This presentation provides an objective approach to make your legacy and custom-built applications agile and infused with intelligence. This allows your apps to utilize new and more substantial data sets as well as apply artificial intelligence and machine learning to take in-the-moment actions.
4. 4Company Confidential
DIGITAL UPSTARTS BEGIN
WITH DATA AND AI
Are you competitive or
falling behind?
Systems of Record (70’s)à Systems of Signals (90’s) à Systems of Intelligence (10’s)
5. 5Company Confidential
IN THE LAST 15 YEARS, 52% OF FORTUNE
500 COMPANIES HAVE DISAPPEARED
Westinghouse, Sears, MCI WorldCom, Toys "R" Us, Tower Records, Motorola, Sun …
6. 6
Oct 30, 2017
The Amazing Ways Spotify Uses Big Data, AI And
Machine Learning To Drive Business Success
Bernard Marr
Enterprise & Cloud
“One example is the Discover Weekly feature on Spotify that
reached 40 million people in its first year.
Every user gets a personalized playlist every week from Spotify
of music that they have not heard before on the service, but that
will be something the listener is expected to enjoy—a modern-
day version of a best friend creating a personalized mix tape.”
Upstart
7. 7
Aug 29, 2017
How Walmart Is Using Machine Learning AI, IoT
And Big Data To Boost Retail Performance
Bernard Marr Enterprise & Cloud
Rejuvenated Incumbent, making big changes
9. 9
THEY RUN INTELLIGENT APPLICATIONS
Sense and take advantage of
signals from many sources
DATA-DRIVEN
Decision-making in the moment
vs. Looking in the rear-view
mirror
IN-THE-MOMENT
Learns to predict, act, and
improve
AI / MACHINE LEARNING
Change resources dynamically
On any cloud, on-premise or
hybrid
ELASTIC AND AGILE
They extend your competitive advantage and provide meaningful impact to the business
11. 11
Your custom
applications are built on
an architecture that
could not have
anticipated the seismic
shifts
YOU EMBARKED ON DIGITAL TRANSFORMATION 1.0
But what is holding you back?
12. 12
LEGACY APPLICATIONS HAVE BEEN OVERLOOKED
Rip and Replace Rewrites Are Overkill
"Gartner predicts that every dollar invested in digital business innovation through to the end of 2020 will require
enterprises to spend at least three times that to continuously modernize the legacy application portfolio."
Gartner, 7 Options to Modernize Legacy Systems, June 1 2018
Cloud Migration Does Not Deliver Intelligence
13. 13
RIP & REPLACE YOUR DATA INFRASTRUCTURE IS
TOO DIFFICULT
OLTPOLAP ML
14. 14
SPLICE MACHINE DISTRIBUTED SQL PLATFORM
Operational
Database
• Scale-out SQL
• OLTP
• Fast
Enterprise Data
Warehouse
• In-Memory
• OLAP
• Massively Parallel
Machine
Learning
• Notebook
• Algorithms
• Model Workflow
• Deployment
INTELLIGENT
APPLICATIONS
ARTIFICIAL
INTELLIGENCE
BUSINESS
INTELLIGENCE
OPERATIONAL
INTELLIGENCE
Converged Architecture
On
Premises
15. 15Company Confidential
Every credit card payment in the world streams onto Splice
Machine in real-time, empowering groundbreaking customer
service.
• Lower fraud loss
• Vastly improved customer
service
• New system can scale out
elastically, future-proofing
the business
IMPACT
• Giant footprint with blazing
speed
• 7PB of data
• 2B records/day
• 2M queries/day < 1sec
SOLUTIONPROBLEM
• Company set out to shorten
dispute resolution from 46-
100 days to 31 days.
• Old system did not have
enough history and was too
slow to enable new call
center and API
PAYMENT LEGACY MODERNIZATION
Global
Payment
Company
16. 16Company Confidential
Migrate Global Claim, Client, & Policy Application To The Cloud and
make intelligent.
• Record Time-To-Market
• No cloud vendor-lock-in
• In-the-moment fraud, AML,
litigation
IMPACT
• Migrate business application from
legacy on-premise DB2
architecture to Distributed SQL on
multiple clouds
• Drive real-time insights with
hybrid OLAP/OLTP
• Incorporate AI/ML in real-time app
SOLUTIONPROBLEM
• Takes too long to open up a
new operating entity due to
data center implementations
• AI/ML models for fraud, AML,
litigation suffer staleness due
to data latency
• Analytical insights not fast
INSURANCE LEGACY MODERNIZATION
Leading
European
Insurance
17. 17Company Confidential
Enhance patient outcome, maximize reimbursable revenue and improve
clinical trials for its member network with AI-based tools that predict
best neurological therapies and patient conditions
• Implement patient-centric
precision medicine using data
driven approach
• Maximize reimbursable
revenues
• Select optimal clinical trial
participants
IMPACT
• Objective multidimensional
quantitative analysis based on
expansive population data across
clinics
• Effectively evaluate neurological
condition, monitor disease
progression and identify therapy
efficacy (add AI/ML)
SOLUTIONPROBLEM
• Diagnosis of neurological
condition is highly complex
• Traditional tests do not
capture critical thresholds of
disease progression
• Clinical trials can fail even if
patient outcomes improve
HEALTHCARE ADVISOR
18. 18
SPLICE MACHINE CORE DIFFERENTIATORS
Scale Out without
Sacrificing SQL
Run Existing Legacy
Applications
Make In-the-
Moment Decisions
Deploy on Any
Cloud or On-
Premise
In-Database ML
Manager
Unify Analytics
Into Your Apps
19. 19
Migrate to
Distributed SQL
MODERNIZATION 2.0
A Less Expensive and Risky Approach
Unify Analytics
Inject AI/ ML
1
2
3
RDBMS
DW
ML
Workbench
Distributed SQL
INTELLIGENT
APPLICATIONS
ARTIFICIAL
INTELLIGENCE
BUSINESS
INTELLIGENCE
OPERATIONAL
INTELLIGENCE
20. 20
RE-WRITES ARE UNTENABLE
Migrate SQL
Migrate
Reports
Create New ML Models
Deploy
Models
Rewrite Application in NoSQL Program Reports in App Create New ML Models
Deploy
Models
Months or Years
• No Changes to ANSI SQL
• Vendor SQL rewritten
• Stored Procedures
convert with tools
• BI Reports leveraged
• Tuning required
• In-DB ML makes ML avoid
ETL
• ML Manager makes it easy
to experiment and deploy
in application
• All business logic touched causing QA nightmare
• Join algorithms have to be written in application – never as
good as 40 years of RDBMS research
• ACID Transactions need to be handled in application – Rollback
hard to do
• BI Reports rewritten in application layer or additional ETL to
separated BI SQL – same as it ever was
• ETL Latency
• ML Engines need to be integrated
• Heavy ETL procedures need to be written
• ETL latency
Splice
NoSQL
Small
Team
Small
Army
22. 22
SUPPORTED BY INDUSTRY ICONS
Roger
Bamford
Mike
Franklin
Jonathan
Goldick
Abinhav
Gupta
Marie-Anne
Neimat
Andy
Pavlo
Ken
Rudin
Bruce Cleveland Bill Ericson Jay Fulcher Drew Harman Ray Lane
Advisory
Board
Board of
Directors
23. 23
Identify Applications That:
• Are slow
• Cannot scale or suffer giant DB costs
• Need new data sources
• Cannot take actions on predictions
• Need hybrid cloud strategies
• Suffer from latency of data movement
• Are batch and should be in the moment
WHERE TO START?