Dynamic Network Workload Control via Server Status Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale network systems with multiple connections face challenges in efficiently managing workload dynamics due to varying request traffic rates across connections, leading to queuing or rejection of requests, as existing solutions are not practical for configuring servers to match the maximum capacity of all clients.
Innovation Solution
A computer-implemented method using a server connection manager to establish and manage concurrent network connections, where a server status handler component detects and handles status information requests, gathering and adding this information to outgoing messages, allowing clients to dynamically control workload by requesting status information only when needed, thus reducing message payload and improving response times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the server is configured to match the maximum capacity of all clients, then the server can handle peak load from any single client, but this leads to large amounts of redundancy and inefficient resource utilization
Solution Approach 1:
The patent implements dynamic workload control where the server connection manager continuously monitors the number of active connections and dynamically adjusts whether to accept new connections based on current server load. This allows the server to adapt its capacity utilization in real-time rather than being statically configured for peak load, eliminating redundancy while maintaining reliability.
Solution Approach 2:
The server connection manager implements a feedback mechanism by monitoring the current number of connections and using this information to make decisions about accepting new connections. This feedback loop allows the server to optimize its resource utilization by matching capacity to actual demand rather than predetermined maximum capacity.
2Productivity
If each connection is configured to support more than its fair-share of the server's overall capacity, then busy clients can route higher rate of requests during times when the server is less busy, but this leads to prolonged periods of queuing or request rejection
Solution Approach 1:
The system uses feedback from the current connection count to dynamically control whether new connections are accepted. When the server is less busy (fewer active connections), it accepts new connections, allowing busy clients to route requests. When the server approaches capacity, it rejects new connections, preventing queuing and rejection delays.
Solution Approach 2:
The connection acceptance policy is dynamic rather than static, adjusting in real-time based on server load conditions. This allows the system to optimize request processing rates by accepting connections when capacity is available and rejecting them when the server is busy, eliminating prolonged queuing periods.
3Adaptability or versatility
If additional software like z/OS WLM or Tivoli Netview is used to discover connection capacity utilization, then clients can monitor server load, but this requires separate configuration and additional software infrastructure
Solution Approach 1:
The patent merges the workload management functionality directly into the connection manager, which is already part of the server infrastructure. This eliminates the need for separate additional software like z/OS WLM or Tivoli Netview, as the connection manager itself provides the capacity discovery and control capabilities.
Solution Approach 2:
The server connection manager performs self-monitoring of its own capacity and makes autonomous decisions about connection acceptance. This self-service approach eliminates the need for external monitoring software, as the system manages its own workload dynamically without requiring separate configuration or additional infrastructure.
Data Source
AI summary
In a network capable of handling a number of concurrent network connections between network nodes, referred to as client and server, the client and server connection managers are customized by adding respective status handler components which add server status information requests, from client to server, and responsive status information, from server to client into message headers of messages being sent between the server and client. The supported server status information types for any given connection are defined when a connection is established through dialogue between the client and server, and then persist for the lifetime of the connection. The customizations of the server and client connection managers are modest and the increase in network traffic over the connection is proportionally very small. Moreover, status information requests can be processed and responded to quickly by attaching to messages that are being sent between the nodes.