Distributed NAT Architecture with Virtual Mapping Servers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing Network Address Translation (NAT) systems face scalability and availability challenges due to IPv4 address exhaustion and the limitations of single hardware-based solutions, requiring more efficient methods to handle increased traffic and failures in cloud networking and data-center architectures.
Innovation Solution
A distributed NAT architecture utilizing both hardware and software capabilities, with fast-path and slow-path servers, and Virtual Mapping Servers (VMS) for load-balancing and failover, allowing for stateful network address translations and dynamic address/port assignments to ensure high availability and minimal latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If hardware-based NAT solutions are used, then high throughput and high availability are achieved, but scalability is limited by single ASIC capabilities
Solution Approach 1:
The patent divides the NAT system into multiple software-based NAT instances running in parallel on commodity servers, replacing the single hardware-based NAT appliance. Each instance handles a portion of the traffic, enabling linear scalability by adding more instances while maintaining high throughput through distributed processing.
2Adaptability or versatility
If software-based NAT solutions are used, then scalability and parallel processing are improved, but failure handling capability deteriorates
Solution Approach 1:
The patent implements pre-configured backup NAT instances that stand by in a hot-standby mode. When a primary NAT instance fails, the backup instance immediately takes over without service interruption. This preliminary preparation of failover mechanisms ensures that software-based solutions can handle failures as effectively as hardware solutions.
3Productivity
If multiple NAT appliances are load-balanced, then throughput is increased, but system complexity increases due to additional load-balancing routers
Solution Approach 1:
The patent combines the NAT functionality with load-balancing capabilities into a unified software-based NAT instance. Each NAT instance independently performs both NAT translation and load-balancing decisions, eliminating the need for separate load-balancing hardware or additional routing complexity while maintaining increased throughput through parallel processing.
Data Source
AI summary
This disclosure describes techniques for providing a distributed scalable architecture for Network Address Translation (NAT) systems with high availability and mitigations for flow breakage during failover events. The NAT servers may include functionality to serve as fast-path servers and/or slow-path servers. A fast-path server may include a NAT worker that includes a cache of NAT mappings to perform stateful network address translation and to forward packets with minimal latency. A slow-path server may include a mapping server that creates new NAT mappings, depreciates old ones, and answers NAT worker state requests. The NAT system may use virtual mapping servers (VMSs) running on primary physical servers with state duplicated VMSs on different physical failover servers. Additionally, the NAT servers may implement failover solutions for dynamically allocated routeable address/port pairs assigned to new sessions by assigning new outbound address/port pairs when a session starts and broadcasting pairing information.