GPU-Enabled Serverless Workflows for Efficient Multimedia Processing

Abstract

Serverless computing has introduced scalable event-driven processing in Cloud infrastructures. However, it is not trivial for multimedia processing to benefit from the elastic capabilities featured by serverless applications. To this aim, this paper introduces the evolution of a framework to support the execution of customized runtime environments in AWS Lambda in order to accommodate workloads that do not satisfy its strict computational requirements: increased execution times and the ability to use GPU-based resources. This has been achieved through the integration of AWS Batch, a managed service to deploy virtual elastic clusters for the execution of containerized jobs. In addition, a Functions Definition Language (FDL) is introduced for the description of data-driven workflows of functions. These workflows can simultaneously leverage both AWS Lambda for the highly-scalable execution of short jobs and AWS Batch, for the execution of compute-intensive jobs that can profit from GPU-based computing. To assess the developed open-source framework, we executed a case study for efficient serverless video processing. The workflow automatically generates subtitles based on the audio and applies GPU-based object recognition to the video frames, thus simultaneously harnessing different computing services. This allows for the creation of cost-effective highly-parallel scale-to-zero serverless workflows in AWS.

A Holistic View on Resource Management in Serverless Computing Environments: Taxonomy and Future Directions

“…For certain application scenarios, serving all the queries using expensive hardware accelerators may not be economically viable (e.g., ML inference queries). For such scenarios, hybrid approaches to opportunistically serve the incoming requests using a mix of GPU-enabled instances and traditional CPU-only instances, could be explored [138], [103]. Research eforts are also directed towards highlighting the adaptability of quantum computing in the design of serverless systems [54], [55].…”

Serverless Computing: A Survey of Opportunities, Challenges, and Applications

“…It achieves 7.9 times lower latency and 17.2 times cost reduction on average compared to that of serverful alternatives. In [126] GPU processing power is harnessed in a serverless setting for video processing. Zhang et al, [168] present a measurement study to extract contributing factors such as the execution duration and monetary cost of serverless video processing approaches.…”

Serverless Workflows for Containerised Applications in the Cloud Continuum

“…For those applications that do not fit within AWS Lambda's computing requirements, SCAR provides a seamless integration with AWS Batch [5] an elastic-cluster as a service offering by AWS which dynamically deploys a cluster in charge of executing jobs packaged as a Docker images and which can grow and shrink depending on the number of jobs queued up at the Local Resource Management System (LRMS). This integration allows to delegate into AWS Batch functions invocations that require longer execution times, larger amount of memory or even GPU resources for accelerated execution, as described in the work by Risco et al [56].…”