Resource-aware compiler for multi-cloud function-as-a-service environment
In a multi-cloud computing environment including a plurality of cloud platforms, wherein each of the cloud platforms supports a different cloud-specific application function model that enables execution of cloud-specific application functions thereon, a method obtains a set of cloud-neutral application functions developed in accordance with a cloud-neutral application function model. The method then translates the set of cloud-neutral application functions into a set of cloud-specific application functions, wherein at least one cloud-specific application function of the set of cloud-specific application functions is executable on a corresponding one of the cloud platforms and another cloud-specific application function of the set of cloud-specific application functions is executable on another corresponding one of the cloud platforms. Further, the corresponding cloud platforms are selected based on one or more resource constraints and/or one or more resource costs.
1. A method comprising:
in a multi-cloud computing environment comprising a plurality of cloud platforms, wherein each of the cloud platforms supports a different cloud-specific application function model that enables execution of cloud-specific application functions thereon;
obtaining a set of cloud-neutral application functions developed in accordance with a cloud-neutral application function model; and
translating, by a compiler, the set of cloud-neutral application functions into a set of cloud-specific application functions that are executable across two or more of the cloud platforms of the multi-cloud computing environment, wherein at least one cloud-specific application function of the set of cloud-specific application functions is executable on a corresponding one of the cloud platforms and another cloud-specific application function of the set of cloud-specific application functions is executable on another corresponding one of the cloud platforms;
wherein translating the set of cloud-neutral application functions into a set of cloud-specific application functions further comprises generating one or more additional cloud-specific applications functions to effectuate interoperability between two or more of the cloud platforms;
wherein the corresponding cloud platforms for which the cloud-neutral application functions are to be translated are selected based on at least one or more resource constraints and one or more resource costs as provided by a resource predicator and a resource mapping logic of the compiler; wherein the resource predicator passes estimated resource usage to the resource mapping logic for recommendation output; and wherein the method is implemented via one or more processing devices each comprising a processor coupled to a memory.
2. The method of claim 1 , further comprising profiling execution of one or more cloud-specific application functions on one or more of the cloud platforms.
3. The method of claim 2 , wherein the profiling step acquires knowledge about how one or more cloud-specific application functions consume resources once deployed on the one or more cloud platforms.
4. The method of claim 2 , further comprising utilizing results of the profiling step to assist in selecting the cloud platforms for which the cloud-neutral application functions are translated.
5. The method of claim 1 , further comprising predicting resource usage for one or more cloud-specific application functions executing on one or more of the cloud platforms.
6. The method of claim 5 , further comprising utilizing the predicted resource usage to assist in selecting the cloud platforms for which the cloud-neutral application functions are translated.
7. The method of claim 1 , further comprising building a resource mapping model that represents resource constraints and pricing for each of the cloud platforms.
8. The method of claim 7 , further comprising utilizing the resource mapping model to assist in selecting the cloud platforms for which the cloud-neutral application functions are translated.
9. The method of claim 1 , wherein the set of cloud-specific application functions are instrumented with profiling code.
10. The method of claim 1 , wherein the resource mapping logic determines an optimal resource overlay.
11. The method of claim 1 , wherein each of the cloud-specific application function models is in the form of a cloud-specific Function-as-a-Service (FaaS) model.
12. The method of claim 1 , wherein the plurality of cloud platforms comprises a combination of private and public cloud platforms.
13. An article of manufacture comprising a non-transitory processor-readable storage medium having stored therein program code of one or more software programs, wherein the program code when executed by at least one processing device causes said at least one processing device to perform steps of:
in a multi-cloud computing environment comprising a plurality of cloud platforms, wherein each of the cloud platforms supports a different cloud-specific application function model that enables execution of cloud-specific application functions thereon;
obtaining a set of cloud-neutral application functions developed in accordance with a cloud-neutral application function model;
translating, by a compiler, the set of cloud-neutral application functions into a set of cloud-specific application functions that are executable across two or more of the cloud platforms of the multi-cloud computing environment, wherein at least one cloud-specific application function of the set of cloud-specific application functions is executable on a corresponding one of the cloud platforms and another cloud-specific application function of the set of cloud-specific application functions is executable on another corresponding one of the cloud platforms;
wherein translating the set of cloud-neutral application functions into a set of cloud-specific application functions further comprises generating one or more additional cloud-specific applications functions to effectuate interoperability between two or more of the recommended cloud platforms and
wherein the corresponding cloud platforms for which the cloud-neutral application functions are to be translated are selected based on at least one or more resource constraints and one or more resource costs as provided by a resource predicator and a resource mapping logic of the compiler; wherein the resource predicator passes estimated resource usage to the resource mapping logic for recommendation output.
14. The article of manufacture of claim 13 , wherein the resource mapping logic determines an optimal resource overlay.
15. A system comprising:
in a multi-cloud computing environment comprising a plurality of cloud platforms, wherein each of the cloud platforms supports a different cloud-specific application function model that enables execution of cloud-specific application functions thereon;
one or more processing devices configured to implement a compiler, wherein the compiler is configured to: obtain a set of cloud-neutral application functions developed in accordance with a cloud-neutral application function model;
translate, by a compiler, the set of cloud-neutral application functions into a set of cloud-specific application functions that are executable across two or more of the cloud platforms of the multi-cloud computing environment, wherein at least one cloud-specific application function of the set of cloud-specific application functions is executable on a corresponding one of the cloud platforms and another cloud-specific application function of the set of cloud-specific application functions is executable on another corresponding one of the cloud platforms;
wherein translating the set of cloud-neutral application functions into a set of cloud-specific application functions further comprises generating one or more additional cloud-specific applications functions to effectuate interoperability between two or more of the recommended cloud platforms and
wherein the corresponding cloud platforms for which the cloud-neutral application functions are to be translated are selected based on at least one or more resource constraints and one or more resource costs as provided by a resource predicator and a resource mapping logic of the compiler; wherein the resource predicator passes estimated resource usage to the resource mapping logic for recommendation output.
16. The system of claim 15 , wherein the compiler is further configured to profile execution of one or more cloud-specific application functions on one or more of the cloud platforms.
17. The system of claim 15 , wherein the compiler is further configured to predict resource usage for one or more cloud-specific application functions executing on one or more of the cloud platforms.
18. The system of claim 15 , wherein the compiler is further configured to build a resource mapping model that represents resource constraints and pricing for each of the cloud platforms.
19. The system of claim 15 , wherein the set of cloud-specific application functions are instrumented with profiling code.
20. The system of claim 15 , wherein the resource mapping logic determines an optimal resource overlay.