Secure Metadata Sharing Among VNFs via RDMA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In cloud native microservice infrastructure, there is no mechanism for secure metadata sharing across different hosts and tenants in a network, limiting the ability of virtualized network functions (VNFs) to improve performance by sharing metadata.
Innovation Solution
A distributed forwarding and metadata sharing (DFMS) framework using remote direct memory access (RDMA) enables secure metadata sharing among VNFs located on different hosts within a cluster, with a DFMS controller registering VNFs into groups and assigning access rights, allowing RDMA requests for metadata access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If VNFs are distributed across different hosts and tenants in cloud native infrastructure, then system scalability and security are improved, but metadata sharing capability deteriorates due to lack of secure sharing mechanism
Solution Approach 1:
The patent introduces a metadata server as an intermediary component that mediates metadata sharing between VNFs on different hosts. The metadata server stores metadata centrally and provides controlled access to authorized VNFs through standardized interfaces, enabling secure metadata sharing across distributed multi-tenant environments without requiring direct peer-to-peer access between VNFs.
2Productivity
If traditional metadata sharing approaches are used without RDMA, then system complexity is reduced, but metadata access speed and efficiency deteriorate
Solution Approach 1:
The patent replaces traditional network-based metadata access mechanisms with Remote Direct Memory Access (RDMA) technology. RDMA enables direct memory-to-memory data transfer between hosts bypassing the operating system and network stack, achieving high-speed metadata access with low latency. The system manages this complexity through standardized RDMA interfaces and a metadata server that abstracts the underlying hardware complexity from VNFs.
Data Source
AI summary
A system and method for securely sharing metadata among virtualized network functions (VNFs) disposed within a cluster of host computers includes a distributed forwarding and metadata sharing (DFMS) framework that forwards received data packets and provides remote direct memory access (RDMA) enabled sharing of secured metadata among the VNFs from more than one host computer in the cluster of host computers, and a DFMS controller that resides in the cluster of host computers as a VNF and registers VNFs in the cluster to at least one group of VNFs in the cluster and assigns access rights to the metadata of VNFs within respective groups of VNFs in the cluster. In operation, each registered VNF in a group in the cluster uses RDMA requests to access the metadata for other registered VNFs in the group in the cluster to which the each registered VNF in the group has access rights.