Runtime VNF Orchestration Across VMs, Containers, and Serverless
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing cloud platforms enforce platform-specific virtualization technologies and software designs, leading to resource inefficiencies, limited scalability, and management challenges for Virtual Network Functions (VNFs), as they do not optimize for runtime conditions and inter-service communication.
Innovation Solution
A method and system for orchestrating VNF deployment across multiple virtualization technologies (VMs, containers, and serverless) in a unified cloud platform, allowing dynamic selection of virtualization and software design at runtime based on resource utilization and optimization objectives, decoupling VNF specification from platform-specific constraints.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single virtualization technology is used per cloud platform, then platform-specific optimization is achieved, but versatility and adaptability across different virtualization technologies are reduced
Solution Approach 1:
The orchestrator is designed to support multiple virtualization technologies (VMs, containers, serverless) within a single cloud platform, enabling a universal deployment system that can adapt to different virtualization requirements without requiring separate platform-specific orchestration systems
Solution Approach 2:
The orchestrator acts as an intermediary layer between the cloud platform and various virtualization technologies, abstracting the complexity of multi-technology management and providing a unified interface for deployment, monitoring, and resource allocation across different virtualization environments
2Reliability
If VNFs are deployed using traditional monolith applications on VMs, then resource isolation and kernel features are provided, but resource footprint and deployment flexibility are increased
Solution Approach 1:
The system dynamically selects the appropriate virtualization technology and application implementation based on runtime conditions, optimization objectives, and resource availability, allowing VNFs to be deployed as VMs, containers, or serverless functions depending on the specific requirements and current infrastructure state
Solution Approach 2:
The orchestrator changes the deployment parameters (virtualization technology type, application architecture style) based on optimization objectives and resource utilization metrics, enabling the same VNF to be deployed with different configurations to balance isolation requirements against resource efficiency
3Productivity
If micro-services architecture is used for VNFs, then deployment granularity and CI/CD support are improved, but management overhead and inter-service communication complexity increase
Solution Approach 1:
The orchestrator segments the management of micro-services into manageable units, coordinating individual service deployments, updates, and scaling operations independently while maintaining overall system coherence, thereby enabling fine-grained control without overwhelming management complexity
Solution Approach 2:
The system implements feedback mechanisms that monitor the state of deployed micro-services and automatically adjust orchestration decisions based on runtime performance, resource utilization, and service dependencies, reducing management overhead through intelligent automation
4Productivity
If platform-specific service design is enforced, then optimal performance on that platform is achieved, but portability and migration capability across platforms are reduced
Solution Approach 1:
The orchestrator provides a universal deployment framework that supports multiple virtualization technologies and application architectures, enabling VNFs to be deployed optimally on different cloud platforms (OpenStack, Kubernetes, serverless) without requiring platform-specific redesign, thereby achieving both performance optimization and portability
Data Source
AI summary
A virtual network function application is to be deployed in a multi-technology virtualization environment. The multi-technology virtualization environment comprises a plurality of virtualization technologies. The application exists in a plurality of application implementations. At least one of the virtualization technologies has a preferred one of the application implementations. The deployment comprises receiving a trigger for a deployment of the virtual network function application. In response to the trigger, information is obtained about at least one optimization objective and about resource utilization in the multi-technology virtualization environment. Based on the received information, one of the plurality of virtualization technologies is selected for said deployment. One of the application implementations is also selected, and the selected application implementation need not be the preferred application implementation for the selected virtualization technology. The selected application implementation of the virtual network function application is then deployed on the selected one of the virtualization technologies.