Nikita Shamgunov, CTO and Co-founder of MemSQL
Driving the On-Demand Economy
with Predictive Analytics
In-Memory
Computing
Scale-out
Imagine scaling a
database on industry
standard hardware.
Need 2x the
performance?
Add 2x the nodes.
Trying to build
scale-out for
a traditional
product
In-Memory and
Scale-out in Action
▪ Every piece of technology is scalable
▪ Analyzing data from hundreds of thousands of
machines
▪ Delivering immense value in real-time
• Real-time code deployment
• Detecting anomalies
• A/B testing results
▪ Fundamentally making the business faster by providing
data at your fingertips
An Insider’s View
▪ An enterprise solution that could scale
▪ Work well with existing tools and infrastructure
▪ Database is only as successful as the ability to
quickly and easily build applications on top of it
An Eye to Adoption
Embrace the tools and projects behind
big data and real-time transformation
Moving to Real-
Time Data
Keeping Pace
On-demand economy Real-Time Data Predictive Analytics
In our world, fresh accurate analytics means live data
▪ We’ll build a pipeline from scratch
We get predictive analytics via real-time scoring and
modeling
▪ I’ll show an example and we’ll see more across the talks
Visualizations make the data consumable
▪ Off-the-shelf options like Tableau, as well as custom
Today in My Talk
What is MemSQL?
Scalable SQL database
Familiar syntax
Really really fast
MemSQL Confidential18
Product or Services Scores
for Operational Data
Warehouse
Critical Capabilities for Data
Warehouse and Data
Management Solutions for
Analytics
Gartner, July 2016
MemSQL Pipelines
Exactly Once
Automatically Distributed
Language Agnostic
CREATE PIPELINE AS
LOAD DATA KAFKA "hostname:9092/tweets"
INTO TABLE tweets
Master Aggregator
Leaf Node Leaf Node Leaf Node Leaf Node
Child Aggregators
Master Aggregator
Leaf Node Leaf Node Leaf Node Leaf Node
Child Aggregators
.sh .sh .sh .sh
Transform
Live Demo
election.memsql.com
Real-time Twitter
Feed
Public Kafka
Load into table "tweets"
Load into table "tweet-sentiment"
MemSQL Pipelines
1. Extract 2. Transform 3. Load
Custom
dashboard
Tableau
dashboard
1. Assume data is already published somewhere in Kafka
2. Create Pipeline and point at Kafka
a. What are the schemas of the table?
b. Sentiment analysis
c. What is the connective tissue between Kafka and applying
sentiment analysis?
Run a set of commands
1. Creating tables
2. Creating pipelines
a. We will see data flowing into MemSQL
3. Build a web app similar to election.memsql.com
a. Quicker alternative: Tableau
Launch Tableau
1. Already streaming data from public Kafka into MemSQL as seen
earlier
2. Connecting Tableau
a. Similar dashboard to election.memsql.com
b. Display sentiment analysis time series
Thank You

CTO View: Driving the On-Demand Economy with Predictive Analytics