analyse en temps réel de bigdata
TRANSCRIPT
L’analyse en temps réel de Big Data Le monitoring de flux par l’exempleAbdellatif BOUCHAMA@a_bouchama
Abdellatif BOUCHAMA@A_BOUCHAMA
• Middleware architect, and passionate about the new technology: Big Data,IoT and Open source
• Co-founder of BusHorn.com
Agenda
Big Data in 2015: Trends & Statistics
The emergence of real time
Use Case: Flow Activity Monitoring
Demo
What is Big Data?
Big Data
Velocity 30KB/s --> 30GB/s· Batch · Real-Time · Streaming
Volume
1-2 Terabytes à ∞
Variety
· Structured · Semi-Structured · UnStructured
Big Data Expectations:Things That Excite Executives About Big Data
It’s the “next oil.” from the Ginni Rometty, CEO of IBM
The potential to revolutionize industries, and change business models (e.g., Uber and Airbnb), with what we learn from the data.
The ability to provide real-time graphing solutions for data relationships.
Real-time operational and business data allows people to make decisions faster thereby saving significant money.
The opportunities it gives us to help clients solve real business problems.
https://dzone.com/articles/14-things-that-excite-executives-about-big-data?oid=big_data
Big Data view by Gartner:
Big Data in 2015:
75% of Companies Are Investing or Planning to Invest in Big Data in the Next Two Years
Goals for Big Data initiatives:1. Enhancing the customer experience2.Streamlining existing processes3.Achieving more targeted marketing and reducing costs
Last year, 37% of big data projects were initiated by the CIO, while 25% were initiated by business unit heads. In 2015, the roles are nearly tied, at 32% and 31 %, respectively.
http://www.gartner.com/newsroom/id/3130817
The emergence of Real-Time
Low
Pure BatchOperationalizing
Near Real TimeOr InteractiveAnalytics
High
Real Time Analytics
• Right time
Real time
• Smart data
Big Data
Use case: Flow Activity Monitoring
Architecture
Constraints & Requirements
Constraints• Non intrusive system• No modification on business flows.• We can plug it and unplug it easily.Requirements• System cost should be mastered
and adaptable• System automatized• Measurement
Why Elasticsearch stack?Open source
Easy to deploy
Distributed
Linear scalability
REST Interface
Kibana & Logstash to complete the Landscape
Data flow and constraints
Collect
JMX
Store & transport
Transform & Access
Modelize & Analyze
Visualize & Predict
JMS
Demo Time!Be prepared for it to fail, because demos always do
What’s Next ?• Enrichment analysis, after adding
business information in the logs
• Integration of backend and frontend applications
• Audit
Thank you