Software Load Balancer Hitless Network Element Upgrade
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing method for hitless upgrades of core network elements is inadequate as it cannot simulate diverse network configurations, leading to potential overall system issues during the upgrade process, resulting in risks to service continuity.
Innovation Solution
A network element upgrade method involving a software load balancer (SLB) that distributes service messages between a first and a second network element, with the second element having the same configuration and version as the first, to ensure seamless switching and minimize service interruptions, using a distribution policy to control message forwarding based on user identifiers and 5-tuple information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all services are switched to the new version network element, then upgrade completion is achieved, but system reliability deteriorates due to undetected problems
Solution Approach 1:
The patent applies partial action by gradually switching services to the new version network element rather than all at once. The SLB distributes service messages proportionally, initially sending fewer messages to the new version and increasing the proportion as stability is confirmed. This allows the upgrade to proceed in controlled stages, maintaining reliability while ultimately achieving full upgrade completion.
Solution Approach 2:
The patent implements preliminary action by deploying the new version network element in parallel with the existing one before full switching. The new version is pre-configured and tested while receiving a portion of service messages, allowing problems to be detected and resolved before complete dependency establishment, thus ensuring service continuity during the transition.
2Reliability
If gradual service switching is implemented, then system reliability is maintained, but upgrade time increases
Solution Approach 1:
The patent maintains continuity of useful action by ensuring that service message distribution continues uninterrupted during the upgrade process. The SLB dynamically adjusts the distribution proportion between old and new version network elements without stopping service flow, allowing the upgrade to proceed continuously while monitoring stability, thus balancing reliability maintenance with time efficiency.
3Reliability
If service messages are distributed to multiple network elements, then risk is reduced, but device complexity increases
Solution Approach 1:
The patent introduces an SLB as an intermediary component that manages the distribution of service messages between the old and new version network elements. This centralized mediator simplifies the overall system architecture by providing a single point of control for traffic distribution, rather than requiring complex peer-to-peer coordination between multiple network elements, thus reducing complexity while maintaining reliability.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6A
AI summary
The present invention discloses a network element upgrade method and a device, and relates to the field of communications technologies, so as to resolve a problem that in an existing upgrade process, after a network element switches all services to a new version, an overall situation may be affected once a problem occurs during the upgrade to the new version. The network element upgrade method provided in the present invention includes: receiving, by an SLB, a distribution policy delivered by an element management system EMS; forwarding, to a first network element according to the distribution policy, some of a plurality of received service messages sent by an external network element, and forwarding remaining service messages to a second network element according to the distribution policy; and after a processing result of the service messages on the second network element meets a preset condition, forwarding, by the SLB to the second network element according to the distribution policy, all of a plurality of subsequently received service messages sent by the external network element.