Key Considerations in Optimizing the Deployment of Big Data Analytics-As-A-Service Utilizing Cloud Architecture and Machine Learning

Cloud computing is usually associated with storage and processing on a back-end server that is accessible over the Internet. Increasingly the Cloud has transcended the boundaries of storage and retrieval and moved into the realms of offering services, performing analytics, undertaking collaborations and much more while ensuring security and privacy of data for the user applications. As a result, deployment of any application or service on the Cloud requires a carefully constructed strategy - especially with respect to managing the dynamicity and balancing of that deployment. This strategic need for Cloud Architecture is further important because of the advent of Big Data and Analytics-as-a-Service (AaaS). The Analytical services utilizing Big Data are not limited to a specific Cloud server. Instead, Big Data Analytics are carried out across the entire spectrum of Internet-based nodes ranging from the back-end Cloud server through to the End-user Internet of Things (IoT) devices and everything in between. The time, location and granularity of Big Data Analytics on and off the Cloud is a crucial strategic question. The Quality of Service (QoS) and security of deployment of Analytics-as-a-Service depends on the key considerations in answering this question. This strategic question relates to the dynamic decision making required to deploy a Cloud-based Big Data Analytics solution – which, in turn, is based on understanding the current conditions – security, volume, performance, criticality, among others – of Cloud-based deployment. The need for automation and intelligence with the dynamic optimization of Analytics on the Cloud requires the application of Machine Learning. This paper explores these key considerations of the strategic aspects of deploying Big Data Analytics using a Cloud Architecture. The practical application of these key considerations is demonstrated through the education domain. Finally, this paper proposes areas of research emanating from the study of Cloud Architecture and Machine Learning for Big Data Analytics