CloudButton ‘democratize’ big data simplifying the programming model as well as the software lifecycle by using serverless technologies.
“Why the cloud doesn’t have a button?”. This simple question, raised by a Berkley professor is the origin of CloudButton idea, as it contains his pupils need of simplifying cloud deployment and execution.
Based on this, CloudButton proposes an architecture based on three pillars:
- To develop the first FaaS (Function as a Service) environment for data analytics using Apache OpenWhisk. This platform operates the serverless architecture paradigm for abstracting resources from the end user. In this way, is the cloud resource provider who manages the application deployment and execution based on user needs.
- To develop mutable data distributed structures using Infinispan. The project has developed a middleware which provides the needed functionalities to ensure the persistence, trust and control of concurrent data in serverless functions.
- To develop the CloudButton Toolkit, which provides the abstraction layer to manage intensive data applications based on the previous points. New tools and methods based on HPC, machine learning and data analysis technologies have been developed to simplify the application migration to the CloudButton platform.
Being a horizontal solution, it is not easy to understand how it works or which are the potential benefits for the end user. For this reason, CloudButton is demonstrated through three different use cases, all of them with high computational needs to analyse big volumes of data:
- Analysis of genomic data: resource limitations in research centres can harm scalability, but this can be solved using serverless technologies improving not only scalability but also productivity while processing big data volumes.
- Analysis of metabolomic data: the proposed solution improves the capacity of analysing non-pre-processed data, while allowing the access and reuse of open data sources.
- Analysis of geospatial data: combining satellite and LiDAR open access data for improving productivity, scalability and performance of environmental applications.
To know more about the project and its potential applications, visit CloudButton website.