Upload
sqlstream-inc
View
489
Download
2
Embed Size (px)
DESCRIPTION
The Internet of Things relies on networks of connected devices at unprecedented scale and complexity. Maximizing the opportunity requires organizations to collect and process high velocity sensor data in real-time. This Webinar presentation shows how real-time operational intelligence is shaping the Internet of Everything, transforming log file and sensor machine data streams into actionable, real-time intelligence. Use cases from Intelligent Transportation, Telematics and M2M are presented.
Citation preview
If you haven’t dialed into the audio portion, please do so now:
U.S.A +1 (646) 307-1721| 789-157-692
T H A N K YO U F O R J O I N I N G ! T H E W E B I N A R I S A B O U T TO S TA R T
Listen to Your Sensors: a Tale of Managing Large-scale, Intelligent Sensor Networks in Real-time
| 2 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Explain real-time Big Data and Operational Intelligence
¤ The principles of streaming data management
¤ Share our thoughts, experiences and use cases
¤ Audience Q&A
PROGRAM MISS ION
| 3 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Introduction (5 min)
¤ Presentation (35 min)
o An Introduction to Operational Intelligence
o Solution Architecture for Real-time Sensor Analytics
o Streaming Big Data management
o Industry Use Cases
o Demonstration (Remote sensor monitoring)
o The Total Cost of Performance
¤ Q | A (20 min)
JULY 16 AGENDA
| 4 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ July 9 2013 |10:00am PST- recording available upon request Analytics, Predictive Analytics, Prescriptive Analytics: The Anatomy of Operational Intelligence
¤ July 16 2013 |11:00am PST Listen to your Sensors: A Tale of Managing Large Scale Sensor Networks in Real-time
¤ July 23 2013 |11:00am PST Predict and Avert: Using Log File Data to Prevent Cybersecurity and Fraud Attacks in Real-time
¤ July 30 2013 |10:00am PST No more CPR for your CDRs: Meet Real-time Traffic Utilization, Billing and Fraud Detection
T he Opera t i ona l I n t e l l i gen ce Se r i e s
| 5 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ 25 years of experience in database and software applications
¤ Senior positions with Oracle, Hyperion Solutions, Information Builders and Algebraix Data
¤ Recent guest speaker at 2013 Sensors Expo
Today ’s P re sen te r : G lenn Hou t SQL s t ream VP fo r t he Amer i ca s
| 6 Copyright © 2013 | +1 877 571 5775 | [email protected]
THE INTERNET OF EVERYTHING Manag i ng La rge - s ca l e Con ne c t ed Se n s o r Ne two r k s
TECHNOLOGY
¤ OPEX & CAPEX
¤ Experience
¤ Loyalty
BUSINESS CASE
¤ Velocity
¤ Volume
¤ Unstructured
REAL-TIME BIG DATA THE NEXT DIGITAL AGE
REAL-TIME
OPERATIONAL
INTELLIGENCE
50 Billion Connected
Devices and Sensors by
2020
(Cisco IBSG)
¤ Wireless
¤ IPv6 Networks
¤ Smart Energy
| 7 Copyright © 2013 | +1 877 571 5775 | [email protected]
ABOUT SQLSTREAM
facts
o Launched 2009
o Over 1.5M lines of code
o Multiple deployments across many industries
capabilities
o Unstructured and structured data
o Accelerates and extends Hadoop & RDBMS
o Not limited to SQL
innovations
o Only true streaming data management platform
o Only true standard SQL streaming engine
o Five patents for stream processing
ü Real-time Operational Intelligence ü High-velocity machine data. ü Streaming Big Data Management Platform
¤ TWEET: during and after the webinar, please
use #RTSensorData for live discussions
¤ DIRECT QUESTIONS: please use the box to the
right of your screen
¤ RECORDINGS: an edited version of the webinar
recording will be emailed upon request
Opera t i ona l I n t e l l i gen ce
| 10 Copyright © 2013 | +1 877 571 5775 | [email protected]
The In ter net -of -Th ings Explos ion | SOURCES E v e r y t h i n g i s b e c o m i n g i n s t r u m e n t e d ; v a r i e t y
ENVIRONMENTAL TRANSPORTATION NETWORKS
Environmental Monitoring Location-based services Machine-to-Machine
Smart Grid Cars as Sensors Logistics
| 11 Copyright © 2013 | +1 877 571 5775 | [email protected]
OPERAT IONAL INTEL L IGENCE I n t eg ra t i ng Opera t i on s and Ana ly t i c s i n Rea l - t ime
As we move toward a real-time business environment, the capability to process data flows swiftly and flexibly will become increasingly important. SQLstream leads the industry in this kind of capability. ”
Robin Bloor Chief Analyst for Bloor Group
Business Intelligence
Operations
Real-time Operational Intelligence Continuous monitoring and analytics
Improve decision-making
Automate operational processes
Environment Smart Grid
Transportation Telematics
M2M Remote Monitoring
”
| 12 Copyright © 2013 | +1 877 571 5775 | [email protected]
T he I n fo r ma t i on Va l ue Cha i n
What is happening?
What might happen?
What just happened?
Make stuff happen!
| 13 Copyright © 2013 | +1 877 571 5775 | [email protected]
COLLECTION: Massive numbers of sensors
o Thousands, tens-of-thousands, hundreds-of-thousands,
millions+ of sources
o Creation of multiple siloes
SCALE: Continuous “fire-hose”
o Massive data volumes and velocities
o Information overload
PROCESSING: Actionable intelligence
o Data “decision” value can drop exponentially with
passing seconds
o Data velocity and rate-of-change critical for decisions
o Lack of real-time processing
B ig Data Explos ion| CHALLENGES A n a l y t i c p r o c e s s i n g a t m a s s i v e s c a l e
BIG DATA
MACHINE DATA
HIGH-VELOCITY, LOW-LATENCY INTELLIGENCE
| 14 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Capabilities
o Health monitoring of key vehicle systems
o Frequent monitoring when “exception” system
events occur
o Real-time “panic” alerts
¤ Objectives
o Reduce vehicle “walkaway” events by 50%
o Improve customer satisfaction
o Reduce warranty costs
o Future: Potential “paid for” service
R EAL - T IME INTEGRATED VEH ICLE HEALTH MANAGEMENT
Warning! Battery draining fast. Please check for lights left on.
| 15 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Real-time vehicle diagnostics & alerting
o Pre-emptive vehicle service
o Critical event monitoring and alerting
¤ Asset tracking for fleet management
o Track vehicles, shipments and transit ETA
¤ Real-time driving log
o Safety, compliance and alerting
¤ Dynamic road tolling
o Assess tolls based on weight, miles, roads used
VEH ICLE TE LEMAT ICS & HEALTH MONITOR ING
Log
| 16 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Insurance Programs
o Breaks on insurance for good drivers
• Immediate “journey” reports to smart
phones
o Real-time analytics and alerting
• Safety, compliance and real-time
monitoring
¤ Geofencing
o Alerts to smartphones when drivers
travel outside prescribed boundaries
INSURANCE & SENSOR MONITOR ING
Driving Analysis: ……… ………………. ………….
| 17 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Real-time health diagnostics, monitoring and alerting
¤ Asset tracking and location-based services
¤ Location- and activity-based advertising
¤ Safety, regulatory and compliance
¤ Customer loyalty solutions
¤ Fleet management solutions
¤ Usage-based applications: insurance, tolling, traffic, etc.
¤ Emergency services and responder alerting
REAL - T IME SENSOR APPL ICAT IONS
| 18 Copyright © 2013 | +1 877 571 5775 | [email protected]
Information coming continuously or in irregular intervals
o Sensor data, log file data, health information (from cars, machinery, hardware, people), etc.
Anything that requires immediate attention
o Alerts that must be acted on now, not in minutes, hours or days
Any process that struggles with batch window latencies
o Service level agreements (SLA’s) with narrow time-action windows
o Batch queries that are continually re-run
o Data cleansing, transformation and loading (ETL)
Massive scale
o Information inflows so large that they must be continuously processed
R EAL - T IME APPL ICAT IONS | IND ICATORS
| 19 Copyright © 2013 | +1 877 571 5775 | [email protected]
Capability Real-Time System Traditional Database / Hadoop
Analytic Processing
• Continuous queries • Incremental calculations
• Batch queries • Expensive, batch recalculations
Latency from event to action
• Milliseconds to seconds • Tens-of-minutes, hours, days • Significant latency
Rate of Change • Monitor data velocity and
complex rates-of-change • Static snapshot of point-in-time
Scalability & Hardware
Requirements
• Millions to tens-of-millions of events/second
• Minimal hardware reqt’s
• Massive data inflow and continual recalculation limits analytic depth
• Significant hardware reqt’s
Iterative Analytics
• Cascading analytics within and across events - pipelining
• Requires myriad views, temporary tables and custom application code
QUERY-THEN-STORE VS. STORE -THEN-QUERY
| 20 Copyright © 2013 | +1 877 571 5775 | [email protected]
Category Operational Intelligence Business Intelligence Vehicle
Diagnostics Is my vehicle about to fail? How many vehicles failed in the last 6 months
in the state of New York?
Sensors ALERT - The airbag just deployed! What is the current location, vehicle and driver information?
How many vehicles had airbags that deployed in the past 2 weeks by vehicle type?
Telematics & Asset Tracking
Where is the vehicle right now? Where has it been and how is it moving on current journey?
What was the average speed of all vehicles on various road segments on prior journeys?
Advertising What is a relevant ad to place based on current location, current activity and historical patterns?
What is a relevant ad to place based on historical patterns?
Security Is an intrusion attempt underway? Where have intrusions happened in the past and from what sources?
Fraud/risk Is the current transaction fraudulent?
How many fraudulent transactions occurred at online electronics retailers during the past 3 months?
R EAL - T IME ANALYT ICS VS. H ISTOR ICAL ANALYT ICS
| 21 Copyright © 2013 | +1 877 571 5775 | [email protected]
R EAL - T IME , CONT INUOUS OPERAT IONAL INTEL L IGENCE
Real-time alerts, action and visualizations
Enhance real-time data by joining it with historical information
Persist both detail and aggregate data to historical archives
| 22 Copyright © 2013 | +1 877 571 5775 | [email protected]
“ R EPLAY” EVENTS TO DEVELOP PRED ICT IVE ALGOR ITHMS
Data Warehouse
DATABA
SE ADA
PTER
Replay events at high-speed
Test & refine predictive algorithms
Measure & report results and real-time exceptions
| 23 Copyright © 2013 | +1 877 571 5775 | [email protected]
R EAL - T IME OPERAT IONAL INTEL L IGENCE C o m p l i m e n t a r y w i t h E x i s t i n g S o l u t i o n s
Sensor Feeds
Data Transmission
Data Collection
Lightweight Agents (listeners)
Existing Archival Database(s)
| 24 Copyright © 2013 | +1 877 571 5775 | [email protected]
¤ Dashboards
o Simple drag-and-drop interface
o Multiple key performance
indicators on a single panel
o Real-time, continuous charting
capability on any metric
¤ Visualizations
o 100% web-based
o Variety of charting types and
customizable alerts
o Smart-phone access
o Stunning graphics capabilities
POWERFUL DASHBOARD ING & V ISUAL IZAT IONS
| 25 Copyright © 2013 | +1 877 571 5775 | [email protected]
Complementary
o Doesn’t replace business intelligence solutions, but provides powerful,
complementary operational intelligence capabilities
Massive scalability
o Ability to provide analytics and alerting on massive volumes of sensor feeds
with minimal infrastructure
Low Total Cost of Ownership (TCO)
o Minimal infrastructure requirements and rapid time-to-value
o Utilize existing business intelligence skill sets for real-time data
SUMMARY
S t ream ing Da ta Managemen t
| 27 Copyright © 2013 | +1 877 571 5775 | [email protected]
H igh - ve lo c i t y B ig Da ta Ana ly t i c s
Historical queries and data
enrichment
Storing valuable derived streams for future access
Ope
ratio
nal I
ntel
ligen
ce
Logs
Sensors
GPS
Networks
Social media
RFIDs
Servers
Telecom
Smart grid
Oil & Gas
Manufacturing
Logistics
M2M
Telematics
Retail
Internet
Banking
Data centers
Automotive
¤ Continuous Queries over Sliding Time Windows ¤ Analysis and Integration of Unstructured and Structured data ¤ Prescriptive Analytics drives Automated Actions
| 28 Copyright © 2013 | +1 877 571 5775 | [email protected]
Rea l - t ime A r c h i t e c t u re Streaming Analysis and Integration for Infinite Flows of Unstructured Data in Real Time
Streaming Agent & Adapter Layer + JDBC API Hadoop Streaming
Query Planner & Optimizer for MPP Execution SQL
Developer Tools
Platform Administration
Streaming SQL Real-time Applications
Real-time Dashboards & Visualization
Logs
Sensors
GPS
Networks Social Media Servers
M2M Telematics
Impala SQL
HBase
HDFS / MR
Hadoop for Stream Persistence, Enrichment & Replay (Optional)
Any external data warehouse, operational system and
enterprise platform
| 29 Copyright © 2013 | +1 877 571 5775 | [email protected]
CLEANING & FILTERING
STREAMING ANALYTICS
STREAMING AGGREGATION
CONTINUOUS INTEGRATION
Geospatial Applications Security
Monitoring Real-time
Dashboards Health
Monitoring
QoS and QoE
AN OPERAT IONAL INTEL L IGENCE P LATFORM
Logs Sensors GPS Networks Social Media Servers M2M Telematics
| 30 Copyright © 2013 | +1 877 571 5775 | [email protected]
T he SQL s t ream s -S t ream ing P roduc t Po r t fo l i o
s-Server Data Management Platform for Streaming Big Data
s-Analyzer Real-Time Visualization for Streaming
Operational Intelligence
s-Transport Geo-Analytics for Location-based
Applications
s-Visualizer Advanced
Visualization
s-Cloud s-Server EC2 AMI Deployment
s-St
udio
D
evel
oper
& A
dmin
Con
sole
To ta l Co s t o f Pe r fo r mance
| 32 Copyright © 2013 | +1 877 571 5775 | [email protected]
R E C O R D S P E R S E C O N D
To ta l Co s t Of Pe r fo r mance ( t o ta l COP ) T h e H i g h - Ve l o c i t y, L ow - L a t e n c y T i p p i n g Po i n t f o r B i g D a t a
Patterns Trends Mining Connections
Searches Inventory Reports Statistics Billing
SOCIAL E-COMM SECURITY TELEMATICS TELECOM
Trading Advertising Alerts Detection Signal
Intelligence
TO
TA
L C
OS
T
| 33 Copyright © 2013 | +1 877 571 5775 | [email protected]
Intelligence
TELECOM
Patterns Trends Mining Connections
Searches Inventory Reports Statistics Billing
Trading Advertising Alerts Detection Signal
SOCIAL E-COMM SECURITY TELEMATICS
R E C O R D S P E R S E C O N D
TO
TA
L C
OS
T To ta l Co s t Of Pe r fo r mance ( t o ta l COP ) T h e H i g h - Ve l o c i t y, L ow - L a t e n c y T i p p i n g Po i n t f o r B i g D a t a
Glenn Hout Email | [email protected]
Phone | 650.343.0864
Website | www.sqlstream.com
Upcoming events | www.sqlstream.com/webinars/
Q | A
| 35 Copyright © 2013 | +1 877 571 5775 | [email protected]
MACH INE -TO-MACH INE (M2M) APPL ICAT IONS Pe r fo r m a n c e , A n a l y t i c s a n d P r e d i c t i v e H e a l t h M o n i t o r i n g
¤ Transportation
Asset tracking and logistics and consumer Journey Time from GPS data
¤ Telematics
Vehicle health and driver insurance from vehicle sensor data
¤ Healthcare
Remote health monitoring applications from low power micro-sensors
¤ M2M Data Feed Monetization
Streaming data aggregation with subscription-based analytics
ü Improve uptime ü Avoid failures ü One platform
| 36 Copyright © 2013 | +1 877 571 5775 | [email protected]
S TREAMING SQL
Example: Compute Average vehicle speed across any subset of network
over rolling time windows from GPS events
- 3 -
In the example shown below, sensor data arriving in the “RoadPositionInfo” stream is used to
calculate average speed per zone over rolling one, five and ten minute windows:
CREATE OR REPLACE VIEW "SpeedZoneStats"
DESCRIPTION ‘Rolling averages for multiple windows partitioned by zone' AS
SELECT STREAM
"zone", -- zone id
"segmentid", -- parent road segment
"speedlimit", -- speed limit for zone
AVG("Speed")OVER last1Min AS "avgSpeed1", -- 1-min running average
AVG("Speed")OVER last5Min AS "avgSpeed5", -- 5-min running average
AVG("Speed")OVER last10Min AS "avgSpeed10" -- 10-mn running average
FROM "RoadPositionInfo"
WINDOW
last1Min AS (PARTITION BY "zone" RANGE INTERVAL '1' MINUTE PRECEDING),
last5Min AS (PARTITION BY "zone" RANGE INTERVAL '5' MINUTE PRECEDING),
last10Min AS (PARTITION BY "zone" RANGE INTERVAL '10' MINUTE PRECEDING);
The output of the application is a continuous stream of exception alerts based on pre-defined
conditions and rolling averages of speed information per road segment over various time windows.
Multiple analytics queries can execute in parallel over any number of different data streams, even
where the arrival rate of each data stream varies.
These simple analytics are used to drive a number of different applications, including real-time
Travel Time predictions, predictive analytics for congestion levels, and real-time alerts for vehicle
speeds in fixed and variable speed limit segments.
LOW LATENCY REQUIREMENTS FOR MULTI-MODAL SOLUTIONS
The case study presented was the first phase of a wider program to deliver a single low latency
platform for all transportation modes. Further real-time monitoring and process automation use
cases have been identified, as well as the integration of traditional data sources, for example:
• Per-segment historical speed/travel-time comparisons, for example, comparisons, in real-time, with the same period yesterday, last week or even last year
• Combine GPS sensor data streams in real-time with roadside camera and signal data, to achieve improved accuracy for congestion and travel time predictions
• Integration with roadside variable speed signs, providing dynamic adjustment of per-segment speed limits in order to respond in real-time to changing traffic conditions
• Driver and user access to the application, providing for example, real-time Travel Time through Smartphones and GPS devices
• Extend Travel Time application to provide end to end journey travel time across multiple transportation modes including heavy vehicles, rail, bus and ferry networks.
36 Copyright © 2012 Proprietary information of SQLstream Inc. All rights reserved