Cloudera Altus: Big Data in the Cloud Made Easy

1© Cloudera, Inc. All rights reserved.
ClouderaAltus: Big Data in the Cloud Made
Easy
David Tishgart | Product Marketing
Jennifer Wu | Product Management

We believe
data can make what is impossible
today, possible tomorrow

We empower
people to transform complex data
into clear and actionable insights
DRIVE
CUSTOMER INSIGHTS
CONNECT
PRODUCTS & SERVICES (IoT)
PROTECT
BUSINESS

We deliver
the modern platform for
machine learning and advanced analytics
RUNS ANYWHERE
Cloud
Multi-cloud
On-prem
SCALABLE
Elastic
Cost-effective
Lower TCO
ENTERPRISE GRADE
Secure
Performant
Compliant

The data-driven enterprise
IoT explosion of new data
30B
connected
devices
440x
more data
Enterprises re-architect to
modernize IT infrastructure
open source
cloud
machine
learning

<1%
of an organization’s
unstructured data is
analyzed or
used at all
<50%
of an organization’s
structured data is
actively used in
making decisions
80%
of analysts’ time
is spent simply
discovering and
preparing data

• Move processing to the cloud without risk
• Focus on your workload, not cluster
operations
• Simplify and unify your analytics

A platform for enabling data-driven decisions
Modern data processing
(ETL) and data governance at
scale
Data
Engineering
Explore, analyze, and
understand all your data
Analytic
Database
Data-driven applications to
deliver real-time insights
Operational
Database
Multi-Storage,
Multi-Environment
Exploratory data science and
machine learning for the
enterprise
Data Science

Data Engineering use cases

Ingest, Process, and Deliver Insights to Drive Decisions
Leverage any source
and format
• Unstructured data
• Structured data
• Social data
• Machine data
• IOT data
Large scale data
processing
• Stream or batch
• Choice of engine:
MapReduce, Spark,
Hive, Hive-on-Spark
• Job SLAs
• Data storage
Analyze, build and
train models
• BI analytic engines
• Data science and
BI tools/libraries
• Real-time analysis
• Report generation
Ingest from data
sources
Process and
transform data
Deliver insights
Make business
decisions
Using data to grow
the business
• Report
consumption
• Apply judgement
• Inform business
decisions

Example: Data Engineering in ML Pipeline
clean, merge, filter
Data
Engineering
Raw Data
● formats
● sources
● volume
Processed
data
● training
● validation
● test model, train, tune
Data Science
processing, execution
Data
Engineering
Validated model
● model
● parameters
Live data
ingest
End results
● insights
● predictions
● results

Goals of an Organization: Data Engineers and IT
Agility
● Time to market for new use
cases
● Reliably meet SLAs for
processed data
Ease-of-use
● Production: run
operationalized workflows;
monitor and troubleshoot
● Development: interactive
workflows and tools
Goals of Data
Engineers
Goals of IT
Administrators
● Cost
● Security
● Standardization
● Self-service LOBs

IT and DE goals
- Agility
- Ease of use
- Cost effectiveness
- Standardization
Public Cloud Properties
- on-demand infrastructure resourcing
- hyperscale storage
- data durable + highly available
Bridging the Gap Between Users and Cloud
Cloudera Altus
providing a bridge since 2017

Cloudera Altus for data engineering workloads
PaaS for ETL, machine learning, and data
processing on AWS
● Managed transient clusters in customer VPC
● Support for MR2, Hive, Spark, Hive-on-Spark
● Workload analytics and troubleshooting

Easy

https://console.altus.cloudera.com

Everything you don’t have to do
• Install any software to start working
• Install any hardware
• Worry about cluster configuration
• Upgrade/reconfigure clusters
• OS upgrades/patching
• Resource Management

Focus on Workloads
• Jobs/Workloads as first class entities
• Clusters as supporting entities
• Abstracts away most infrastructure and cluster details

Workload troubleshooting and analytics
● Troubleshoot jobs after cluster termination
through job log and configuration browsing
● Insight into causes of job failure
● Identification and root cause analysis of slow
jobs

Agile

• The same CDH is present regardless of deployment model
• Simplified application migration
• Minimizes cloud migration risk
• Core components open-source
• Reduces risk for lock-in
• Reduces unknowns for third-party partners
One Platform Everywhere

Embrace Transience for
Lower Costs
Compartmentalize for
Optimal Performance
Hyperscale Cloud Store
Grow Storage and Compute
Discretely for Efficiency
STORE
COMPUTE
Cloud-native architecture

Unified

• Specialized clusters provide optimized
computation
• Data and Metadata need to be
consistently accessible across all clusters
• Whether CDH clusters are managed by
Altus or not
No Data Silos
Object Store

• S3 eventual consistency model can lead to errors
• S3Guard provides a consistent view of S3 data across clusters
• Table schemas are often relevant across clusters
• For simple scenarios, inline DDL in jobs is sufficient
• For complex scenarios, share a persistent backing database for the metadata
store
Multi-cluster data and metadata consistency

Altus feature overview
Low cost
• Per-node/per-hour pricing
• Terminate clusters when not in
use
• Spot with self-healing
End-user focused
• Manages your cluster so you don’t
have to
• Job submission CLI/API
• Workload troubleshooting
Easy to use
• Self-service for end-users
• Cloud console + familiar tools
• Cluster provisioning in minutes
Cloud-native
• Decouple storage and compute
• R/W to/from Amazon S3
• Spin EC2 clusters up and down
Integrated Platform
• Same Cloudera platform on-
premises and in the cloud
• Feed cleaned data into Impala
clusters for BI analytics
• Share metadata across clusters
Secure
• Integrated with AWS security
• No Cloudera access to customer
data
• Support for multi-AWS accounts
• Admin and end-user accounts

Altus Demo

Demo “Analysis: Hospital Death Rates vs. State GDP”
Data
Engineering
Data Science
Well-formatted
data
Hospital data
Death Rates
Complications
cleaning
merging
filtering
Linear
Regression
Statistical
AnalysisPer-capita State
GDP
[State, Condition,
Hospital, Type,
Average Score,
GDP]
Insights
Data Science Workbench

BUILT FOR THE
ENTERPRISE
SIMPLIFIED &
OPTIMIZED
OPEN
ARCHITECTURE
Your choice for the cloud
● Portable for hybrid
and multi-cloud
● Rich ISV partner
ecosystem
● Open source,
supported by
community
● Low TCO on cloud-
native infrastructure
● Deployed as-a-service
with minimal cluster
management
UNIFIED
PLATFORM
● Minimize data silos
● Simplify operations
● Improve ease of
development and
cloud migration
● Security
● Availability
● Performance
● Governance
CLOUDERA
● Disjointed services
create data silos
● Confusing and
disjoined services,
data, and pricing
models
● Vendor lock in
OTHER CLOUD
DATA
PLATFORMS
● No common
metadata, security,
lineage across apps

Visit
www.cloudera.com/altus
for access, and ask about
our free trial

Questions?

Thank you
David Tishgart
Jennifer Wu

Cloudera Altus: Big Data in the Cloud Made Easy

Recommended

More Related Content

What's hot (20)

Similar to Cloudera Altus: Big Data in the Cloud Made Easy (20)

More from Cloudera, Inc. (20)

Recently uploaded (20)

Cloudera Altus: Big Data in the Cloud Made Easy