Enabling scatter-gather communication between serverless function on a novel framework

Kumar, Saurabh (2020) Enabling scatter-gather communication between serverless function on a novel framework. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (826kB) | Preview

Abstract

Serverless computing is a concept in the cloud computing world where the developer writes code in any of the high level languages with all its dependencies pushes it to the cloud platform for execution. The code written by the developer is called as serverless functions, these functions are executed on the trigger of an event. There are many providers of the serverless platform like the lambdas by Amazon Web Services, azure functions by Microsoft Azure and cloud functions by Google Cloud Platform are the well-known commercial enterprise level providers of the serverless architecture. The most prevalent use case of the serverless platform is IoT and microservices. However, the serverless platform is incapable of handling distributed computing tasks that involve communication between the serverless functions. The communication patterns like scatter-gather and point to point communication are difficult to establish. As the address of the function spawned is unknown to the developer, communication between the functions is impossible. Since the cloud computing infrastructure has access to vast resources and incomprehensive computing power, the power of distributed computation can be leveraged on the resources provided by the platform. As a part of the thesis, an artifact is developed which solves the scatter and gather problem on the serverless platform. The application could launch containers on uploading a file and distributed its chunks to the containers to process it parallelly and then get results. Using this approach, the serverless platform can be used for solving scatter and gather communication problems. This artifact shows that the address of the spawned nodes can be known to a master process. This artifact is a library that shows that it is possible to establish communication between serverless functions, but it lacks features of container creation, keeping the system highly available and fault-tolerant.

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Cloud computing
Divisions:	School of Computing > Master of Science in Cloud Computing
Depositing User:	Dan English
Date Deposited:	28 Jan 2021 14:59
Last Modified:	28 Jan 2021 14:59
URI:	https://norma.ncirl.ie/id/eprint/4541

Actions (login required)

View Item