#dataoncloud london event
TRANSCRIPT
#DataOnCloud
TAME DATA ON CLOUD.
Welcome to #DataOnCloud
From this very room Winston Churchill said, “This is the room from which I will direct the war.”
NETWORK. BRAINSTORM. TAME DATA.
Agenda
Data On Cloud Data problems. Why Cloud. Myth busting. Solution RoadmapBy Wade Wegner, CTO, Aditi Technologies
Platform ChoicesThe latest on Windows AzureBy Simon Karn, Azure Platform Partner, Microsoft
Q&A PanelDiscover Risks, Strategies & Roadmap for Cloud adoption
ACCELERATE AND DE-RISK JOURNEY TO CLOUD
DIFFERENTIATE WITH DESIGN AND USER EXPERIENCE
DELIVER SCALE AND AGILITY
WE HELP OUR CLIENTS…
MOVE THEIR
BUSINESS TO THE
CLOUD
Our Promise
You Have a Data Problem
Quality of Data
Derive Valuable Insights
Massive Amounts of Data
Budgets for Data Growth
Are You Experiencing …
High Volume Data Growth
Quality of Data
Increased Frequency of
Data Collection
Data Beyond Relational
From Where Does this Data Come?
Device + Sensors
Social Feeds
Relational Databases
Trading Desks Web Logs
Document Stores
No SQL or Table Storage
SQL
How Do We Use this Data?
KPI Dashboards
Trading Stations
Personalized Web
BusinessAlertsand Notifications
What’s the Opportunity?
Power and Utility
Manufacturing
Sensor on Plant Floor10,000 events/sec
Web Analytics
Click-Stream Data etc.Personalized pages100,000 events/sec.
Financial Services
Fraud DetectionAlgorithmic Trading100,000 events/sec.
Energy ConsumptionSmart Grids100,000 events/sec.
So, What Exactly is the Data Problem?
VarietyVelocityVolume Veracity
Case STUDY # 1: HealthCare Company Improves Hospital Hygiene Using Sensor Data
Aggregate and report “Hand Hygiene Compliance” for hospitals
Identify increased patient risk, provide notification in real-time
Azure based storage for unstructured data - RFID data and video collection for 100+ doctors
Azure based computation scalability and push live notifications
Single sign-on authentication using Active Directory
Render reports using predefined views
Storage scalability
Reduced costs
Streamlining unstructured data
Case STUDY # 2: Big European Travel Conglomerate Optimizes Product Pricing
Price their products better based on competitor data using web crawlers
Collect web logs to analyze customer behavior and deliver better pricing
Deliver predictive models to forecast future prices of products during holiday seasons
Utilize Cloud storage to capture data from web crawlers as raw data
Use Hadoop to segment and identify best price from logs and crawler data
Scheduling and computing using the cloud Render KPIs using visualizations Run machine learning algorithms
Storage scalability
Idea to production in six weeks
High performance computing
What Do We Mean By Cloud?
• On-demand self service• Broad network access• Resource pooling• Rapid elasticity• Measured serviced
• Software as a Service• Platform as a Service• Infrastructure as a Service
Characteristics Service Models
How Does Cloud Solve the 4V’s?
VarietyVelocityVolume Veracity
How Cloud Helps Solve the Data Problem
↑ Ability to add storage dynamically
↑ Increase computing power on demand
↑ Use global distributed data centers for localized processing
High Volume Data Growth
VOLUME
How Cloud Helps Solve the Data Problem
↑ Use Azure networks to collect data with very low latency
↑ Leverage CEP on Azure to do real time event processing
↑ Distribute notifications and alerts
VELOCITYIncreased
Frequency of Data
Collection
How Cloud Helps Solve the Data Problem
↑ Azure supports Relational, No SQL and Blob locally
↑ Ability to process and enrich all kinds of data using HDInsights
↑ Combine relational and non relational data in one service
VARIETY
Data Beyond Relational
How Cloud Helps Solve the Data Problem
↑ Ability to add storage dynamically
↑ Increase computing power on demand
↑ Use global distributed data centers for localized processing
VERACITY
Quality of Data
Approach for USING DATA with the CLOUD
Aggregate AnalyzeEnrich
Aggregate
Fragmented information
sourcesNon
relational information
Unclean data DATA SOURCE
Relational historic data
DATA INJECTIONUse Data hub to load data into Azure Blob
storage
Classify data into tables, blobs, SQL
Azure
Enable the blob storage as HDFS for
HDInsights
Enrich
Filter data using
MAPREDUCEREFINE
TRANSFORM
CLEANSE
Apply transformations
Segment data based on multiple variables
Remove duplicates
Eliminate non required information
Leverage HIVE to use HDInsights as a
DW
Prepare and load it into relational format if
required
Load data into clusters using
PIG
Analyze
ANALYZE
VISUALIZEAccess HDFS data using Excel data
explorer
Implement Embedded visualizations using Power
view
Leverage machine learning
Deliver alerts and notifications
Implement statistical algorithms like Naïve
baiyes,Clustering
Process real time business events using StreamInsight
Microsoft’s Investment in Data Services
Power ViewExcel with Data Explorer AlertsPredictive Analytics
APPsLOBCRMERP
SQL Server Relational DW
Statistical models CEP
Devices CrawlersSensors Bots
HDInsights
Challenges & Mitigations
Compliance ComplexityData Security & Privacy
How Do We Make Sense of this Data?
Right PersonRight TimeRight Data
Starting the Journey
Data & Cloud Quickstart• Half-day with an Architect• Detailed review of data
challenges and cloud maturity
Additional Quickstarts
• Cloud Application Portfolio Assessment• HA SQL Server in the Cloud• Migrating SharePoint Workloads to the Cloud• Cloud-Based Dev/Test Environments• Cloud-Based Core Infrastructure• AD/IAM in the Cloud
Web | Blog | Facebook | Twitter | LinkedIn