Cloud Orchestration for Deterministic Time-Sensitive Networking in Microservices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Microservices architecture lacks deterministic packet delivery, which is crucial for time-sensitive networking requirements, especially in IoT applications, as existing time-sensitive networking models are not adapted to handle containers effectively.
Innovation Solution
A cloud orchestration platform determines an optimal path through microservices, computes necessary resources, and reserves timeslots and resources to ensure deterministic packet delivery by communicating with a scheduling driver to coordinate workload scheduling and resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional time-sensitive networking models are used, then deterministic packet delivery is achieved in traditional networks, but the models cannot be adapted to achieve deterministic packet delivery in microservices architecture with containers
Solution Approach 1:
The patent segments the networking model into container-level and microservice-level components. Each container is assigned specific timeslots for packet transmission, and this segmentation is extended to the microservice architecture level where service meshes and gateways are also assigned timeslots. This hierarchical segmentation allows deterministic networking to function at multiple levels of the microservices stack.
Solution Approach 2:
The patent creates a universal timeslot assignment mechanism that works across different networking layers (container network, microservice mesh, gateway). The same deterministic networking principles are applied universally across traditional networks and microservices architecture, making the model adaptable while maintaining deterministic packet delivery guarantees.
2Reliability
If containers are used to isolate applications in microservices architecture, then application isolation is achieved, but predictable latency and jitter cannot be guaranteed across the network
Solution Approach 1:
The patent applies preliminary action by pre-assigning timeslots to containers and microservices before packet transmission occurs. The cloud orchestration platform determines the path through microservices in advance and reserves timeslots for each segment of the path. This preliminary scheduling ensures that when packets are transmitted, the timing is already determined, guaranteeing predictable latency and jitter while maintaining container isolation.
3Reliability
If cloud orchestration platform determines paths and reserves timeslots, then deterministic networking is achieved, but system complexity increases
Solution Approach 1:
The patent introduces a cloud orchestration platform as an intermediary that manages the complexity of path determination and timeslot reservation. This intermediary abstracts the complex scheduling decisions from individual containers and microservices, centralizing the deterministic networking logic in a manageable location. The orchestration platform communicates scheduling requirements to scheduling drivers in a standardized manner, reducing overall system complexity while achieving deterministic networking.
Data Source
AI summary
A cloud orchestration platform obtains from a policy controller, application flow requirements for an application to be deployed in a container network that includes a plurality of microservices. The cloud orchestration platform determines a path through at least a subset of the plurality of microservices based on the application flow requirements, and computes information describing compute resources for workloads associated with the path through the plurality of microservices needed to support the application flow requirements. The cloud orchestration platform creates and/or reserves the workloads among the plurality of microservices. The cloud orchestration platform communicates scheduling requirements to a scheduling driver function associated with the plurality of microservices, the scheduling deriver function using the scheduling requirements to coordinate scheduling of workloads based on the path. The cloud orchestration platform directs traffic associated with the application into the container network.