Resource orchestration for microservices-based 5G applications
A method for performing resource orchestration for microservices-based 5G applications in a dynamic, heterogenous, multi-tiered compute and network environment is presented. The method includes managing compute requirements and network requirements of a microservices-based application jointly by positioning computing nodes distributed across multiple layers, across edges and at a central cloud, identifying and modeling coupling relationships between compute and network resources for a plurality of microservices, when only application-level requirements are provided, to build coupling functions, solving a multi-objective optimization problem to identify how each of the plurality of microservices are deployed in the dynamic, heterogenous, multi-tiered compute and network environment by employing the coupling functions to jointly optimize resource usage of the compute and network resources across different compute and network slices, and deriving optimal joint network and compute resource allocation and function placement decisions.
1. A method for performing resource orchestration for microservices-based 5G applications in a dynamic, heterogenous, multi-tiered compute and network environment, the method comprising:
managing compute requirements and network requirements of a microservices-based application jointly by:
positioning computing nodes distributed across multiple layers, across edges and at a central cloud;
identifying and modeling coupling relationships between compute and network resources for a plurality of microservices, when only application-level requirements are provided, to build coupling functions;
solving a multi-objective optimization problem to identify how each of the plurality of microservices are deployed in the dynamic, heterogenous, multi-tiered compute and network environment by employing the coupling functions to jointly optimize resource usage of the compute and network resources across different compute and network slices; and
deriving optimal joint network and compute resource allocation and function placement decisions.
2. The method of claim 1 , wherein inputs to the resource orchestration include application structure, end-to-end requirements, a state of infrastructure resources, and camera environment or video content.
3. The method of claim 1 , wherein the coupling functions are derived from historical data collected for a specific microservices-based application.
4. The method of claim 1 , wherein the multi-objective optimization problem includes a set of binary variables for function placement decision, a set of continuous variables for allocation of resources to each function, and a single continuous variable denoting performance of the microservices-based application.
5. The method of claim 1 , wherein solving the multi-objective optimization problem further enables reduction in overall usage and maximization of performance while satisfying application end-to-end latency and throughput.
6. The method of claim 1 , wherein the coupling relationships and limitations of the compute and network resources are modeled as hard constraints.
7. The method of claim 1 , wherein the multi-objective optimization problem is a mixed integer nonlinear program (MINLP).
8. A non-transitory computer-readable storage medium comprising a computer-readable program for performing resource orchestration for microservices-based 5G applications in a dynamic, heterogenous, multi-tiered compute and network environment, wherein the computer-readable program when executed on a computer causes the computer to perform the steps of:
managing compute requirements and network requirements of a microservices-based application jointly by:
positioning computing nodes distributed across multiple layers, across edges and at a central cloud;
identifying and modeling coupling relationships between compute and network resources for a plurality of microservices, when only application-level requirements are provided, to build coupling functions;
solving a multi-objective optimization problem to identify how each of the plurality of microservices are deployed in the dynamic, heterogenous, multi-tiered compute and network environment by employing the coupling functions to jointly optimize resource usage of the compute and network resources across different compute and network slices; and
deriving optimal joint network and compute resource allocation and function placement decisions.
9. The non-transitory computer-readable storage medium of claim 8 , wherein inputs to the resource orchestration include application structure, end-to-end requirements, a state of infrastructure resources, and camera environment or video content.
10. The non-transitory computer-readable storage medium of claim 8 , wherein the coupling functions are derived from historical data collected for a specific microservices-based application.
11. The non-transitory computer-readable storage medium of claim 8 , wherein the multi-objective optimization problem includes a set of binary variables for function placement decision, a set of continuous variables for allocation of resources to each function, and a single continuous variable denoting performance of the microservices-based application.
12. The non-transitory computer-readable storage medium of claim 8 , wherein solving the multi-objective optimization problem further enables reduction in overall usage and maximization of performance while satisfying application end-to-end latency and throughput.
13. The non-transitory computer-readable storage medium of claim 8 , wherein the coupling relationships and limitations of the compute and network resources are modeled as hard constraints.
14. The non-transitory computer-readable storage medium of claim 8 , wherein the multi-objective optimization problem is a mixed integer nonlinear program (MINLP).
15. A system for performing resource orchestration for microservices-based 5G applications in a dynamic, heterogenous, multi-tiered compute and network environment, the system comprising:
a memory; and
one or more processors in communication with the memory configured to:
manage compute requirements and network requirements of a microservices-based application jointly by:
positioning computing nodes distributed across multiple layers, across edges and at a central cloud;
identifying and modeling coupling relationships between compute and network resources for a plurality of microservices, when only application-level requirements are provided, to build coupling functions;
solving a multi-objective optimization problem to identify how each of the plurality of microservices are deployed in the dynamic, heterogenous, multi-tiered compute and network environment by employing the coupling functions to jointly optimize resource usage of the compute and network resources across different compute and network slices; and
deriving optimal joint network and compute resource allocation and function placement decisions.
16. The system of claim 15 , wherein inputs to the resource orchestration include application structure, end-to-end requirements, a state of infrastructure resources, and camera environment or video content.
17. The system of claim 15 , wherein the coupling functions are derived from historical data collected for a specific microservices-based application.
18. The system of claim 15 , wherein the multi-objective optimization problem includes a set of binary variables for function placement decision, a set of continuous variables for allocation of resources to each function, and a single continuous variable denoting performance of the microservices-based application.
19. The system of claim 15 , wherein solving the multi-objective optimization problem further enables reduction in overall usage and maximization of performance while satisfying application end-to-end latency and throughput.
20. The system of claim 15 , wherein the coupling relationships and limitations of the compute and network resources are modeled as hard constraints.