Upload
inside-analysis
View
482
Download
2
Tags:
Embed Size (px)
Citation preview
H T Technologies 2013
HOST: Eric Kavanagh
THIS YEAR is…
INVESTIGATIVE ANALYTICS =
� Seeking previously unknown patterns in data
� Extracting real-‐time insight from machine generated data
� Being able to query data streams, i.e., mobile data, web logs, geospatial data, social media data, etc.
ANALYST:
Philip Howard Research Director, Bloor Research
ANALYST:
Robin Bloor Chief Analyst, The Bloor Group
GUEST:
Don DeLoach CEO & President, Infobright TH
E LINE UP
INTRODUCING
Philip Howard
Exploiting the Internet of Things with Investigative Analytics
Philip Howard Research Director, Bloor Research
telling the right story Confidential © Bloor Research 2013
Internet of Things
+ trains, golf courses, icebergs, ATMs, pipeline networks ….
telling the right story Confidential © Bloor Research 2013
Investigative Analytics
" What happened? " Why did it happen? Is this part of a pattern that indicates that it
might happen again? " How are we going to react? If it is part of a pattern how can we
can leverage this for business purposes in the future?
telling the right story Confidential © Bloor Research 2013
Some use cases
IBM X-Force survey
telling the right story Confidential © Bloor Research 2013
Requirements
telling the right story Confidential © Bloor Research 2013
INTRODUCING
Robin Bloor
INVESTIGATIVE ANALYTICS +
THE INTERNET OF THINGS
It Begins With State…
People, objects, systems, system
components, etc.
Things can report state
RFID tags, sensors, log files, tweets, etc.
Such snippets of data are events
EVERYTHING HAS STATE
Transactional Event Based
� Corresponds to a system change
� Process heavy/data light � Analysis happens
downstream � Flows as part of a
business process � Fast
� Corresponds to a state change
� Process light/data heavy � Analysis can happen
pre-transaction � Can be a trigger in a
business process � Faster
Transactions v Events
The Technology March
The Three Latencies
Time to develop
Time to deploy
User experience
Boiling It Down
It is all about TIME TO INSIGHT – as long as that is followed by ACTION
INTRODUCING
Don DeLoach
Infobright: Investigative Analytics for The Internet of Things
Don DeLoach, CEO, Infobright [email protected]
Internet of Things
Graphic from Sensor Mania! The Internet of Things, Wearable Computing, Objective Metrics, and the Quantified Self 2.0,JSAN, Nov. 2102
Requirements for Practical Investigative Analytics
LOW TOUCH HIGH AVAILABILITY
AFFORDABILITY TCO
AD HOC PERFORMANCE SCALABILITY
COMPRESSION
LOAD SPEEDS
§ Data management § Hadoop transforming this area
§ Transparent analytic stack § Operational, investigative, predictive § Machine-generated, text
§ User consumption: § Real-time, interactive visualization & query creation
Emerging Data Analytics Stack: Days of One-Size-Fits All Are Gone
“Yesterday’s BI-‐ETL-‐EDW stack is wrong-‐sided for tomorrow’s needs, and quickly becoming irrelevant.” Gigamon
Intelligence Not Hardware: Knowledge Grid
• Stores it in the Knowledge Grid (KG) • KG is loaded into memory • Less than 1% of total compressed data size
Creates informa?on (metadata) about the
data upon load, automa?cally
• The less data that needs to be accessed, the faster the response
• Sub-‐second responses when answered by the KG
Uses the metadata when processing a query to
eliminate / reduce need to access data
• No need to par??on data, create/maintain indexes, projec?ons or tune for performance
• Ad-‐hoc queries are as fast as sta?c queries, so users have total flexibility
Architecture Benefits
Infobright Analytic Suite
Investigative Analytics for Machine-generated Data: § High performance ad-hoc query capabilities—enabling real-time information insights at
the speed of business
§ Extremely efficient (footprint, compression, data load) analytic engine designed for enterprise software deployments, OEM/embedded configurations and enterprise-ready appliance configurations proven in production
§ Install to analytics in hours: Infobright is designed for time to value
§ Integrated with the leading Hadoop, BI and ETL players
Operational Simplicity
High Performance
Efficient Form Factor
Infobright sets the bar for query performance, form
factor, and analytics business impact
AFTER BEFORE
What is needed today (and tomorrow)?
MACHINE DATA
MACHINE DATA
DATABASE ADMINISTRATORS
HARDWARE
HARDWARE
APP
LIC
ATIO
N
APP
LIC
ATIO
N
Embedded Database for M2M/Internet of Things
Low Admin: Do not want to force users to require DBAs to keep solution running
Load Speeds: Ingestion rates continue to increase, placing heavy burden on solutions
High Compression: Want to keep longer histories in less space
Lower TCO: Resulting in better value for customers, better margins for providers
Stripped Away “DBA” tax requirement required by previous versions
Ingesting over 1TB/Hour, with significant headroom beyond that
Over 3X the retention period and a 5X simultaneous reduction in storage requirement
Lower TCO for users, higher margins for JDSU
Little to No Admin
Fast Load
Speeds
20:1+ Compression
Exceptional Ad
Hoc Query Performance
Very Low TCO
REQUIREMENTS EXAMPLE: JDSU
Embedded Database for M2M/Internet of Things
Low Admin: Looking for would ensure customers have fast access to data
Load Speeds: Handle projected 70% growth rate in mobile messaging
High Compression: Need to increase data stored without increase in storage requirements
Lower TCO: Competitive flexibility of lower cost with higher value-add services
No indexes, data partitioning or manual tuning. No need for dedicated DBAs.
100,000 records per second at peak making data available for analysis within minutes Increased to 90 days of data stored in less hardware due to drastic compression
TCO only 20% of the cost of competitors. Major wireless carrier wins with this solution
Little to No Admin
Fast Load
Speeds
20:1+ Compression
Exceptional Ad
Hoc Query Performance
Very Low TCO
REQUIREMENTS EXAMPLE: MAVENIR
Embedded Database for M2M/Internet of Things
Low Admin: Do not want to force users to require DBAs to keep solution running
Fast Query Performance: Customers depend on this analysis to tune networks
High Compression and Fast Load Speeds: Need to meet business growth projections
Lower TCO: Resulting in better value for customers, better margins for providers
Low touch administration reduces friction and latency for queries
Sub-second web-based queries critical to customers to tune the network
High data compression rates and load speed allow for projected growth rate of data volume
Low OPEX = better margins and more confidence planning capacity to meet growth
Little to No Admin
Fast Load
Speeds
20:1+ Compression
Exceptional Ad
Hoc Query Performance
Very Low TCO
REQUIREMENTS EXAMPLE: POLYSTAR
Embedded Database for M2M/Internet of Things
High Compression: Projected data growth outpacing storage capacity
Ad hoc Query: Utilities want to drive customer participation in efficiency-related programs
Fast Load Speeds: Need to integrate several data streams quickly
Lower TCO: Solution needs to affordably meet business needs
No additional hardware or manual set-up in the form of data indexing or partitioning
Fast flexible reporting (20K reports in first 3 months) help utilities better drive business
Better business answers due to combined analysis of behavioral, demographic and log data
Low TCO translates to better pricing and stronger competitive positioning
Little to No Admin
Fast Load
Speeds
20:1+ Compression
Exceptional Ad
Hoc Query Performance
Very Low TCO
REQUIREMENTS EXAMPLE: OPOWER
Momentum in the M2M/Internet of Things
Applications in the Internet of Things will all require Low Touch, High Capacity and High Density; and Low Cost deployments
Smart Grids
Smart Vehicles,
Smart Cities
Mobile Health
Others..
BEFORE
MACHINE DATA
DATABASE ADMINISTRATORS
HARDWARE
APP
LIC
ATIO
N
AFTER
MACHINE DATA
HARDWARE APP
LIC
ATIO
N
Thank You
The Archive Trifecta: • Inside Analysis www.insideanalysis.com • SlideShare www.slideshare.net/InsideAnalysis • YouTube www.youtube.com/user/BloorGroup
THANK YOU!