Non relational data approaches applied effective can result in massive cost reduction and performance improvement compared to an infrastructure of legacy enterprise hardware and software solutions. While still not totally without risk on an enterprise scale some tech savy early adopters are realizing tens of millions of dollars in total cost savings. Astute Corporate IT Buyers should include this on their roadmaps if for nothing else to leverage Legacy IT providers
IT Category Purchasing Managers Opportunity for Savings with Non Relational Systems such as Hadoop
1. Open Source Non Relational Storage Systems
A Strategic Cost Savings Opportunity for Purchasing IT Category
Managers
Bill Kohnen
IT Procurement Forum Discussion
San Francisco CA
2. With growing volumes of data and increasing requirements to
process and analyze data even faster, organizations are faced with
several options:
1. To add more hardware and/or horsepower to their
existing infrastructure and operational systems. Very
expensive and especially with legacy hardware and ERP
providers (HP, Cisco, EMC, Oracle, SAP etc.) Also
performance only scales but does not improve
2. Consider alternative ways to manage their data.
3. Do nothing. Organizations must ask themselves is all
data important and should they try to capture all of it to
process, analyze and discover greater insights in it?
3. Benefits of using Open Source Non Relational
Storage Systems
• Open source software - Lowers Cost
• Running on commodity hardware Lowers Cost
• Performance is better than that of traditional databases
• Decades of data can now be stored more easily and cost-effectively.
• Data does not need to be destroyed after its regulatory life to save
on storage
• Analysis can be conducted on larger set of data
4. Cost Savings
• Hardware 60%
• Development 30%
• Other software 50% (ETL, Enterprise BI
solutions)
• Enterprise Software Maintenance 100% for
some applications
• Enterprise Hardware Maintenance 80%
Even Mid Sized Companies Can Save Millions over several years.
The most aggressive early adopters like Facebook have saved hundreds of
millions combined total cost.
5. Enterprise Data Warehouse (EDW) - Simplified, Traditional setup:
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
6. Open Source Non Relational File System - Simplified setup:
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
7. Typically Run Parallel
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
8. Ways to Shift the Current Corporate Data Paradigm with Open
Source Non Relational Systems such as Hadoop
• Stage structured data
• Process structured data
• Process non-integrated & unstructured data
• Archive all data
• Access all data via the EDW
• Access all data via Hadoop
9. Stage Structured Data
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
Reduces Storage Costs, Frees Enterprise Processing Power, and de bottlenecks ETL
10. Process Structured Data
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
Reduces Storage Costs, Frees Enterprise Processing Power, and de bottlenecks ETL
Even if you do not currently have massive big data sources
11. Process non-integrated & unstructured data
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
Reduces Storage Costs, Frees Enterprise Processing Power, and de bottlenecks ETL
When you say all data is important but want it available in both systems
12. Archive All Data
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
Eliminates need to Purge Data so can Analyze Big Sets of Data
13. Access all data via the EDW
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
More Cost Effective way of maintaining Legacy System as “system of record”
14. Access all data via Non Relational Database such
as Hadoop
Structured
Data
Enterprise
Data
Warehouse
BI
Analytics
Un
Structured
Data
Open Source
Non
Relational File
System
Big Data
APs
Paradigm Shifting Approach
15. Summary
• IT Category Managers should work with corporate IT
to evaluate potential incorporation of non relational
database approaches as a major cost and
performance improvement
• There are still near term risks as the technology on
an enterprise scale is still maturing
• A major threat to Legacy hardware and software
providers and even new BI tool market