Real time big data neworking
Download as PPT, PDF0 likes352 views
This document discusses top-down and bottom-up approaches to big data analytics. Top-down is periodic, repetitive, reflective, and backwards-looking, while bottom-up is real-time, predictive, actionable, and exploratory. The interactive investigation of new data using these approaches can create new knowledge and business opportunities. Big data analytics turns available operational information into real-time business advantages through interactive and experimental analysis. The platform discussed provides engines for processing large batches of data and real-time event streams to enable timely, actionable analytics and automated actions.
Real time big data neworking
- 1. How did I get here? Ericsson IP@Service™
- 2. › Top Down is – Periodic – Repetitive – Reflective – Backwards Looking / Prior Results Oriented › Bottom Up is – Real Time – Predictive – Actionable – Exploratory What if / Scenario Capable big data analytics is more than just business intelligence This top-down and bottom-up interactive investigation leverages new data, creating new knowledge, leading to new business scenarios / opportunities Turning available information on all aspects of the Operations into a real-time Business Advantage – Interactive – Experimental
- 3. Big data networking “(Yet another) Big Data definition: when the data volume become so big that it can no longer be economically back-hauled to one centralized database” Networking
- 4. Platform: interactive & Real time • Interactive analysis – Generate cube/models using customizable logic – Process, store & correlate huge amount of data – Horizontally scalable & batch oriented • Real time analysis – Independent & simpler processing of events – Analysis produce real time actionable events/results – Vertically scalable & quick response Timely & Actionable Analytics Automated Actions Data at Rest • Pricing guides • Product catalogs • Customer CRM • Asset & inventory • Employee data Data in Motion › Platform provides two core engines for such needs – Stream Engine processes data in small batches – Real time engine processes events in real time
- 5. Network Cloud Storage Processing Apps MapR/Batch MapR2.0/ Interactive Real time/ streaming SDN NFV Public Hybrid Private HDFS/DAS NAS“Software Def. Storage” A long term reference model DBMS/ Interactive
- 6. Real-time apache hadoop tools Google moved from MapReduce to Dremel 2 years ago
- 7. Real-time apache hadoop tools Google moved from MapReduce to Dremel 2 years ago
Editor's Notes
- #3: 2013-02-25 Big Data Analytics