Improve Time to Market with Real-Time Analytics on Time-Series Data

John Kain, Worldwide Capital Markets Business Development, Amazon Web Services
Domenic Ravita, Field CTA and VP of Product Marketing, MemSQL
Vin Dahake, VP Cloud Transformation, Virtusa
Improve Time to Market
with Real-Time Analytics
on Time-Series Data
© 2020, Amazon Web Services, Inc. or its Affiliates. All rights reserved.

Agenda
• Why AWS for financial services?
• Hosting time-series databases on the cloud
• MemSQL & Virtusa
• Building an operational analytics database with MemSQL
• Deploying MemSQL with Virtusa
• Customer Success Story: Global Investment Firm
• Getting Started
• Q&A

Financial Services on AWS

Why AWS for financial services?
Modernize legacy
systems for
improved agility
and scale
Drive business
growth by
harnessing data
and innovation
Meet rapidly
changing
behaviors and
expectations
Build with confidence
on a secure,
compliant, and
resilient cloud

Hosting time-series databases on AWS
• Maximizes the data that power financial services with
fast cloud performance
• Powers fundamental operations like investment
strategies, risk assessments, and fraud detections
• Highly secure, scalable, reliable, and performing
infrastructure to handle data-intensive workloads

Get the speed and scale needed for time-series data
Virtually unlimited
infrastructure
scales elastically as your
data needs evolve, with
on-demand burst
capacity for data-
intensive workloads
Faster
input/output
with a fully managed
parallel file system
Improved
performance
to handle data-intensive
operational analytics
with Amazon EC2
instances clustered
together
Optimized
cost
with flexible pay-as-you-
go pricing models

Financial services time-series data segments
Banking and payments
Perform stress test modelling
and calculate capital
requirements by running
concurrent simulations on time-
series data
Capital markets
Manipulate and analyze
time-series data for
trading and risk analytics,
forecasting, and
compliance assessments
Insurance
Scale databases to retain
long-term insurance data
and meet regulatory
obligations to policy holders

MemSQL & Virtusa

Why MemSQL for operational analytics
and time-series data?
• Scalable, relational SQL that delivers speed and high-concurrency, without sacrificing simplicity and power
• Unifies unstructured, semi-structured, and structured time-series data
• Supports for both OLTP and OLAP workloads
• Optimizes every step of the data processing lifecycle, from ingestion to storage

MemSQL – Operational databases built for speed and scale
MemSQL on AWS provides a modern data architecture to
handle operational analytics databases and time-series data
for financial services
Empowers operational analytics to power modern financial
services applications and data-intensive operational
analytics systems
MemSQL enables organizations to become real-time
businesses by delivering speed, scale, and SQL for
operational analytics and time-series data

MemSQL and Sagemaker support Operational Analytics
MemSQL becomes the de facto store for “always on” data and analytics with an SLA
Amazon S3
Amazon SageMaker
Data Intergration through
MemSQL Pipelines
Data science
and model
building
Kafka, Spark,
CDC
Operational and
frequently-accessed data
Data for cold storage
“Restful API Endpoint”
Dashboards,
applications,
operational
AI/ML

MemSQL provides the ability to handle diverse workloads

MemSQL offers distinct performance and scalability
• Distributed, highly scalable SQL database
• Includes MemSQL Studio, a visual user interface
• Ingests data continuously to perform operational analytics on billions of rows of data in
varying formats
• Handles both transactional workloads and analytical workloads
• Provides a compelling platform for building real-time applications on AWS

Why run MemSQL on AWS?
Faster
time-to-market
and reduced latency with
eliminated data duplication
across multiple database
types, saving on storage
capacity and cost
Better customer
experiences
with faster query
responses backed by
compression, storage,
and query optimization
capabilities
Improved developer
experiences
with flexible deployment
options (AWS Marketplace,
AMIs, AWS Kubernetes
Services, SaaS) and direct
support for AWS native
services

MemSQL on AWS – Features and capabilities
Fast event-to-insight
performance to
deliver against
stringent SLAs with
real-time query
processing
Built for the
enterprise to support
mission-critical
applications and
sensitive data
environments
Extensive
scalability with an
elastic scale-out
architecture
Drop-in compatibility
to plug-in directly
with existing
technologies and skills

MemSQL tools for data ingestion and operational analytics
MemSQL
Studio
Provides a visual user
interface that allows you to
easily monitor, debug, and
interact with all of your
MemSQL clusters. Available
in the MemSQL-studio
package.
MemSQL-
client
A lightweight client
application that allows you to
connect to your MemSQL
cluster and run queries.
Available in the MemSQL-
client package.
MemSQL
Replicate
A data ingestion tool for
replicating data from a source
type (such as Oracle) into
MemSQL. Filters and maps
can be used to tailor how the
data is replicated.

MemSQL on AWS
Amazon SageMaker
Virtusa Accelerate
Amazon S3
Data Store
Virtusa DQ
Engine
Amazon Kinesis
Data Firehose
AWS Database
Migration Service
External Files
Heterogenous
Databases
Streaming Data
MemSQL File Loads
MemSQL Cluster
having Aggregator
and Leaf Nodes
MemSQL S3 Pipeline
for Extract and
Optional Transform
MemSQL Replicate
(Java based)
Real Time Operational Analytics

Typical reference architecture – MemSQL on AWS
Amazon S3
Data
Store (Raw)
Amazon Kinesis Data
Firehose Process
and Load data
AWS Glue
Virtusa PySpark
Engine
Amazon EMR
AWS Lake Formation
Refined
Publish
Exploratory
Virtusa PySpark
Engine
Data Warehouse
Amazon Redshift
Cluster
Virtusa DQ
Engine
Amazon Athena
memSQL Cluster
having Aggregator
and Leaf Nodes
AWS Glue
Data Catalog
Crawler
AWS Database
Migration Service
External Files
OnPrem Databases
Streaming Data
AWS Tools
and SDKs
AWS Data Sources
MemSQL Replicate (Java based)
Real Time Operational Analytics
Amazon SageMaker
Virtusa Accelerate
S3
Pipeline
Amazon S3
Data Lake

MemSQL implementation with Virtusa
Leveraging a trusted Service Integration Partner for a
successful deployment

Virtusa – Digital Innovation Engineering
Leading provider of digital engineering services to help organizations modernize their IT
landscapes on AWS
Designed to accelerate application modernization and build new cloud-native solutions
Over 1100 AWS certifications across all tracks, uniquely position to help modernize
applications on AWS

Benefits of a MemSQL implementation with Virtusa
30% quicker implementation
time with Virtusa’s in-house
accelerators
First time right during migration
based on knowledge of
commercial databases and a
strong DB practice, 20+ years of
data transformation experience
Strong experience building
out data platforms and
data migration on AWS

MemSQL and Virtusa on AWS
for financial services
Trade latency
monitoring
Portfolio
analytics
Fraud
prevention
Risk
analytics

Transforming a global investment firm
Customer success story

About the global investment firm
Globally integrated fixed-income manager, sourcing
ideas and investment solutions worldwide
Operated on legacy databases that hindered investors’
abilities to make real-time decisions

Moving away from on-premises legacy datacenters
• Operated on a sluggish on-premises legacy datacenter
• Failed or incomplete overnight batch processing hindered their decision-making abilities
• Time-sensitive data was becoming stale as the team was unable to run reports on the
current day’s data
• Investment teams were lacking up-to-date insights needed to make accurate and effective
decisions as the trading day started
• Needed a dedicated solution to run investigative queries and self-service reports intra-daily

Implementing a scalable, relational database
Moved away from their traditional Oracle data warehouse and adopted a relational database
with MemSQL on AWS to scale and dramatically increase performance
Transactional systems fed into MemSQL, where time-series data is blended, joined, and sliced on
the fly
The solution enables real-time, time-series reporting through MemSQL Pipelines and time-series
specific functions

Becoming a real-time enterprise
• Broke down data silos that hindered their ability to accurately report
• Increased ability for investors to make faster, real-time decisions
on investments
• Greater business agility to forge a competitive edge
• Improved ability to handle data-intensive mixed workloads, including time-series data
• Moved away from being a business of managing infrastructure, to a business managing
real-time investment decisions

Please submit your questions in the chat panel
Q & A

Thank you!

Improve Time to Market with Real-Time Analytics on Time-Series Data

More Related Content

What's hot (6)

Similar to Improve Time to Market with Real-Time Analytics on Time-Series Data (17)

Recently uploaded (20)

Improve Time to Market with Real-Time Analytics on Time-Series Data