Optimizing serverless computing using a distributed computing framework
Aspects of the technology provide improvements to a Serverless Computing (SLC) workflow by determining when and how to optimize SLC jobs for computing in a Distributed Computing Framework (DCF). DCF optimization can be performed by abstracting SLC tasks into different workflow configurations to determined optimal arrangements for execution in a DCF environment. A process of the technology can include steps for receiving an SLC job including one or more SLC tasks, executing one or more of the tasks to determine a latency metric and a throughput metric for the SLC tasks, and determining if the SLC tasks should be converted to a Distributed Computing Framework (DCF) format based on the latency metric and the throughput metric. Systems and machine-readable media are also provided.
1. A computer-implemented method for optimizing a Serverless Computing (SLC) workflow, comprising:
receiving a SLC job comprising SLC tasks;
executing the SLC tasks in the SLC job to determine a latency metric and a throughput metric for the SLC tasks;
converting at least one but not all of the SLC tasks to a Distributed Computing Framework (DCF) format based on the latency metric and the throughput metric; and
processing, in hybrid SLC/DCF pipelines, the SLC job as a combination of converted ones of the SLC tasks in DCF format and unconverted ones of the SLC tasks.
2. The computer-implemented method of claim 1 , wherein the converting comprises automatically converting the at least one but not all of the SLC tasks to DCF format.
3. The computer-implemented method of claim 2 , wherein automatically converting the at least one but not all of the SLC tasks to DCF format further comprises:
performing a map reduce function on the SLC tasks.
4. The computer-implemented method of claim 1 , wherein the latency metric is determined based on a time period required to fully execute the SLC tasks.
5. The computer-implemented method of claim 1 , wherein the throughput metric is determined based on a frequency that SLC tasks are triggered.
6. The computer-implemented method of claim 1 , wherein the converting further comprises:
comparing the latency metric to a predetermined latency threshold; and
converting the at least one but not all of the SLC tasks to DCF format if the latency metric exceeds the predetermined latency threshold.
7. The computer-implemented method of claim 1 , wherein the converting further comprises:
comparing the throughput metric to a predetermined throughput threshold; and
converting the at least one but not all of the SLC tasks to DCF format if the throughput metric exceeds the predetermined throughput threshold.
8. A system for optimizing a Serverless Computing (SLC) workflow, comprising:
one or more processors; and
a computer-readable medium comprising instructions stored therein, which when executed by the processors, cause the processors to perform operations comprising:
receiving a SLC job comprising SLC tasks;
executing of the SLC tasks to determine at least a latency metric and a throughput metric for the SLC tasks;
converting at least one but not all of the SLC tasks to a Distributed Computing Framework (DCF) format based on the latency metric and the throughput metric; and
processing, in hybrid SLC/DCF pipelines, the SLC job as a combination of converted ones of the SLC task in DCF format and unconverted ones of the SLC tasks.
9. The system of claim 8 , wherein the converting comprises automatically converting the at least one but not all of the SLC tasks to DCF format.
10. The system of claim 9 , wherein automatically converting the at least one but not all of the SLC tasks to DCF format further comprises:
performing a map reduce function on the at least one but not all of the SLC tasks of the SLC job.
11. The system of claim 8 , wherein the latency metric is determined based on a time period required to fully execute the SLC tasks.
12. The system of claim 8 , wherein the throughput metric is determined based on a frequency that the SLC tasks are triggered.
13. The system of claim 8 , wherein the converting further comprises:
comparing the latency metric to a predetermined latency threshold; and
converting the at least one but not all of the SLC tasks to DCF format if the latency metric exceeds the predetermined latency threshold.
14. The system of claim 8 , wherein the converting further comprises:
comparing the throughput metric to a predetermined throughput threshold; and
converting the SLC tasks to DCF format if the throughput metric exceeds the predetermined throughput threshold.
15. A non-transitory computer-readable storage medium comprising instructions stored therein, which when executed by one or more processors, cause the processors to perform operations comprising:
receiving a SLC job comprising SLC tasks;
executing SLC tasks to determine at least a latency metric and a throughput metric for the SLC tasks;
converting at least one but not all of the SLC tasks to a Distributed Computing Framework (DCF) format based on the latency metric and the throughput metric; and
processing, in hybrid SLC/DCF pipelines, the SLC job as a combination of converted ones of the SLC task in DCF format and unconverted ones of the SLC tasks.
16. The non-transitory computer-readable storage medium claim 15 , wherein
the converting comprises automatically converting the at least one but not all of the SLC tasks to DCF format.
17. The non-transitory computer-readable storage medium comprising claim 16 , wherein automatically converting the at least one but not all of the SLC tasks to DCF format further comprises:
performing a map reduce function on the at least one but not all of the SLC tasks of the SLC job.
18. The non-transitory computer-readable storage medium claim 15 , wherein the latency metric is determined based on a time period required to fully execute the SLC tasks.
19. The non-transitory computer-readable storage medium claim 15 , wherein the throughput metric is determined based on a frequency that the SLC tasks are triggered.
20. The non-transitory computer-readable storage medium claim 15 , wherein the converting further comprises:
comparing the latency metric to a predetermined latency threshold; and
converting the at least one but not all of the SLC tasks to DCF format if the latency metric exceeds the predetermined latency threshold.