Policy Decision Offload Accelerator for Cloud Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud computing environments, policy decision points (PDPs) can cause bottlenecks and service disruptions if they fail to respond to access requests, leading to inefficient scaling and potential denial of service due to the concentration of decision-making workload.
Innovation Solution
A policy decision offload accelerator is introduced, utilizing machine-learning techniques to predict responses to access requests and offload decisions to a processor-executable veto rules module, allowing for faster access management even when the PDP is unavailable, by using a request-response association model trained by a learning engine.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If a centralized policy decision point (PDP) is used to manage access requests, then policy decision consistency is improved, but system reliability deteriorates due to single point of failure and bottleneck effects
Solution Approach 1:
The patent segments the centralized PDP into multiple distributed PDP instances across different computing nodes. Each PDP instance independently processes access requests for its associated storage system, eliminating the single point of failure while maintaining policy decision consistency through distributed architecture. This segmentation allows the system to continue operating even if one PDP instance fails.
Solution Approach 2:
The patent introduces a policy decision offload accelerator as an intermediary component between storage systems and PDPs. This accelerator uses machine learning models to predict access patterns and pre-make policy decisions, reducing the workload on PDPs and enabling faster response times. The intermediary handles routine decisions locally while PDPs focus on complex policy evaluation.
2Measurement precision
If all policy decisions are processed by the PDP, then decision-making accuracy is improved, but processing speed deteriorates due to centralized workload concentration
Solution Approach 1:
The patent implements preliminary action by using machine learning models to predict access requests before they occur. The system analyzes historical access patterns and pre-determines policy decisions for anticipated requests. This allows the system to prepare responses in advance and reduces the processing time when actual access requests arrive, as the ML model can quickly provide predicted outcomes.
Solution Approach 2:
The policy decision offload accelerator serves as an intermediary that handles routine and predictable policy decisions locally using machine learning predictions. This offloads computational workload from the centralized PDP, allowing PDPs to focus on complex policy evaluation while the accelerator handles high-volume routine requests, thereby improving overall processing speed without sacrificing decision accuracy.
3Reliability
If the PDP responds to every access request, then policy compliance is improved, but system scalability deteriorates due to linear increase in processing load
Solution Approach 1:
The patent introduces policy decision offload accelerators as intermediary components that handle routine access requests independently. These accelerators use machine learning models to make policy decisions without requiring PDP intervention, thereby reducing the processing load on PDPs. This allows the system to scale horizontally by adding more accelerators rather than increasing PDP capacity linearly.
Solution Approach 2:
The policy decision offload accelerator enables self-service by allowing the system to autonomously make policy decisions for routine access requests using machine learning predictions. The accelerator independently evaluates access patterns, predicts outcomes, and enforces policies without requiring continuous PDP involvement. This self-service capability maintains policy compliance while significantly improving system scalability.
Data Source
AI summary
Methods and systems for a networked computing system are provided. One method includes detecting that a processor executable, policy decision point (PDP) has not responded to a request for accessing data associated with a storage system; predicting a response to the request using a machine-learned, request-response association maintained by a processor executable training device; and presenting the predicted response to a processor executable, policy enforcement point (PEP) for granting access to the data and denying access to the data, based on the predicted response.