Stateless Transformation Replication for Message Flow Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems face challenges in optimizing the placement of computational components of event stream processing applications on a network of servers to minimize end-to-end latency, as existing methods do not effectively exploit replication and workload distribution across brokers.
Innovation Solution
The system replicates stateless transformations onto multiple brokers, assigns varying workloads, and determines the number and location of replicas using a placement algorithm that considers message properties and broker capacities to minimize overall end-to-end latency, enhancing placement flexibility and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If replication is applied to stateless transformations, then placement flexibility and performance are improved, but system complexity increases
Solution Approach 1:
The system segments stateless transformations into multiple independent replicas that can be distributed across different brokers. Each replica is an identical copy that can process messages independently, allowing flexible placement decisions while maintaining functional equivalence. This segmentation enables the placement algorithm to consider multiple brokers for the same transformation, improving placement flexibility.
Solution Approach 2:
The patent creates virtual copies of stateless transformations as replicas. These copies are identical to the original transformation but can be placed on different brokers. The copying approach allows the system to explore multiple placement configurations without creating new transformation logic, thereby improving flexibility while controlling complexity through replication rather than variation.
2Loss of time
If multiple replicas are created for stateless transformations, then end-to-end latency is reduced, but computational overhead increases
Solution Approach 1:
The system creates multiple replicas of stateless transformations, which is excessive action compared to having a single instance. This excessive replication allows the placement algorithm to distribute workloads more effectively and reduce end-to-end latency by providing more parallel processing paths. The marginal increase in computational overhead is justified by the significant reduction in latency.
Solution Approach 2:
The patent changes the parameter of transformation instances from single to multiple replicas. This parameter change enables the system to leverage parallel processing and reduce latency. The placement algorithm optimizes the number and distribution of replicas based on broker capacities and message flow characteristics, balancing the trade-off between latency reduction and computational overhead.
3Adaptability or versatility
If stateless transformations are labeled as replicable, then workload distribution options increase, but placement algorithm complexity increases
Solution Approach 1:
The system performs preliminary action by pre-labeling transformations as replicable or non-replicable before the placement process. This classification is done based on transformation characteristics (stateless vs. stateful) and can be configured by administrators. By doing this preliminary classification, the placement algorithm doesn't need to evaluate all transformations for replication eligibility during placement, reducing its complexity while still providing flexible workload distribution options.
Solution Approach 2:
The patent applies local quality by treating different transformations differently based on their characteristics. Stateless transformations are labeled as replicable and given the option to be replicated, while stateful transformations are labeled as non-replicable. This localized differentiation allows the placement algorithm to apply specific rules only where needed, reducing overall algorithm complexity while maintaining flexible workload distribution for applicable transformations.
Data Source
AI summary
Systems, methods and computer program products for improving overall end-to-end runtime latency of flow graphs of message transformations which are placed onto an overlay network of broker machines by aggressively replicating stateless transformations. Exemplary embodiments include a system having a process including defining a message transformation graph, receiving information about measured and estimated properties of a message flow, receiving information about physical brokers and links in the overlay network onto which the message transformation graph is deployed, labeling each of a plurality of stateless transformations associated with the flow graph as replicable, heuristically determining a number of replicas and a corresponding load partitioning ratios among the number of replicas for each stateless transformation, converting the message transformation graph into an enhanced flow graph, running a placement algorithm with the enhanced flow graph and consolidating each of the plurality of virtual replicas that are assigned to a common message broker.