Microservice Session State Modeling via Bayesian Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Microservice architectures in cloud computing pose challenges in modeling session states and predicting defects, leading to inefficiencies in resource provisioning and security threats due to complex interactions and distributed deployment risks.
Innovation Solution
Implementing an additive Weibull distribution for modeling session states and using a Bayesian statistical framework to predict defects, along with monitoring network packets to detect security threats, enabling intelligent resource scaling and real-time threat detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If microservices are deployed in cloud infrastructure to enable service-oriented architecture, then system flexibility and scalability are improved, but security risks and session state modeling complexity increase
Solution Approach 1:
The patent segments the complex session state modeling problem by analyzing individual microservices independently. Each microservice is assigned its own session state model that captures only the relevant states for that specific service, rather than attempting to model the entire distributed system as a single complex unit. This segmentation reduces the complexity of individual models while maintaining overall system flexibility.
Solution Approach 2:
The patent introduces an intermediary session state modeling layer between the microservices and the resource provisioning system. This intermediary layer abstracts the complexity of microservice interactions by translating diverse session states into a unified representation that can be processed by the provisioning system, thereby reducing the complexity burden on individual services while maintaining system adaptability.
2Ease of operation
If traditional monolithic architecture is used, then session state modeling is straightforward, but system scalability and resource efficiency are limited
Solution Approach 1:
The patent applies dynamics by transitioning from static monolithic session modeling to dynamic microservice-specific session state modeling. Each microservice maintains its own session state model that can be independently updated and scaled, allowing the system to adapt resource allocation dynamically based on actual microservice usage patterns rather than relying on fixed monolithic structures.
Solution Approach 2:
The patent changes the parameters of session state modeling from a single unified model to multiple specialized models, each with its own parameters tailored to specific microservices. This parameter transformation enables more precise resource provisioning decisions by considering the specific characteristics of each microservice rather than applying a one-size-fits-all approach.
3Adaptability or versatility
If microservices are deployed to enable distributed architecture, then system capability is enhanced, but security threats from malicious users increase
Solution Approach 1:
The patent applies preliminary action by implementing session state modeling and resource provisioning decisions before actual attacks occur. By continuously monitoring session states and predicting resource needs in advance, the system can proactively allocate resources and adjust configurations to prevent malicious users from exploiting system vulnerabilities, rather than reacting to attacks after they manifest.
Solution Approach 2:
The patent introduces feedback mechanisms that continuously monitor microservice performance and session states, using this information to adjust resource allocation and security policies in real-time. This feedback loop enables the system to detect anomalous patterns that may indicate malicious activity and respond by adjusting resource provisioning or triggering security protocols.
Data Source
AI summary
Methods, systems, and computer-readable storage media for provisioning resources for microservice-based applications hosted in cloud infrastructures, implementations including actions of receiving event data from each of a plurality of microservices hosted in a cloud infrastructure, for each microservice, determining a load metric based on a probability distribution, comparing each load metric to one or more thresholds to effect respective comparisons, and selectively initiating provisioning of one or more instances of at least one microservice based on a respective comparison.