view that the load balancer can use to make informed routing decisions, bridging the gap between platform diversity and load balancing requirements.
2Measurement precision
If unified monitoring is implemented across disparate platforms, then load balancing decisions can be informed by combined tier metrics, but system complexity increases due to need to monitor and integrate multiple platform types
Solution Approach 1:
The workload manager implements a universal monitoring framework that can collect and process metrics from multiple platform types (mainframe, distributed systems, cloud platforms) through a common interface. It provides multi-functional capabilities including metric collection, normalization, aggregation, and analysis across heterogeneous environments, reducing the complexity that would otherwise arise from platform-specific monitoring implementations.
Solution Approach 2:
The system transforms diverse platform-specific parameters into a standardized set of workload metrics that can be uniformly processed. It changes the representation of system state from platform-specific formats to a common metric space, enabling accurate workload assessment without requiring complex platform-specific handling for each metric type.
3Productivity
If load balancing decisions consider only application tier metrics, then decision-making is simpler, but resource utilization is inefficient because data tier capacity and health are not accounted for
Solution Approach 1:
The workload manager performs preliminary aggregation and analysis of both application tier and data tier metrics before load balancing decisions are made. It proactively computes a unified workload health score that incorporates both tiers' current state, capacity, and trends, so that when the load balancer needs to make a routing decision, the assessment is already complete and ready for immediate use.
Data Source
AI summary
A method, apparatus and computer program product for improved load balancing provides for the grouping under a same workload of both application instances in an application tier, and data sharing members in a data tier. This grouping enables a workload manager to make recommendations (to load balancer appliances) about how to distribute workload connections, e.g., based on metrics gathered from both the application and data tiers. In this approach, both applications and data sources are grouped into a workload grouping, and health, status and capacity information about both of these tiers (application and data) is then used to determine an overall distribution policy for the workload. These different tiers can reside on the same or different operating system environments.