Date: December 2014
Developer and Presenter: Farhana Zulkernine
Supported By: SOSCIP Consortium and IBM Canada
Project Team Members: Patrick Martin (Primary Investigator), Michael Bauer, and Ying Zou.
Cloud-based Analytics-as-a-Service (CLAaaS) or simply Analytics-as-a-Service (AaaS) facilitates big data analytics for all group of users. CLAaaS is hosted on the Cloud and provides web-based access to a wide variety of analytics tools and convenience of performing data analytics without the worries about resource management for big data. Based on the service level agreement (SLA), multiple user roles are supported by the service so that a data scientist can upload the data on the cloud and define multi-step data analytic workflows and a management user can execute the workflows or simply access the analyzed data or reports. With the software tools such as Cognos or BigInsights provided by IBM on the cloud, dashboards are created to demonstrate how information can be shared with others in the organization. WINGS or Workflow Instance Generation and Selection is an open source research tool, which is integrated in CLAaaS to allow users to create meta data for the uploaded data and define complex workflows with custom tools and components. These workflow definitions can be stored and executed at a later point with newer data. WINGS separates the definition and execution of workflows, and thereby, enables integration of 3rd party execution engines such as Pegasus for distributed and parallel execution of the workflows. The front-end of CLAaaS provides a software-as-a-service on any mobile device allowing the user to log in and start execution of a predefined workflow. Once signed in, data scientists can access various analytic tools through platform-as-a-service. Results can be published using the tools hosted on the cloud for convenient data sharing.
We present the demo in multiple parts starting with an overview of CLAaaS or AaaS. The subsequent parts demonstrate the different functionality provided by the service starting with registering with a SLA, data uploading and accessing analytic tools using platform-as-a-service, and defining and executing complex multi-component workflows using the WINGS interfaces.
Click here to view screenshots of the platform.