Bigdata implementations-caseStudy.pptx
-
Upload
vivek-biradar -
Category
Documents
-
view
228 -
download
0
Transcript of Bigdata implementations-caseStudy.pptx
Early 2012, client management decided to build a single platform for printer data analysis, capable to handle huge data volume in open source technologies which ensures low cost and high performance
Build an application that will accept the granular level attributes from the printer feeds and store in a Hadoop based Data Repository
The layered architecture wasExtract – JBossTransformation – HadoopLoad – Vertica Advanced analytics were
planned in Informatica & Vertica ( Later offloaded to Hadoop )
Project Overview
Big data Implementation : Case Study
The client is an American multinational information technology corporation. It provides products, technologies, software, solutions and services to consumers, small- and medium-sized businesses (SMBs) and large enterprises.
Client Overview
Process huge volume of semi structured live data in the order of 2 TB / day.
Data from different sources. Complex transformations & algorithms. Variety of data association & aggregations. Rapidly growing data volumes. Cost Efficient Hardware Scalability. Integration and Quality Assurance different
sources (Jboss & Vertica) Systems capability to reprocess subset of data Variety of reports to be generated Multi-Layer Architecture Infrastructure readiness & Network
connectivity Issues
Driver/ Issues/Challenges
Created an architecture to address client’s single platforms in Big dataCreated a generic parser framework layer to convert un/semi structured , multi formatted data to a structured dataCreated a validation framework to validate the data as per the business requirementsCreated a transformation layer to execute complex business transformationsCreated data formatter layer to format data in to customer expected formatSuccessfully implemented Big data solution addressing the below business area’s Active variety of Ink Jet Printers Passive variety of Ink Jet Printers Laser Printer ( in progress ) Web Based Printer ( in progress ) Advanced transformations for all printer’s
TCS Solution
Single Platform for all printer types. Optimal hardware resource utilization Data Analytics with historical, daily data. Maintained an Invalid data logging
mechanism. Which help customer to identify data lose.
Fully automated data processing and alerting mechanism
Advanced analytics also offloaded to Hadoop because of performance satisfaction factor
New business transformations are offloading to Hadoop
.
Results and /Business Benefits