migrating paraccel/matrix workloads to the...
TRANSCRIPT
Migrating ParAccel/Matrix
Workloads to the Cloud
Topics
• Introductions
• The Situation
• Cloud Data Warehousing: Considerations and Case Studies
• Data Warehousing as a Service
• Q&A
About Cazena
Enterprise Heritage
Founded by former
Netezza leaders
Created enterprise data
warehousing appliance
category, 800+ customers
Started Cazena in 2013,
launched 2015
Strong Backing Founding Vision
Big Data On Demand
About MagnusData
Enterprise
Big Data Experts
Served Customers in
Wealth Management,
Digital Marketing, and
Social Media
MIT
Mission Vision
Big Data Cloud Services
Data That Creates Value
-
We Believe in Making
Technology an Asset
For
Business Partners
Speakers
Paul Wolmering Cazena Solution Architecture Lead
Lokesh Khosla Principal and Partner, Magnus Data
The Situation For ParAccel/Matrix Shops
An unexpected, unbudgeted challenge needing fast resolution
Actian Support Ends March 30
What do Organizations Need to Do?
• Equal or better performance
• Minimal impact on infrastructure
• No impact on users
• Cost-effective
• Safe and supported
Required: move to a platform that minimizes cost and risk
Possible: also lay a foundation for evolution and innovation
What Options Exist?
Pros Cons
Stay on ParAccel/Matrix • Cost-effective
• Leverage existing skills
• No disruption
• Risk and lack of sleep
• Lack of effective support
• Lack of investment
Move to new EDW on-
prem
• Leverage existing skills
• Too slow
• Expensive
New Hadoop • Cost-effective for some
workloads
• Not be suitable for all
workloads
• Not something to be rushed
Migrate DW to Cloud • Cost-effective
• Speed
• Flexibility
• New skills, development
required, or a partner
• Security and other concerns
On-Premise vs. Cloud Based Infrastructure
On-Premise Infrastructure Cloud Based Infrastructure
• Costly
• Hinders innovation
• Lacks support for
business growth
• Requires backups
• Business contingency
plans in case of
disasters
• Speed
• Agility
• Elasticity
• Focus
• Lower TCO
• Secure
• Shared responsibility
• Global access
• Innovation friendly
HPC for Big Data on AWS
• Customer challenge:
– Unable to scale use of AWS tools to support business growth
• Magnus solution:
– Helped with architecting an HPC Big Data system using AWS tools
• Customer benefits:
– Faster than the 5 second SLA for analyst queries
– Aligned with IT governance standards
– Improved security
Netezza Migration to AWS
• Customer challenge:
– Existing system lacked flexibility
– Incurring huge cost
• Magnus solution:
– Moved Netezza workloads to Hadoop on AWS
• Customer benefits:
– Reduced costs 80%
– Exceeded performance standards
– Increased flexibility
Marketing Analytics Company
Data Warehousing as a Service
Requirements for Data Warehousing in the Cloud
Security and Compliance Implement cloud security controls, compliance, governance
Production Operations, Monitoring & Support Develop processes for monitoring, maintenance, upgrades, support
Data Movement Deploy tools and process to securely move data to/from the cloud
Analytic Platform and Infrastructure Provision and deploy multi-cloud multi-database support
Enterprise Integration Securely connect cloud to existing enterprise tools, systems
Production development and operations requirements
Workload Engine
Benchmarking
Cloud Infrastructure
Workload
Engines
Security, Compliance
SLA Optimization,
Operations
3
Cazena
Gateway Data
Movers
Cloud Sockets
How Cazena Works
Data
Sources,
ETL & Warehouses
BI /
Analytics
Tools & Apps
External or
Cloud
Data Sources
Partner or
Customer
Analytics Tools
Enterprise Datacenter(s)
All-in-One Service
“As a Service” Includes:
• Cazena Gateway
• Workload Engine Licenses
• Cloud Infrastructure Fees
• Operations and Upgrades
• Monitoring and Security
• SLA Optimization
• Fast Managed Setup
• 24 X 7 Support – One Call
Data Lake
as a Service
Data Mart
as a Service
How Cazena Fits In
Data Mart/Warehouse (MPP SQL)
as a Service
Futures in evaluation by
R&D team: Search, etc.
Data Movement, Integration with Data Flow
Operations, Monitoring, Backup/Recovery
Security, Governance, Encryption
Data Lake (Hadoop or Spark)
as a Service
Best-of-Breed Workload Engines, Benchmarking
Infrastructure (Amazon Web Services, Microsoft Azure)
Data Science / Analytics BI / Visualization Applications
Data Architecture | ETL/MDM | Modeling
Cazena
Big Data
as a
Service
Enterprise/
Partner
Cazena: Pre-Built and Production-Ready
Do it Yourself
~200+ tasks, many components,
Multiple vendors, interfaces, SKUs
Must integrate with enterprise
3 – 12+ month deployment
Multiple support organizations
Ongoing R&D and optimization
Cazena Big Data as a Service
One platform, one interface
One vendor, one interface, one SKU
Easy enterprise integration
2 week deployment
Single support contact (24 x 7)
Optimized, future proof
Next Steps and Q&A
Next Steps
• Learn more at cazena.com
• Contact us for a free migration assessment
• Validate with a one-month pilot