Storage and Analysis of Big Data Tools for Sessionized Data

The Oracle database currently used to mine data at PEGGY is approaching end-of-life and a new infrastructure overhaul is required. It has also been identified that a critical business requirement is the need to load and store very large historical data sets. These data sets contain raw electronic consumer events and interactions from a website such as page views, clicks, downloads, return visits, length of time spent on pages, and how they got to the site / originated. This project will be focused on finding a tool to analyze and measure sessionized data, which is a unit of measurement in web analytics that captures either a user\u27s actions within a particular time period, or the process of segmenting user activity of each user into sessions, each representing a single visit to the site. This sessionized data can be used as the input for a variety of data mining tasks such as clustering, association rule mining, sequence mining etc (Ansari. 2011) This sessionized data must be delivered in a reorganized and readable format timely enough to make informed go-to-market decisions as it relates to the current and existing industry trends. It is also pertinent to understand any development work required and the burden on the resources. Legacy on-premise data warehouse solutions are becoming more expensive, less efficient, less dynamic, and unscalable when compared to current Cloud Infrastructure as a Service (IaaS) that offer real time, on-demand, pay-as-you-go solutions . Therefore, this study will examine the total cost of ownership (TCO) by considering, researching, and analyzing the following factors against a system wide upgrade of the current on-premise Oracle Real Application Cluster (RAC) System: High performance: real-time (or as close to as possible) query speed against sessionized data SQL compliance Cloud based or, at least a hybrid (read: on-premise paired with cloud) Security: encryption preferred Cost structure: cost-effective pay-as-you-go pricing model and resources required for the migration and operations. These technologies analyzed against the current Oracle database are: Amazon Redshift Google Bigquery Hadoop Hadoop + Hive The cost of building an on-premise data warehouse is substantial. The project will determine the performance capabilities and affordability of Amazon Redshift, when compared to other emerging highly ranked solutions, for running e-commerce standard analytics queries on terabytes of sessionized data. Rather than redesigning, upgrading, or over purchasing infrastructure at a high cost for an on-premise data warehouse, this project considers data warehousing solutions through cloud based infrastructure as a service (IaaS) solutions. The proposed objective of this project is to determine the most cost-effective high performer between Amazon Redshift, Apache Hadoop, and Google BigQuery when running e-commerce standard analytics queries on terabytes of sessionized data