© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE
DIFFERENTIATORS
© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE KEY DIFFERENTIATORS
Snowflake Architecture
Near Zero Maintenance
Data Sharing
Ease of Use
Security and Governance
Business Continuity
Customer Satisfaction
Snowflake Use Cases
Snowflake and Generative AI
© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE
ARCHITECTURE
© 2022 Snowflake Inc. All Rights Reserved
Snowflake’s multi-cluster, shared data architecture is built from the ground up for the cloud, providing unparalleled performance,
simplicity, and automation.
• The Centralized Storage layer uses the cloud platform’s storage. It’s the fastest, most cost-effective, and most scalable storage
option. Snowflake automatically manages the partitioning and distribution of the data in the storage layer.
• The Multi-Cluster Compute layer represents the elastic compute clusters that process queries and updates. Snowflake
uniquely separates storage from compute so you can run multiple workloads across multiple teams without resource contention.
The compute layer allows workloads to scale up and out as needed, so workloads also leverage precisely the compute needed
when needed and release the compute resources when they’re not. Combined with per-second billing for compute cycles, this
provides an unmatched control over the cost of the cloud data warehouse.
• The Scale-Out Services layer is the brain of the entire architecture. As workloads hit Snowflake, the scale-out services layer
determines the unique requirements for getting the processing done in the most performant and cost-effective manner. The
services layer takes care of all the optimization, security, transactional consistency and metadata management automatically
with near-zero administration required.
• Cross-Cloud Snowgrid enables you to distribute your data across regions or even across cloud providers while maintaining the
same Snowflake experience regardless of complexity. This enables customers to operate a multi-cloud strategy, including a
cross-cloud approach to failover and disaster recovery, to mix and match clouds as you see fit, and share data with your partners
regardless of the cloud preferences.
SNOWFLAKE ARCHITECTURE
© 2022 Snowflake Inc. All Rights Reserved
Snowflake Platform Architecture
© 2023 Snowflake Inc. All Rights Reserved 6
structured semi-structured unstructured
On-Prem1
OBJECT
STORAGE
Cloud
Open2
OBJECT STORAGE
2 Private Preview
1 Public Preview
© 2023 Snowflake Inc. All Rights Reserved 7
INDEPENDENTLY
SCALE
WORKLOADS.
NO RESOURCE
CONTENTION.
Data Science
ETL
BI/Visualization
Dev/QA
© 2023 Snowflake Inc. All Rights Reserved 8
GLOBAL, CROSS-CLOUD
AWS
GCP
Azure
SNOWFLAKE REGIONS
SNOWGRID
© 2022 Snowflake Inc. All Rights Reserved
NEAR-ZERO
MANAGEMENT
© 2022 Snowflake Inc. All Rights Reserved
Snowflake is a true SaaS offering with near zero management. More specifically:
● There is no hardware (virtual or physical) to select, install, configure, or manage
● There is virtually no software to install, configure, or manage
● Ongoing maintenance, management, upgrades, and tuning are handled by Snowflake
● No outage windows - we use blue/green deployments to roll out new releases seamlessly with no disruption to your workloads
● Data storage is automatically optimized, meaning that you don’t have to worry about partitioning, indexing or redistributing your
data
● Storage is completely decoupled from compute, which means that the amount of data you store in Snowflake and the number of
queries you run can grow independently with virtually no limits. All you have to do is load your data and start using it
● Per-second consumption model for compute (after the first minute) and pass-through costs for data storage ensure that you only
pay for what you use, making Snowflake a highly cost-effective solution
NEAR ZERO MANAGEMENT
© 2020 Snowflake Inc. All Rights Reserved
NEAR-ZERO MANAGEMENT
NO Infrastructure
NO Tuning
NO Optimization
NO Indexing
NO Storage worries
NO Vacuuming
NO Partitioning
NO Required sorting
NO Workload mgmt.
NO Manual backups
Free time to focus on higher value tasks.
Add new datasets. Refine existing ones. Enable new teams.
© 2023 Snowflake Inc. All Rights Reserved 12
SELF-MANAGED, AUTOMATED ADMINISTRATION
OPTIMIZED STORAGE
INTEGRATION & TRANSFORMATION
Snowpipe Streams Tasks
Auto-
Clustering
Search
Optimization
Materialized
Views
Query
Acceleration1
Snowpipe
Streaming1
Dynamic
Tables2
2 Private Preview
1 Public Preview
FASTER QUERIES
© 2022 Snowflake Inc. All Rights Reserved
DATA SHARING
© 2021 Snowflake Inc. All Rights Reserved
Single live, copy of the data
Secure Data Sharing
Live data, no file copies
No ETL
Share personalized data
Share business logic
Cross-cloud and cross-region
Governed, revocable access
Differentiators
SNOWFLAKE SECURE DATA SHARING
TECHNOLOGY
© 2023 Snowflake Inc. All Rights Reserved
Live, ready-to-use data and Snowflake Native Apps* cross-cloud and cross-region. No ETL.
Governed, privacy-preserving collaboration for every scenario
Collaboration in the Data Cloud
Your Account
Share with Companies
Not Yet on Snowflake
Analyze Data without
Exposing It via Data Clean
Rooms
15
Managed
or Referral
Accounts
Discover and Monetize via
Snowflake Marketplace
Share Across Your
Business Ecosystem
*Snowflake Native Apps are available in Public Preview as of June, 2023
CATEGORY PROVIDER DATASET / DATA SERVICE
Operational & Competitive
Integrate operational & competitive datasets
Consumer Travel Data
Travel data including cruise lines, airline ticketing, rental car
spend.
Global Flight Information
Global Airline Flight Schedules from OAG, provides users
with flight schedules, updated in near real time.
Airline Reporting Travel Demand
Global daily flight data including air travel purchases,
cancellations, to/from locations, and fares.
Public Health
Incorporate consumer location and travel
patterns
COVID-19 Epidemiological Data
Analytics ready data on COVID-19 including cases &
vaccine distribution
COVID-19 Data Atlas
30+ public COVID-19 pandemic-related datasets from the
WHO, JHU, ECDC, CDC, and other authoritative sources
Weather WeatherSource
Weather and climate data worldwide
SNOWFLAKE MARKETPLACE
1,000s of curated data ready to be used
© 2022 Snowflake Inc. All Rights Reserved
EASE OF USE
© 2022 Snowflake Inc. All Rights Reserved
Ease of use is a key Snowflake value. Ease of use is important because simplicity scales and complexity does not. Below are some of
the ways in which Snowflake simplifies the tasks of acquiring, managing, and acting on the data:
● Snowflake is delivered as a service: Management and infrastructure are automatically managed by the platform so customers
can focus on getting value from the data instead of tuning infrastructure.
● Able to store any type of business data: Snowflake natively handles diverse types of data without requiring complex
transformations before loading that data into the data platform. This includes structured, semi-structured and unstructured data.
● Instantly scalable for flexible performance and concurrency: Snowflake is able to scale up, down, out and in at any time
without disruption. Snowflake also scales out to as many different use cases as needed. No need to plan for capacity ahead of
time.
● A seamless fit with existing skills and tools: Snowflake provides full support for standard SQL, leveraging the most widely
available skill within the data community. Analysts and data scientists who prefer to use R or Python can tap into the power of
Snowflake through native connectors/connectors and Snowpark (more on that later). Business Intelligence users can connect to
Snowflake from BI tools such as Tableau and Power BI and push down queries directly to Snowflake, without writing a single line
of code.
● Complex operations are simplified in Snowflake through tools such as Snowflake Streaming, Dynamic Tables, ML-Powered
Functions and much more. For example, geocoding, PII detection and masking, sentiment analysis and time series forecasting
can all be run directly in Snowflake inside a simple SQL statement.
EASE OF USE
© 2022 Snowflake Inc. All Rights Reserved
● SQL users find themselves right at home at Snowflake with full ANSI SQL support
● Snowsight is a feature-rich UI that enables data exploration, discovery and analysis
FIT WITH EXISTING SKILL SETS: SQL
ANALYZE
INTERACTIVE SQL EDITOR WITH CHARTS DASHBOARDS
DISCOVER MANAGE
© 2022 Snowflake Inc. All Rights Reserved
Snowflake + Tableau Integration
Tableau connects to Snowflake via a native ODBC connector. Push down query logic to execute on
live, up-to-date data. RBAC and governance policies are enforced no matter which tool you connect
with.
© 2022 Snowflake Inc. All Rights Reserved
Snowpipe Streaming
In Dev Private Public GA
SOON
Description
Serverless auto-ingest of streaming data
Value
Simplify architectures by ingesting streaming data
directly into Snowflake, without complexity
Functionality
Standard ingestion framework that supports rowset
ingestion. Leveraged by Snowpipe, Java client library,
Kafka connector, and open to partner ecosystem for
further development
© 2022 Snowflake Inc. All Rights Reserved
Dynamic Tables
In Dev Private Public GA
CREATE DYNAMIC TABLE
<name>
LAG = <duration>
AS <select>
Stores Results
Automatic Refreshes
Any Query!
Description
Declarative approach to
transformations and simple data
pipeline creation
Value
Automate incremental data refresh
with low latency using easy-to-use
declarative pipelines
Functionality
Join and aggregate across multiple
source objects and incrementally
update results as sources change
© 2023 Snowflake Inc. All Rights Reserved
ML-Powered Functions
In Dev Private Public GA
Description
SQL functions that abstract complexity of ML
frameworks with algorithms for Forecasting,
Anomaly Detection, and Contribution Explorer
Value
Empower analysts to enhance decision speed and
quality with easy to use, yet powerful SQL functions
and share insights in integrated analytics / BI tools
Functionality
Scale from one to millions of ML-powered insights
with the elasticity and near-zero operations of
Snowflake’s engine
© 2022 Snowflake Inc. All Rights Reserved
.
External Functions
25
External functions enable customers to tap into hundreds of APIs and services, such as document
processing, PII detection and geocoding, directly from a SQL query - in a very cost-effective manner
© 2022 Snowflake Inc. All Rights Reserved
Geography Type
● Store and analyze POINTs, LINESTRINGs,
POLYGONs
● Input/Output as GeoJSON, Well-Known Text
(WKT), and Well-Known Binary (WKB)
● Spherical model of earth
OGC-Compliant Functions
● CONTAINS, INTERSECTS, DISTANCE,
DWITHIN, and more
Performance
● Pruning and joins on geospatial predicates
Ecosystem
● Integrates with BI tools for visualization
● Works with spatial ETL tools for data integration
GEOSPATIAL SUPPORT
26
Snowflake geospatial data types and functions enable complex spatial calculations directly inside
Snowflake - no need to move the data out. Geocoding can be performed at scale via external
functions.
© 2023 Snowflake Inc. All Rights Reserved
Streamlit in Snowflake
In Dev Private Public GA
SOON
What is it?
Streamlit is an open-source Python library for app
development natively integrated into Snowflake for
scalable, reliable, and secure deployment
Value
Use Python to turn data and ML models into interactive
applications that empower stakeholders to self-serve
insights and gain trust in results
How it works
Quickly add, adjust or remove components in code
editor, visualize changes in the preview screen, and
deploy to share URL with coworkers with one click
© 2021 Snowflake Computing Inc. All Rights Reserved
CODE THE SAME WAY, EXECUTE
FASTER WITH SNOWPARK
28
SQL
SNOWFLAKE PROCESSING ENGINE
SCALA
JAVA JAVASCRIPT
PYTHON
PYTHON JAVA OTHER
EXTERNAL
SNOWPARK
CLIENT-SIDE
SERVER-SIDE
SQL
© 2022 Snowflake Inc. All Rights Reserved
SECURITY AND
GOVERNANCE
© 2022 Snowflake Inc. All Rights Reserved
Security is built into Snowflake. The platform is secure by design.
● All ingested data is automatically encrypted using AES-256 strong encryption. All communications are secured using TLS
1.2 and are controlled by network policies.
● Identity and access are managed through such enterprise features as SAML 2.0 based SSO, Oauth 2.0 delegated
authorization, and SCIM for user management. For native Snowflake credentials, supported features include password policies,
MFA, and key-pair authentication with key rotation.
● Robust role-based access control (RBAC) helps ensure that data and information can be accessed or operated on only by
authorized users within an organization. Access control is applied to all database objects including tables, schemas, secure
views, secure user-defined functions (secure UDFs), and virtual warehouses.
● Snowflake provides a comprehensive audit trail of all activities by all users.
● Governance features include dynamic data masking, row-access policies, tagging, classification, access history, and more.
● Snowflake’s security validations include SOC 1 Type II, SOC 2 Type II, HIPAA, HITRUST, PCI DDS, FedRAMP Moderate (in
specified US Government regions)
SECURITY AND GOVERNANCE
Summary of Security Features Summary of Data Governance Features
© 2023 Snowflake Inc. All Rights Reserved
Snowflake Security Product Documentation
Built in multi-factor,
integration with your
federated SSO, easy
user management
End-to-End Encryption
Always-
encrypted client
communications,
plus integration
with cloud provider
private networking
Strong Authentication
Data at rest is
always encrypted
while handled by the
Snowflake drivers
and systems
Fully Encrypted Storage
We give you options
to ensure your data
can be recovered in
case of an accident
or worse
Recovery
Full Auditing
Track every login, every
transaction, every data
transfer, and export to
your security tools
All objects, actions,
and even compute
usage can be
controlled with roles
Role-Based Access Control
Protecting Your Data in Snowflake
© 2023 Snowflake Inc. All Rights Reserved
Third party attestations and certifications
Snowflake Will Be 100% Transparent
Snowflake Security and Trust Center: https://www.snowflake.com/product/security-and-trust-center/
Snowflake Security Policy: https://www.snowflake.com/legal/ (first link)
Self-Assessments
CAIQ, SIG Lite, Pen Test Results
© 2023 Snowflake Inc. All Rights Reserved 33
UNIFIED GOVERNANCE
Tag-Based
Policies
Dat
Know Your Data Protect Your Data Connect Your Ecosystem
Access
History
Account
Usage
What
Where
Who Conditional Masking
Object Tagging
Classification
Object Dependencies External
Tokenization
Tag-Based Policies
Row Access
Policies
Dynamic Data
Masking
Direct Secure Sharing
Data Cleanrooms
Marketplace
Pre-built Partner
Integrations to
Manage Entire
Data Estate
© 2022 Snowflake Inc. All Rights Reserved
BUSINESS CONTINUITY
© 2022 Snowflake Inc. All Rights Reserved
BUSINESS CONTINUITY
The first rule of IT is that everything breaks all the time. Failures can occur due to user errors, software failures,
hardware failures and more. The following Snowflake features protect you from possible failures and ensure that
your users will continue to have access to critical data:
● User errors: Time Travel and Fail Safe enable you to access historical data (i.e., data that has been changed
or deleted.
● Infrastructure failures:
○ Storage failure: Snowflake relies on native object storage services (AWS S3, Azure Blob Storage and
Google Cloud Storage), which provide 99.999999999% of data durability and 99.99% availability.
○ Single instance failure: Snowflake automatically reprovisions new compute nodes and retries for failed
parts of the query, with virtually no latency and no intervention from you.
○ Zone failure within cloud region: Snowflake automatically fails over to one of the available zones with
virtually no latency and no intervention from you.
○ Cloud provider’s region or multi-region failure: You can configure cross-region/cross-cloud
replication and failover, including redirection of client connections. For example, you can configure
failover from Azure to AWS. Cross-cloud replication and failover is a feature unique to Snowflake.
© 2022 Snowflake Inc. All Rights Reserved
Snowflake enables business continuity
Possible failures and how Snowflake mitigates these
Failure Mitigation
Customer
Error
Snowflake Features
Time Travel
Fail-safe
Single Instance
Failure
Snowflake Built-in Redundancy
Triple-redundancy for critical services
Automatic retries for failed parts of a query
Zone
Failure
Snowflake Built-in Redundancy
Using Availability Zones on AWS, Azure, GCP
Using Availability Sets on Azure
Region
Failure
Snowflake Features
Cross-Region Replication
Cross-Region Failover
Multi-Region
Failure
Snowflake Features
Cross-Cloud Replication
Cross-Cloud Failover
© 2022 Snowflake Inc. All Rights Reserved
AWS GCP
Azure
Snowflake Regions
Maintain global business continuity
Eliminate disruptions, deliver better experiences,
and comply with changing regulations through
unique cross-cloud, cross-region connectivity.
Share data with no ETL or silos
Remove the barriers to data, regardless of cloud, region,
workload, or organizational domains. Get instant access
and distribution through a single copy of data.
Cross-cloud governance controls
Simplify governance at scale with flexible policies
that follow the data for consistent enforcement across
users and workloads.
Tap into the extended ecosystem
Enrich insights with a network of third-party data.
Discover and run new functions for extended workflows.
Snowgrid
Image not inclusive of all deployments
© 2022 Snowflake Inc. All Rights Reserved
CUSTOMER
SATISFACTION
© 2022 Snowflake Inc. All Rights Reserved
CUSTOMER SATISFACTION
Customers is Snowflake’s #1 priority in every decision the company makes. And customers love
Snowflake! Some statistics to highlight:
● In January 2022, 100% of Snowflake customer participants in the Dresner’s market
survey said they would recommend Snowflake to other organizations, for the fifth
consecutive year.
● This year, Snowflake received a Net Promoter Score of 72, a score more than three
times the industry average of 21, based on the Qualtrics 2021 NPS Industry
Benchmarking Report.
● As of July 2022, Snowflake reports 171% net revenue retention rate, which means
customers are seeing the benefits, adding the use cases and spending $0.73 more per
each dollar they spent the previous year.
https://www.snowflake.com/blog/customer-experience-report-2022/
© 2023 Snowflake Inc. All Rights Reserved 40
NET PROMOTER SCORE
Put Customers First
As of June 2022. If a customer fails to (i) respond to each required question in the survey or (ii) submit a complete set of responses by the end of the survey period, we consider that customer’s survey
incomplete. Starting with our NPS as of June 2022, we exclude incomplete survey responses from the calculation.
© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE
REFERENCES
© 2022 Snowflake Inc. All Rights Reserved
CUSTOMER STORIES
This deck includes a small selection of customer stories relevant to the SFO
use case.
Dozens of additional customer stories can be found on:
https://www.snowflake.com/en/why-snowflake/customers/
43
PROBLEM
● Drive customer loyalty by
delivering personalized
experiences
● Understand customer
preferences to apply to
offers and services
● Use data science including
ML models to inform
decisions in areas such as
customer experience and
fuel efficiency
● Reduce time spent on
system deployment and
maintenance
SOLUTION
● Unify company’s data lake
and data warehouse on the
cloud-agnostic Snowflake
Data Cloud
● Integrate DevOps, CI/CD
pipelines and ingestion
pipelines on Snowflake
● Drive administration
efficiencies with automation
RESULTS
● Enabled Emirates to deliver
data-driven offers and
passenger-focused service
● Allowed for personalizing
experiences such as inflight
entertainment
● Significantly reduced the
weeks or months previously
needed to deploy a data
analytics environment
● Enabled data team to focus
on business features
instead of deployment and
maintenance.
UNDERSTANDING CUSTOMER PREFERENCES
TO DELIVER PERSONALIZED EXPERIENCES
“We love what we can do
now because of the
integration with DevOps,
our CI/CD pipelines, and
our ingestion pipelines—
everything is more
powerful.”
Naveed Memon
Program Director, Data and
Analytics
View Full Case Study Here
44
PROBLEM
● OAG needed to make better
use of Artificial Intelligence
(AI) to help its customers
answer new questions.
● Data needed to be
accessible and visible to
customers in one place.
SOLUTION
● OAG used the Snowflake
platform to broaden its
offering to include new APIs,
alerts and a deeper layer of
data science.
● OAG has introduced OAG
Metis, an open platform
powered by Microsoft Azure
and Snowflake’s platform.
RESULTS
More powerful data
Through its partnership with
Snowflake, OAG has unlocked
the true power of its
intelligence and improved its
data, content and
solutions—both derived and
proprietary.
An incorporated view of
information
OAG Metis gives customers
access to an incorporated,
tailored view of flight
information.
ENHANCED DATA VISIBILITY @ OAG
“Our customers are now
able to directly access
and analyze the freshest
data in real-time. This
leads to more creative
use cases and helps
businesses solve harder
problems.”
PHIL CALLOW,
CEO
View Full Press Release Here
45
PROBLEM
● Provide data products and
services that empower the
travel industry to collaborate,
grow, and thrive
● Serve airlines with greater
speed and consistency so
they can be more responsive
to travelers
● Establish a scalable
architecture with the
flexibility to shift as needed
with evolving travel demands
SOLUTION
● Establish a scalable data
lake in Snowflake as a
single source of truth for all
global airline sales data
● Use a modular approach to
allow for an incremental
migration, while providing a
performant environment
● Leverage a data mart as an
intermediate step between
the data lake and delivering
customer-facing products
RESULTS
● Provided high-quality data
to help airlines, travel
agencies, and technology
providers work together to
deliver better traveler
experiences
● Gained the ability to
provision customers and
fulfill requests in minutes
● Accelerated report delivery
from hours to minutes
● Fostered a new culture of
collaboration among ARC
teams
EMPOWERING THE TRAVEL INDUSTRY TO
DELIVER BETTER CUSTOMER EXPERIENCES
“Snowflake has
accelerated our ability to
deliver value to our
customers with a
high-performance
architecture.”
Mostafa Ghazi
Solutions Architect
Click for video link
© 2022 Snowflake Inc. All Rights Reserved 47
PROBLEM
● High maintenance required
the use of a dedicated DBA
role
● Performance suffered due to
downtime issues related to
workload management
● Costly and time-consuming
to scale the existing
warehouse up or down
● Needed a data warehouse
that could handle
semi-structured JSON data
SOLUTION
● Zero maintenance, built for
the cloud data warehouse
● Quick, easy, and affordable
scalability with the ability to
scale storage and compute
separately
● Enhanced performance
requiring no workload
management
● Ability to extract data more
quickly with Tableau
● Role-based user control to
support GDPR and CCPA
compliance
RESULTS
Single Source of Truth
All teams have access to the
same data and can leverage it
for competitive business
advantages
Cost Savings
Zero maintenance and flexible
scaling reduced costs while
boosting performance
GDPR & CCPA Compliance
Industry-standard security and
privacy reduced GDPR and
CCPA compliance costs
Machine Learning Platform
Enhanced machine learning
capabilities
DRIVING ANALYTICS AND MACHINE LEARNING
AT CARRENTALS BY EXPEDIA
“We are now far better
at understanding our
business data on
Snowflake than we ever
were on our previous
cloud solution. It’s a big
difference!”
Dave Zabinski,
Senior Director, Product
Management
View Full Case Study Here
© 2022 Snowflake Inc. All Rights Reserved
OBJECTIVES
According to the National Association of State Chief Information Officers, CIOs across the U.S. have
identified data and information management and consolidation/optimization among their top 10
priorities for 2022. Chief Technology Officer for the City of Tacoma Grace Bronson aimed to
establish a city-wide data analytics program.
CHALLENGES
● Needed to combine financial, utilities, and billing data sources
● Aimed to develop user friendly self-service capabilities enabling users to create their own
reporting
● Their newly formed analytics team spent the majority of their time on data preparation
OUTCOME
● Single Source of Truth - The city now leverages a single source of data, drastically reducing
the amount of time connecting disparate sources from utilities , and eliminating out-of-date
data. This transparency builds public trust and public visualizations help citizens understand
their usage patterns over time
● Enabled Real-Time Resource and Revenue Management - Citizens have greater
transparency to utility services
48
“For the first time ever we can combine SAP
data with other sources like Esri to put
financials together with demographic and
geographic information. We were able to turn
a highly manual process into one that is fully
automated and has brought massive
efficiency to our whole ecosystem.”
GRACE BRONSON, Chief Technology Officer, City of
Tacoma
26
City departments
connected
Created 1,000 data
assets for the analytics
team to leverage
BY THE NUMBERS
Case Study: City of
Tacoma Builds Data
Analytics Program for
Financial Transparency and
Proactive Citizen Outreach
KEY RESOURCES
CITY OF TACOMA ACHIEVES FINANCIAL
ANALYTICS FOR SELF-SERVICE
© 2022 Snowflake Inc. All Rights Reserved
OBJECTIVES
According to a 2021 U.S. Department of Health Report on Improper Payments, the
improper payment rate error in the annual Department of Health & Human Services Agency
Financial Report (HHS AFR) was 6.26% of all claims which equates to $25 Billion. The
agency wanted to detect and prevent fraudulent claims.
CHALLENGES
● Needed to identify fraudulent claims at the time of submission
● Poor scale and concurrency, resulting in delayed, limited analytics
OUTCOME
● Elasticity - The federal agency and Peraton decided upon Snowflake Data Cloud on
AWS as their modern data platform. As petabytes scale, the federal agency is now able
to aggregate and report on claims processing data hours after payment from over 50
data sources.
● Machine Learning - Notifications about unusual transactions to help detect anomalies
49
“If everybody were on Snowflake, it’d unlock
the greater mission many federal agencies
have, which is for Americans to use their
services without any fear of fraud.”
“I don’t see a future where Snowflake goes
away. Jobs that once took hours with Spark
are done in less than half the time. With
Snowflake, we have effectively taken away the
technical limitation.”
AMIR DRUSBOSKY, Vice President of Technology,
Peraton
50
data sources
Ensuring claim accuracy for
> 63 million citizens
BY THE NUMBERS
Rise of the Data Cloud,
Episode 27
Blog: How Capital One
Leverages the Power of
Data in the Cloud
KEY RESOURCES
LARGE FEDERAL AGENCY
© 2022 Snowflake Inc. All Rights Reserved
OBJECTIVES
According to a 2021 U.S. Department of Health Report on Improper Payments, the
improper payment rate error in the annual Department of Health & Human Services Agency
Financial Report (HHS AFR) was 6.26% of all claims which equates to $25 Billion. The
agency wanted to detect and prevent fraudulent claims.
CHALLENGES
● Needed to identify fraudulent claims at the time of submission
● Poor scale and concurrency, resulting in delayed, limited analytics
OUTCOME
● Elasticity - The federal agency and Peraton decided upon Snowflake Data Cloud on
AWS as their modern data platform. As petabytes scale, the federal agency is now able
to aggregate and report on claims processing data hours after payment from over 50
data sources.
● Machine Learning - Notifications about unusual transactions to help detect anomalies
50
“If everybody were on Snowflake, it’d unlock
the greater mission many federal agencies
have, which is for Americans to use their
services without any fear of fraud.”
“I don’t see a future where Snowflake goes
away. Jobs that once took hours with Spark
are done in less than half the time. With
Snowflake, we have effectively taken away the
technical limitation.”
AMIR DRUSBOSKY, Vice President of Technology,
Peraton
50
data sources
Ensuring claim accuracy for
> 63 million citizens
BY THE NUMBERS
Rise of the Data Cloud,
Episode 27
Blog: How Capital One
Leverages the Power of
Data in the Cloud
KEY RESOURCES
LARGE FEDERAL AGENCY
© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE AND
GENERATIVE AI
© 2022 Snowflake Inc. All Rights Reserved
SNOWFLAKE AND GENERATIVE AI
Snowflake’s innovations and product direction with regards to Generative AI and Large
Language Models (LLM) is described here:
https://www.snowflake.com/blog/generative-ai-llms-summit-2023/
The three key themes for Snowflake include:
● The ability to run and fine-tune leading LLMs in Snowflake using Snowpark Container Services.
This new feature* drastically simplifies the management of infrastructure typically needed for this
task
● Get smarter about your data with built-in LLMs. As an example, Snowflake incorporated an LLM
into our new Document AI feature*
● Boost productivity with LLM-powered experiences, such as using natural language to query your
data**
*Currently in private preview
**In development
© 2023 Snowflake Inc. All Rights Reserved 53
Snowflake: Platform for LLMs
* Private Preview ** In Dev
SNOWFLAKE DATA
Fine-Tuning
Inference
SNOWPARK CONTAINER SERVICES *
PARTNER LLMS
OPEN SOURCE LLMS SNOWFLAKE LLMS
LLM-POWERED EXPERIENCES
Document AI *
SNOWFLAKE NATIVE APPS *
EXTERNAL
LLMS
EXTERNAL
FUNCTIONS
OR
SNOWPARK
EXTERNAL
ACCESS *
Text-to-Code **
Semantic Search **
STREAMLIT
COMMUNITY
STREAMLIT
Front-End for LLMs
© 2023 Snowflake Inc. All Rights Reserved
Document AI
Analyze and extract value from unstructured data
Description
A new interface to easily extract content from
documents via a built-in large language model
Value
Easier to extract analytical value from documents
without requiring machine learning expertise
Functionality
Ask questions in natural language, automatically
get answers, optionally fine-tune results
In Dev Private Public GA
© 2023 Snowflake Inc. All Rights Reserved
Snowpark Container Services
In Dev Private Public GA
What is it?
Additional Snowpark runtime that helps developers register
and deploy container images in Snowflake
Value
● Language & Hardware Flexibility: Build in any
programming language, package as a container
image, and deploy in configurable CPUs & GPUs
● Unified Services Experience: Effortlessly deploy with
integrated image registry, elastic compute infrastructure,
and managed Kubernetes-based cluster
● Bring Sophisticated Apps to the Data: Install entire
containerized third-party Snowflake Native Apps via
Snowflake Marketplace
SNOWFLAKE GOVERNED DATA
Compute
Image
Registry
Internally-developed Container
services/jobs/service functions
SNOWFLAKE
NATIVE APP
END-CONSUMER’S SNOWFLAKE ACCOUNT
CPU GPU
Kubernetes-based Cluster
…
SNOWPARK CONTAINER SERVICES
© 2022 Snowflake Inc. All Rights Reserved
THANK YOU

Snowflake Data Cloud Differentiators !!!

  • 1.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE DIFFERENTIATORS
  • 2.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE KEY DIFFERENTIATORS Snowflake Architecture Near Zero Maintenance Data Sharing Ease of Use Security and Governance Business Continuity Customer Satisfaction Snowflake Use Cases Snowflake and Generative AI
  • 3.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE ARCHITECTURE
  • 4.
    © 2022 SnowflakeInc. All Rights Reserved Snowflake’s multi-cluster, shared data architecture is built from the ground up for the cloud, providing unparalleled performance, simplicity, and automation. • The Centralized Storage layer uses the cloud platform’s storage. It’s the fastest, most cost-effective, and most scalable storage option. Snowflake automatically manages the partitioning and distribution of the data in the storage layer. • The Multi-Cluster Compute layer represents the elastic compute clusters that process queries and updates. Snowflake uniquely separates storage from compute so you can run multiple workloads across multiple teams without resource contention. The compute layer allows workloads to scale up and out as needed, so workloads also leverage precisely the compute needed when needed and release the compute resources when they’re not. Combined with per-second billing for compute cycles, this provides an unmatched control over the cost of the cloud data warehouse. • The Scale-Out Services layer is the brain of the entire architecture. As workloads hit Snowflake, the scale-out services layer determines the unique requirements for getting the processing done in the most performant and cost-effective manner. The services layer takes care of all the optimization, security, transactional consistency and metadata management automatically with near-zero administration required. • Cross-Cloud Snowgrid enables you to distribute your data across regions or even across cloud providers while maintaining the same Snowflake experience regardless of complexity. This enables customers to operate a multi-cloud strategy, including a cross-cloud approach to failover and disaster recovery, to mix and match clouds as you see fit, and share data with your partners regardless of the cloud preferences. SNOWFLAKE ARCHITECTURE
  • 5.
    © 2022 SnowflakeInc. All Rights Reserved Snowflake Platform Architecture
  • 6.
    © 2023 SnowflakeInc. All Rights Reserved 6 structured semi-structured unstructured On-Prem1 OBJECT STORAGE Cloud Open2 OBJECT STORAGE 2 Private Preview 1 Public Preview
  • 7.
    © 2023 SnowflakeInc. All Rights Reserved 7 INDEPENDENTLY SCALE WORKLOADS. NO RESOURCE CONTENTION. Data Science ETL BI/Visualization Dev/QA
  • 8.
    © 2023 SnowflakeInc. All Rights Reserved 8 GLOBAL, CROSS-CLOUD AWS GCP Azure SNOWFLAKE REGIONS SNOWGRID
  • 9.
    © 2022 SnowflakeInc. All Rights Reserved NEAR-ZERO MANAGEMENT
  • 10.
    © 2022 SnowflakeInc. All Rights Reserved Snowflake is a true SaaS offering with near zero management. More specifically: ● There is no hardware (virtual or physical) to select, install, configure, or manage ● There is virtually no software to install, configure, or manage ● Ongoing maintenance, management, upgrades, and tuning are handled by Snowflake ● No outage windows - we use blue/green deployments to roll out new releases seamlessly with no disruption to your workloads ● Data storage is automatically optimized, meaning that you don’t have to worry about partitioning, indexing or redistributing your data ● Storage is completely decoupled from compute, which means that the amount of data you store in Snowflake and the number of queries you run can grow independently with virtually no limits. All you have to do is load your data and start using it ● Per-second consumption model for compute (after the first minute) and pass-through costs for data storage ensure that you only pay for what you use, making Snowflake a highly cost-effective solution NEAR ZERO MANAGEMENT
  • 11.
    © 2020 SnowflakeInc. All Rights Reserved NEAR-ZERO MANAGEMENT NO Infrastructure NO Tuning NO Optimization NO Indexing NO Storage worries NO Vacuuming NO Partitioning NO Required sorting NO Workload mgmt. NO Manual backups Free time to focus on higher value tasks. Add new datasets. Refine existing ones. Enable new teams.
  • 12.
    © 2023 SnowflakeInc. All Rights Reserved 12 SELF-MANAGED, AUTOMATED ADMINISTRATION OPTIMIZED STORAGE INTEGRATION & TRANSFORMATION Snowpipe Streams Tasks Auto- Clustering Search Optimization Materialized Views Query Acceleration1 Snowpipe Streaming1 Dynamic Tables2 2 Private Preview 1 Public Preview FASTER QUERIES
  • 13.
    © 2022 SnowflakeInc. All Rights Reserved DATA SHARING
  • 14.
    © 2021 SnowflakeInc. All Rights Reserved Single live, copy of the data Secure Data Sharing Live data, no file copies No ETL Share personalized data Share business logic Cross-cloud and cross-region Governed, revocable access Differentiators SNOWFLAKE SECURE DATA SHARING TECHNOLOGY
  • 15.
    © 2023 SnowflakeInc. All Rights Reserved Live, ready-to-use data and Snowflake Native Apps* cross-cloud and cross-region. No ETL. Governed, privacy-preserving collaboration for every scenario Collaboration in the Data Cloud Your Account Share with Companies Not Yet on Snowflake Analyze Data without Exposing It via Data Clean Rooms 15 Managed or Referral Accounts Discover and Monetize via Snowflake Marketplace Share Across Your Business Ecosystem *Snowflake Native Apps are available in Public Preview as of June, 2023
  • 16.
    CATEGORY PROVIDER DATASET/ DATA SERVICE Operational & Competitive Integrate operational & competitive datasets Consumer Travel Data Travel data including cruise lines, airline ticketing, rental car spend. Global Flight Information Global Airline Flight Schedules from OAG, provides users with flight schedules, updated in near real time. Airline Reporting Travel Demand Global daily flight data including air travel purchases, cancellations, to/from locations, and fares. Public Health Incorporate consumer location and travel patterns COVID-19 Epidemiological Data Analytics ready data on COVID-19 including cases & vaccine distribution COVID-19 Data Atlas 30+ public COVID-19 pandemic-related datasets from the WHO, JHU, ECDC, CDC, and other authoritative sources Weather WeatherSource Weather and climate data worldwide SNOWFLAKE MARKETPLACE 1,000s of curated data ready to be used
  • 18.
    © 2022 SnowflakeInc. All Rights Reserved EASE OF USE
  • 19.
    © 2022 SnowflakeInc. All Rights Reserved Ease of use is a key Snowflake value. Ease of use is important because simplicity scales and complexity does not. Below are some of the ways in which Snowflake simplifies the tasks of acquiring, managing, and acting on the data: ● Snowflake is delivered as a service: Management and infrastructure are automatically managed by the platform so customers can focus on getting value from the data instead of tuning infrastructure. ● Able to store any type of business data: Snowflake natively handles diverse types of data without requiring complex transformations before loading that data into the data platform. This includes structured, semi-structured and unstructured data. ● Instantly scalable for flexible performance and concurrency: Snowflake is able to scale up, down, out and in at any time without disruption. Snowflake also scales out to as many different use cases as needed. No need to plan for capacity ahead of time. ● A seamless fit with existing skills and tools: Snowflake provides full support for standard SQL, leveraging the most widely available skill within the data community. Analysts and data scientists who prefer to use R or Python can tap into the power of Snowflake through native connectors/connectors and Snowpark (more on that later). Business Intelligence users can connect to Snowflake from BI tools such as Tableau and Power BI and push down queries directly to Snowflake, without writing a single line of code. ● Complex operations are simplified in Snowflake through tools such as Snowflake Streaming, Dynamic Tables, ML-Powered Functions and much more. For example, geocoding, PII detection and masking, sentiment analysis and time series forecasting can all be run directly in Snowflake inside a simple SQL statement. EASE OF USE
  • 20.
    © 2022 SnowflakeInc. All Rights Reserved ● SQL users find themselves right at home at Snowflake with full ANSI SQL support ● Snowsight is a feature-rich UI that enables data exploration, discovery and analysis FIT WITH EXISTING SKILL SETS: SQL ANALYZE INTERACTIVE SQL EDITOR WITH CHARTS DASHBOARDS DISCOVER MANAGE
  • 21.
    © 2022 SnowflakeInc. All Rights Reserved Snowflake + Tableau Integration Tableau connects to Snowflake via a native ODBC connector. Push down query logic to execute on live, up-to-date data. RBAC and governance policies are enforced no matter which tool you connect with.
  • 22.
    © 2022 SnowflakeInc. All Rights Reserved Snowpipe Streaming In Dev Private Public GA SOON Description Serverless auto-ingest of streaming data Value Simplify architectures by ingesting streaming data directly into Snowflake, without complexity Functionality Standard ingestion framework that supports rowset ingestion. Leveraged by Snowpipe, Java client library, Kafka connector, and open to partner ecosystem for further development
  • 23.
    © 2022 SnowflakeInc. All Rights Reserved Dynamic Tables In Dev Private Public GA CREATE DYNAMIC TABLE <name> LAG = <duration> AS <select> Stores Results Automatic Refreshes Any Query! Description Declarative approach to transformations and simple data pipeline creation Value Automate incremental data refresh with low latency using easy-to-use declarative pipelines Functionality Join and aggregate across multiple source objects and incrementally update results as sources change
  • 24.
    © 2023 SnowflakeInc. All Rights Reserved ML-Powered Functions In Dev Private Public GA Description SQL functions that abstract complexity of ML frameworks with algorithms for Forecasting, Anomaly Detection, and Contribution Explorer Value Empower analysts to enhance decision speed and quality with easy to use, yet powerful SQL functions and share insights in integrated analytics / BI tools Functionality Scale from one to millions of ML-powered insights with the elasticity and near-zero operations of Snowflake’s engine
  • 25.
    © 2022 SnowflakeInc. All Rights Reserved . External Functions 25 External functions enable customers to tap into hundreds of APIs and services, such as document processing, PII detection and geocoding, directly from a SQL query - in a very cost-effective manner
  • 26.
    © 2022 SnowflakeInc. All Rights Reserved Geography Type ● Store and analyze POINTs, LINESTRINGs, POLYGONs ● Input/Output as GeoJSON, Well-Known Text (WKT), and Well-Known Binary (WKB) ● Spherical model of earth OGC-Compliant Functions ● CONTAINS, INTERSECTS, DISTANCE, DWITHIN, and more Performance ● Pruning and joins on geospatial predicates Ecosystem ● Integrates with BI tools for visualization ● Works with spatial ETL tools for data integration GEOSPATIAL SUPPORT 26 Snowflake geospatial data types and functions enable complex spatial calculations directly inside Snowflake - no need to move the data out. Geocoding can be performed at scale via external functions.
  • 27.
    © 2023 SnowflakeInc. All Rights Reserved Streamlit in Snowflake In Dev Private Public GA SOON What is it? Streamlit is an open-source Python library for app development natively integrated into Snowflake for scalable, reliable, and secure deployment Value Use Python to turn data and ML models into interactive applications that empower stakeholders to self-serve insights and gain trust in results How it works Quickly add, adjust or remove components in code editor, visualize changes in the preview screen, and deploy to share URL with coworkers with one click
  • 28.
    © 2021 SnowflakeComputing Inc. All Rights Reserved CODE THE SAME WAY, EXECUTE FASTER WITH SNOWPARK 28 SQL SNOWFLAKE PROCESSING ENGINE SCALA JAVA JAVASCRIPT PYTHON PYTHON JAVA OTHER EXTERNAL SNOWPARK CLIENT-SIDE SERVER-SIDE SQL
  • 29.
    © 2022 SnowflakeInc. All Rights Reserved SECURITY AND GOVERNANCE
  • 30.
    © 2022 SnowflakeInc. All Rights Reserved Security is built into Snowflake. The platform is secure by design. ● All ingested data is automatically encrypted using AES-256 strong encryption. All communications are secured using TLS 1.2 and are controlled by network policies. ● Identity and access are managed through such enterprise features as SAML 2.0 based SSO, Oauth 2.0 delegated authorization, and SCIM for user management. For native Snowflake credentials, supported features include password policies, MFA, and key-pair authentication with key rotation. ● Robust role-based access control (RBAC) helps ensure that data and information can be accessed or operated on only by authorized users within an organization. Access control is applied to all database objects including tables, schemas, secure views, secure user-defined functions (secure UDFs), and virtual warehouses. ● Snowflake provides a comprehensive audit trail of all activities by all users. ● Governance features include dynamic data masking, row-access policies, tagging, classification, access history, and more. ● Snowflake’s security validations include SOC 1 Type II, SOC 2 Type II, HIPAA, HITRUST, PCI DDS, FedRAMP Moderate (in specified US Government regions) SECURITY AND GOVERNANCE Summary of Security Features Summary of Data Governance Features
  • 31.
    © 2023 SnowflakeInc. All Rights Reserved Snowflake Security Product Documentation Built in multi-factor, integration with your federated SSO, easy user management End-to-End Encryption Always- encrypted client communications, plus integration with cloud provider private networking Strong Authentication Data at rest is always encrypted while handled by the Snowflake drivers and systems Fully Encrypted Storage We give you options to ensure your data can be recovered in case of an accident or worse Recovery Full Auditing Track every login, every transaction, every data transfer, and export to your security tools All objects, actions, and even compute usage can be controlled with roles Role-Based Access Control Protecting Your Data in Snowflake
  • 32.
    © 2023 SnowflakeInc. All Rights Reserved Third party attestations and certifications Snowflake Will Be 100% Transparent Snowflake Security and Trust Center: https://www.snowflake.com/product/security-and-trust-center/ Snowflake Security Policy: https://www.snowflake.com/legal/ (first link) Self-Assessments CAIQ, SIG Lite, Pen Test Results
  • 33.
    © 2023 SnowflakeInc. All Rights Reserved 33 UNIFIED GOVERNANCE Tag-Based Policies Dat Know Your Data Protect Your Data Connect Your Ecosystem Access History Account Usage What Where Who Conditional Masking Object Tagging Classification Object Dependencies External Tokenization Tag-Based Policies Row Access Policies Dynamic Data Masking Direct Secure Sharing Data Cleanrooms Marketplace Pre-built Partner Integrations to Manage Entire Data Estate
  • 34.
    © 2022 SnowflakeInc. All Rights Reserved BUSINESS CONTINUITY
  • 35.
    © 2022 SnowflakeInc. All Rights Reserved BUSINESS CONTINUITY The first rule of IT is that everything breaks all the time. Failures can occur due to user errors, software failures, hardware failures and more. The following Snowflake features protect you from possible failures and ensure that your users will continue to have access to critical data: ● User errors: Time Travel and Fail Safe enable you to access historical data (i.e., data that has been changed or deleted. ● Infrastructure failures: ○ Storage failure: Snowflake relies on native object storage services (AWS S3, Azure Blob Storage and Google Cloud Storage), which provide 99.999999999% of data durability and 99.99% availability. ○ Single instance failure: Snowflake automatically reprovisions new compute nodes and retries for failed parts of the query, with virtually no latency and no intervention from you. ○ Zone failure within cloud region: Snowflake automatically fails over to one of the available zones with virtually no latency and no intervention from you. ○ Cloud provider’s region or multi-region failure: You can configure cross-region/cross-cloud replication and failover, including redirection of client connections. For example, you can configure failover from Azure to AWS. Cross-cloud replication and failover is a feature unique to Snowflake.
  • 36.
    © 2022 SnowflakeInc. All Rights Reserved Snowflake enables business continuity Possible failures and how Snowflake mitigates these Failure Mitigation Customer Error Snowflake Features Time Travel Fail-safe Single Instance Failure Snowflake Built-in Redundancy Triple-redundancy for critical services Automatic retries for failed parts of a query Zone Failure Snowflake Built-in Redundancy Using Availability Zones on AWS, Azure, GCP Using Availability Sets on Azure Region Failure Snowflake Features Cross-Region Replication Cross-Region Failover Multi-Region Failure Snowflake Features Cross-Cloud Replication Cross-Cloud Failover
  • 37.
    © 2022 SnowflakeInc. All Rights Reserved AWS GCP Azure Snowflake Regions Maintain global business continuity Eliminate disruptions, deliver better experiences, and comply with changing regulations through unique cross-cloud, cross-region connectivity. Share data with no ETL or silos Remove the barriers to data, regardless of cloud, region, workload, or organizational domains. Get instant access and distribution through a single copy of data. Cross-cloud governance controls Simplify governance at scale with flexible policies that follow the data for consistent enforcement across users and workloads. Tap into the extended ecosystem Enrich insights with a network of third-party data. Discover and run new functions for extended workflows. Snowgrid Image not inclusive of all deployments
  • 38.
    © 2022 SnowflakeInc. All Rights Reserved CUSTOMER SATISFACTION
  • 39.
    © 2022 SnowflakeInc. All Rights Reserved CUSTOMER SATISFACTION Customers is Snowflake’s #1 priority in every decision the company makes. And customers love Snowflake! Some statistics to highlight: ● In January 2022, 100% of Snowflake customer participants in the Dresner’s market survey said they would recommend Snowflake to other organizations, for the fifth consecutive year. ● This year, Snowflake received a Net Promoter Score of 72, a score more than three times the industry average of 21, based on the Qualtrics 2021 NPS Industry Benchmarking Report. ● As of July 2022, Snowflake reports 171% net revenue retention rate, which means customers are seeing the benefits, adding the use cases and spending $0.73 more per each dollar they spent the previous year. https://www.snowflake.com/blog/customer-experience-report-2022/
  • 40.
    © 2023 SnowflakeInc. All Rights Reserved 40 NET PROMOTER SCORE Put Customers First As of June 2022. If a customer fails to (i) respond to each required question in the survey or (ii) submit a complete set of responses by the end of the survey period, we consider that customer’s survey incomplete. Starting with our NPS as of June 2022, we exclude incomplete survey responses from the calculation.
  • 41.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE REFERENCES
  • 42.
    © 2022 SnowflakeInc. All Rights Reserved CUSTOMER STORIES This deck includes a small selection of customer stories relevant to the SFO use case. Dozens of additional customer stories can be found on: https://www.snowflake.com/en/why-snowflake/customers/
  • 43.
    43 PROBLEM ● Drive customerloyalty by delivering personalized experiences ● Understand customer preferences to apply to offers and services ● Use data science including ML models to inform decisions in areas such as customer experience and fuel efficiency ● Reduce time spent on system deployment and maintenance SOLUTION ● Unify company’s data lake and data warehouse on the cloud-agnostic Snowflake Data Cloud ● Integrate DevOps, CI/CD pipelines and ingestion pipelines on Snowflake ● Drive administration efficiencies with automation RESULTS ● Enabled Emirates to deliver data-driven offers and passenger-focused service ● Allowed for personalizing experiences such as inflight entertainment ● Significantly reduced the weeks or months previously needed to deploy a data analytics environment ● Enabled data team to focus on business features instead of deployment and maintenance. UNDERSTANDING CUSTOMER PREFERENCES TO DELIVER PERSONALIZED EXPERIENCES “We love what we can do now because of the integration with DevOps, our CI/CD pipelines, and our ingestion pipelines— everything is more powerful.” Naveed Memon Program Director, Data and Analytics View Full Case Study Here
  • 44.
    44 PROBLEM ● OAG neededto make better use of Artificial Intelligence (AI) to help its customers answer new questions. ● Data needed to be accessible and visible to customers in one place. SOLUTION ● OAG used the Snowflake platform to broaden its offering to include new APIs, alerts and a deeper layer of data science. ● OAG has introduced OAG Metis, an open platform powered by Microsoft Azure and Snowflake’s platform. RESULTS More powerful data Through its partnership with Snowflake, OAG has unlocked the true power of its intelligence and improved its data, content and solutions—both derived and proprietary. An incorporated view of information OAG Metis gives customers access to an incorporated, tailored view of flight information. ENHANCED DATA VISIBILITY @ OAG “Our customers are now able to directly access and analyze the freshest data in real-time. This leads to more creative use cases and helps businesses solve harder problems.” PHIL CALLOW, CEO View Full Press Release Here
  • 45.
    45 PROBLEM ● Provide dataproducts and services that empower the travel industry to collaborate, grow, and thrive ● Serve airlines with greater speed and consistency so they can be more responsive to travelers ● Establish a scalable architecture with the flexibility to shift as needed with evolving travel demands SOLUTION ● Establish a scalable data lake in Snowflake as a single source of truth for all global airline sales data ● Use a modular approach to allow for an incremental migration, while providing a performant environment ● Leverage a data mart as an intermediate step between the data lake and delivering customer-facing products RESULTS ● Provided high-quality data to help airlines, travel agencies, and technology providers work together to deliver better traveler experiences ● Gained the ability to provision customers and fulfill requests in minutes ● Accelerated report delivery from hours to minutes ● Fostered a new culture of collaboration among ARC teams EMPOWERING THE TRAVEL INDUSTRY TO DELIVER BETTER CUSTOMER EXPERIENCES “Snowflake has accelerated our ability to deliver value to our customers with a high-performance architecture.” Mostafa Ghazi Solutions Architect
  • 46.
  • 47.
    © 2022 SnowflakeInc. All Rights Reserved 47 PROBLEM ● High maintenance required the use of a dedicated DBA role ● Performance suffered due to downtime issues related to workload management ● Costly and time-consuming to scale the existing warehouse up or down ● Needed a data warehouse that could handle semi-structured JSON data SOLUTION ● Zero maintenance, built for the cloud data warehouse ● Quick, easy, and affordable scalability with the ability to scale storage and compute separately ● Enhanced performance requiring no workload management ● Ability to extract data more quickly with Tableau ● Role-based user control to support GDPR and CCPA compliance RESULTS Single Source of Truth All teams have access to the same data and can leverage it for competitive business advantages Cost Savings Zero maintenance and flexible scaling reduced costs while boosting performance GDPR & CCPA Compliance Industry-standard security and privacy reduced GDPR and CCPA compliance costs Machine Learning Platform Enhanced machine learning capabilities DRIVING ANALYTICS AND MACHINE LEARNING AT CARRENTALS BY EXPEDIA “We are now far better at understanding our business data on Snowflake than we ever were on our previous cloud solution. It’s a big difference!” Dave Zabinski, Senior Director, Product Management View Full Case Study Here
  • 48.
    © 2022 SnowflakeInc. All Rights Reserved OBJECTIVES According to the National Association of State Chief Information Officers, CIOs across the U.S. have identified data and information management and consolidation/optimization among their top 10 priorities for 2022. Chief Technology Officer for the City of Tacoma Grace Bronson aimed to establish a city-wide data analytics program. CHALLENGES ● Needed to combine financial, utilities, and billing data sources ● Aimed to develop user friendly self-service capabilities enabling users to create their own reporting ● Their newly formed analytics team spent the majority of their time on data preparation OUTCOME ● Single Source of Truth - The city now leverages a single source of data, drastically reducing the amount of time connecting disparate sources from utilities , and eliminating out-of-date data. This transparency builds public trust and public visualizations help citizens understand their usage patterns over time ● Enabled Real-Time Resource and Revenue Management - Citizens have greater transparency to utility services 48 “For the first time ever we can combine SAP data with other sources like Esri to put financials together with demographic and geographic information. We were able to turn a highly manual process into one that is fully automated and has brought massive efficiency to our whole ecosystem.” GRACE BRONSON, Chief Technology Officer, City of Tacoma 26 City departments connected Created 1,000 data assets for the analytics team to leverage BY THE NUMBERS Case Study: City of Tacoma Builds Data Analytics Program for Financial Transparency and Proactive Citizen Outreach KEY RESOURCES CITY OF TACOMA ACHIEVES FINANCIAL ANALYTICS FOR SELF-SERVICE
  • 49.
    © 2022 SnowflakeInc. All Rights Reserved OBJECTIVES According to a 2021 U.S. Department of Health Report on Improper Payments, the improper payment rate error in the annual Department of Health & Human Services Agency Financial Report (HHS AFR) was 6.26% of all claims which equates to $25 Billion. The agency wanted to detect and prevent fraudulent claims. CHALLENGES ● Needed to identify fraudulent claims at the time of submission ● Poor scale and concurrency, resulting in delayed, limited analytics OUTCOME ● Elasticity - The federal agency and Peraton decided upon Snowflake Data Cloud on AWS as their modern data platform. As petabytes scale, the federal agency is now able to aggregate and report on claims processing data hours after payment from over 50 data sources. ● Machine Learning - Notifications about unusual transactions to help detect anomalies 49 “If everybody were on Snowflake, it’d unlock the greater mission many federal agencies have, which is for Americans to use their services without any fear of fraud.” “I don’t see a future where Snowflake goes away. Jobs that once took hours with Spark are done in less than half the time. With Snowflake, we have effectively taken away the technical limitation.” AMIR DRUSBOSKY, Vice President of Technology, Peraton 50 data sources Ensuring claim accuracy for > 63 million citizens BY THE NUMBERS Rise of the Data Cloud, Episode 27 Blog: How Capital One Leverages the Power of Data in the Cloud KEY RESOURCES LARGE FEDERAL AGENCY
  • 50.
    © 2022 SnowflakeInc. All Rights Reserved OBJECTIVES According to a 2021 U.S. Department of Health Report on Improper Payments, the improper payment rate error in the annual Department of Health & Human Services Agency Financial Report (HHS AFR) was 6.26% of all claims which equates to $25 Billion. The agency wanted to detect and prevent fraudulent claims. CHALLENGES ● Needed to identify fraudulent claims at the time of submission ● Poor scale and concurrency, resulting in delayed, limited analytics OUTCOME ● Elasticity - The federal agency and Peraton decided upon Snowflake Data Cloud on AWS as their modern data platform. As petabytes scale, the federal agency is now able to aggregate and report on claims processing data hours after payment from over 50 data sources. ● Machine Learning - Notifications about unusual transactions to help detect anomalies 50 “If everybody were on Snowflake, it’d unlock the greater mission many federal agencies have, which is for Americans to use their services without any fear of fraud.” “I don’t see a future where Snowflake goes away. Jobs that once took hours with Spark are done in less than half the time. With Snowflake, we have effectively taken away the technical limitation.” AMIR DRUSBOSKY, Vice President of Technology, Peraton 50 data sources Ensuring claim accuracy for > 63 million citizens BY THE NUMBERS Rise of the Data Cloud, Episode 27 Blog: How Capital One Leverages the Power of Data in the Cloud KEY RESOURCES LARGE FEDERAL AGENCY
  • 51.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE AND GENERATIVE AI
  • 52.
    © 2022 SnowflakeInc. All Rights Reserved SNOWFLAKE AND GENERATIVE AI Snowflake’s innovations and product direction with regards to Generative AI and Large Language Models (LLM) is described here: https://www.snowflake.com/blog/generative-ai-llms-summit-2023/ The three key themes for Snowflake include: ● The ability to run and fine-tune leading LLMs in Snowflake using Snowpark Container Services. This new feature* drastically simplifies the management of infrastructure typically needed for this task ● Get smarter about your data with built-in LLMs. As an example, Snowflake incorporated an LLM into our new Document AI feature* ● Boost productivity with LLM-powered experiences, such as using natural language to query your data** *Currently in private preview **In development
  • 53.
    © 2023 SnowflakeInc. All Rights Reserved 53 Snowflake: Platform for LLMs * Private Preview ** In Dev SNOWFLAKE DATA Fine-Tuning Inference SNOWPARK CONTAINER SERVICES * PARTNER LLMS OPEN SOURCE LLMS SNOWFLAKE LLMS LLM-POWERED EXPERIENCES Document AI * SNOWFLAKE NATIVE APPS * EXTERNAL LLMS EXTERNAL FUNCTIONS OR SNOWPARK EXTERNAL ACCESS * Text-to-Code ** Semantic Search ** STREAMLIT COMMUNITY STREAMLIT Front-End for LLMs
  • 54.
    © 2023 SnowflakeInc. All Rights Reserved Document AI Analyze and extract value from unstructured data Description A new interface to easily extract content from documents via a built-in large language model Value Easier to extract analytical value from documents without requiring machine learning expertise Functionality Ask questions in natural language, automatically get answers, optionally fine-tune results In Dev Private Public GA
  • 55.
    © 2023 SnowflakeInc. All Rights Reserved Snowpark Container Services In Dev Private Public GA What is it? Additional Snowpark runtime that helps developers register and deploy container images in Snowflake Value ● Language & Hardware Flexibility: Build in any programming language, package as a container image, and deploy in configurable CPUs & GPUs ● Unified Services Experience: Effortlessly deploy with integrated image registry, elastic compute infrastructure, and managed Kubernetes-based cluster ● Bring Sophisticated Apps to the Data: Install entire containerized third-party Snowflake Native Apps via Snowflake Marketplace SNOWFLAKE GOVERNED DATA Compute Image Registry Internally-developed Container services/jobs/service functions SNOWFLAKE NATIVE APP END-CONSUMER’S SNOWFLAKE ACCOUNT CPU GPU Kubernetes-based Cluster … SNOWPARK CONTAINER SERVICES
  • 56.
    © 2022 SnowflakeInc. All Rights Reserved THANK YOU