3. …but in recent times business expectations from data have changed
drastically…
3
Nightly batch
Real-time
VS
Linear Scale
Exponential Scale
VS
Structured
Polymorphic
VS
4. Traditional databases are unable to meet the new age needs of our
customers.
4
Top 3 pain points with traditional Relational Databases
CostScalability Agility
5. 5
How do we cater to these needs
without compromising on what
RDBMS is best at
6. The databases of the modern age cater to these needs
Features of Modern
Databases
Schema-free and unstructured
data formats
Flexibility to accommodate
changes and various data types
Horizontal scaling on
commodity servers
Low cost Low Complexity
Consistent Multi platform
experience Avoids platform
lock-in Aligns to Next Gen
Architecture
Denormalized data
Higher speed of retrieval
Higher performance
Built in Replication,
High Availability and
Automated Failover
No Add-on's Low
Complexity
Open Source
Low License & Storage Cost
Lower Cost
6
7. …but the choices are many
Key-Value
Databases
Caching Data
User Session and
Preferences
Shopping Cart Data
Graph
Databases
Social and other networks
Real time Routing
Fraud Detection
Document
Databases
Agile / Web App
Product Catalog
Transactional systems
Columnar
Databases
Large Data Set /Big
Data
IoT/ Sensor data
New Age
Applications
7
8. MongoDB stands out as one of the best alternatives.
8
Familiar relational syntax | Ease to add to any application | Multiple documents in 1 or many collections
Similarity to relational transactions
Multi Document ACID
guarantee with Mongo 4.0
− Snapshot isolation,
− All or nothing execution to
maintain consistency
− No performance impact for
non-transactional operations
Expressive
Query Language
Strong
Consistency
Secondary
Indexes
Flexibility
Scalability
Cost
9. …its agile and flexible architecture helps accelerate speed to value
9
Speed & Flexibility to Develop Speed to Production Speed to Insight
10. However it is important to
understand and mitigate the
migration challenges…
11. • Stakeholder alignment
• Application knowledge
with migration team
• Lack of skills on new
technologies
There are key challenges to such migrations…
PROCESS
PEOPLE TECHNOLOGY
• A radical change on how
you look at data models
• Limitation of automated
tools for migration
• Maintain equal or better
performance metrics
• Long downtime for applications • Inaccuracies in the migration
impact assessments
• Lack of documented application
knowledge
…and above all, it is extremely difficult to come out of complex contracts
12. Infosys’s robust migration methodology addresses these challenges
12
02
Benefit
Realization
Model
05
Tools and
Accelerators
03
Impact
Analysis
04
Migration
Roadmap
01
Risk
Management
06
Validation
Framework
DATA
MODERNIZATION
FRAMEWORK
Use case Suitability Analysis
Applications profiling
Risk Identification and mitigation strategy
Evaluate the current cost of ownership
Estimate the realization and proposed
cost to organization
Benefit Realization Report
Analysis on the impacts of changes within
the applications and supporting apps
Validation framework to evaluate the
success of data modernization and app
remediation
iCIA – Infosys Code Impact Analyzer - for
application remediation
NoSQL modeler – Recommends the data model
iDSS - Infosys Data Service Suite – for Data
Migration
iDTW - Infosys Data Testing Workbench – for test
& validation
TCO Calculator – To evaluate the benefit
realization model
Infosys Data Migration validator for MongoDB –
Validation of schema and data migrated
Create a migration plan and roadmap
evaluating the risk and benefit
13. …and automated through tools and accelerators
13
Analysis Tools
Infosys NoSQL
Modeler
Migration Tools
iDSS - Infosys Data
Service Suite
Testing Tools
iDTW - Infosys Data
Testing Workbench
Project Management
Accelerators
Codified Project and
Risk Management
Frameworks
20-30%
savings
On data
model design
60-70%
savings
On data
migration
30-40%
savings
On testing
30-40%
savings
On project
management
14. Infosys NoSQL Modeler – overview
14
EXTRACTION ANALYSIS PERSISTENCE PROCESSING DEPLOYMENT
Source
RDBMS
Entity Design
Query Pattern
Entity
Cardinality
Entity and
Relationship
Read and
Write Query
Entity Change
frequency and
Cardinality
Rules
Drools
SQL Lite
Target
Data-Model
Generation
Deployment
Scripts
Target
NoSQL
Process Rules
15.
16. 16
Step 1 – RDBMS TO MONGO
Step 2 – MONGO TO MONGO(Aggregation)
Infosys Data Services Suite (iDSS) overview
17.
18. We have done this for many
of our clients…
18
A fortune 100 technology
conglomerate
Order Orchestration Platform for a
leading CSP
Leading telecommunication service
provider
• Built a single consolidated view of
customers, devices and contracts
• Real time view of customer entitlements
across all categories
• Increased end customer loyalty by
providing an a unified, real time view of
all their entitlements
• Supports 15+ million transactions per
day
• Enables richer data insights on customer
behavior
• Oracle licensing and support had become
expensive
• Infosys Architect team performed the
product evaluation and comparison for the
Oracle replacement. MongoDB was
selected as target data store
• Enhanced Order orchestration platform that
can support extreme order volume during
events like new Apple IPhone launch where
order volume is 50x.
• Reduce client dependency on Oracle
licensing. Move towards sunset of Oracle
DB over next couple of years.
• Client has the requirement to store call data
records which can be easily retrieved to be
presented to police and justice department
for judicial investigation
• Current data model was rigid and
enhancements require frequently changing
the database schema
• Developed and implemented a solution to
migrate more than 1TB of data from oracle
to Mongodb using Kafka