Legacy App Deployment on Hyperscalers via Service Mesh Translation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Legacy applications face challenges when deploying on new generation hyperscalers due to incompatibility with static physical host names and static port addresses, as well as limitations in deploying customized code, which hinders seamless operation across different cloud platforms.
Innovation Solution
A method that converts topology descriptions with static physical host names and port addresses into deployment files compatible with cloud platforms, using a control script and key-lock process to limit pre-start database operations to a single node, enabling deployment on hyperscalers and orchestrators that do not support static naming and addresses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If legacy applications use static physical host names and static port addresses, then they can maintain stable connections and configurations, but they become incompatible with new generation hyperscalers that do not support static naming and addresses
Solution Approach 1:
The patent introduces a service mesh as an intermediary layer between legacy applications and the hyperscaler infrastructure. The service mesh translates static host names and port addresses into dynamic service endpoints, allowing legacy applications to maintain their static naming conventions while communicating with dynamically provisioned containers on the hyperscaler platform. This mediator resolves the incompatibility without requiring changes to the legacy application code.
Solution Approach 2:
The patent transforms the networking parameters from static physical host names and port addresses to dynamic service names and mesh-provided endpoints. By changing the addressing paradigm from physical to logical/service-based, the system maintains connection stability through service identity while adapting to the dynamic nature of hyperscaler container provisioning.
2Productivity
If customized code is deployed to multiple nodes for parallel processing, then productivity increases, but pre-start database operations may execute concurrently causing conflicts and data inconsistency
Solution Approach 1:
The patent implements a control script that executes pre-start database operations as a preliminary action on a designated leader node before other nodes join the cluster. This ensures that database initialization, schema creation, and other critical setup operations are performed exactly once in the correct sequence, preventing conflicts that would arise from concurrent execution while still allowing parallel processing of application workloads across multiple nodes.
Solution Approach 2:
The patent assigns different roles and execution permissions to different nodes. The leader node has the specific quality of executing pre-start database operations, while follower nodes have the quality of joining the cluster and processing workloads. This local differentiation of node capabilities ensures data consistency while maintaining parallel processing productivity.
Data Source
AI summary
In some implementations, there is provided a method including determining whether a topology description includes a static physical host name and a static port address; in response to determining the topology description includes the static physical host name and the static port address, converting at least the topology description into a deployment file; including, in a container, the deployment file, code, and a control script that limits pre-start database operations to a single node of a plurality of nodes of a cloud platform; and deploying, the container including the deployment file, the code, and the control script, to the cloud platform to form, at the cloud platform, the plurality of nodes including the single node. Related systems and articles of manufacture are also disclosed.