Dynamic Bandwidth Allocation for Data Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current bandwidth provisioning for data mirroring and replication in wide area networks often results in over-provisioning, leading to unnecessary costs and performance issues due to infrequent peak demand, as it is based on projected peak data demand rather than real-time requirements.
Innovation Solution
A method that dynamically adjusts bandwidth allocation by monitoring parameters such as average compression ratio, throughput, and round trip time, triggering increases or decreases based on threshold violations to match current and expected demand, thereby optimizing bandwidth usage and avoiding over-provisioning and under-provisioning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If bandwidth is provisioned to meet projected peak data demand, then application performance is ensured during peak times, but bandwidth cost increases due to over-provisioning during normal operation
Solution Approach 1:
The patent implements dynamic bandwidth adjustment by continuously monitoring actual data replication parameters (compression ratio, throughput, round trip time) and automatically modifying bandwidth allocation based on real-time conditions. This transitions from static projected-demand-based provisioning to dynamic actual-demand-based provisioning, resolving the contradiction between ensuring performance and reducing cost.
Solution Approach 2:
The system establishes a feedback loop where performance parameters (compression ratio, throughput, round trip time) are continuously measured and fed back to the bandwidth management system. This feedback mechanism enables the system to adjust bandwidth allocation based on actual replication conditions, preventing both over-provisioning during normal operation and under-provisioning during peak demand.
2Reliability
If bandwidth is provisioned based on projected peak demand, then future peak demand is satisfied, but actual peak demand may exceed projected demand causing application slowdown
Solution Approach 1:
The system performs preliminary monitoring of replication parameters and proactively adjusts bandwidth before actual peak demand occurs. By analyzing trends in compression ratio, throughput, and round trip time, the system can anticipate peak conditions and increase bandwidth allocation in advance, ensuring that actual peak demand is met without causing application slowdown.
Solution Approach 2:
The continuous feedback mechanism monitors actual replication performance and triggers bandwidth adjustments when parameters indicate approaching peak conditions. This real-time feedback enables the system to respond dynamically to actual demand patterns, ensuring both future peak demand satisfaction and maintaining application throughput.
3Device complexity
If static bandwidth allocation is used, then bandwidth management is simple, but bandwidth cannot adapt to changing data replication patterns
Solution Approach 1:
The system implements self-service bandwidth management where the replication process automatically monitors its own performance parameters (compression ratio, throughput, round trip time) and triggers self-adjustments to bandwidth allocation. This eliminates the need for complex external bandwidth management systems while achieving adaptive bandwidth allocation that responds to changing data replication patterns.
Solution Approach 2:
The feedback mechanism enables the system to automatically adapt bandwidth allocation based on real-time replication conditions. By continuously monitoring performance parameters and adjusting bandwidth accordingly, the system achieves high adaptability without requiring complex manual intervention or external management systems.
Data Source
AI summary
An approach is provided for managing an allocation of a bandwidth of a dedicated channel in a network being utilized by an application performing a replication of data from a first to a second storage resource. Checks of parameters specified by the application are determined. The parameters include, but are not limited to, average compression ratio, compression ratio trend, throughput, throughput trend, and round trip time. Respective high and low threshold values for the parameters are determined. At least one of the checks of the parameters is executed, which determines a violation of a threshold value for one of the parameters. If the violation of the high threshold value is determined, an increase in the bandwidth is triggered. If the violation of the low threshold value is determined, a decrease in the bandwidth is triggered.