Netezza Vs Teradata
Topics
Key Design Principles
Key Technology Requirements
Key Physical Architecture Principles
Parameter for Comparison
Evaluation Consideration
Conclusion
2
Key Design Principles
3
Criteria
Parallelism across the
database platform
Dedicated Physical
Architecture
Proprietary intelligent
system inter-connect
NA
Equal Distribution of
Data across the Data
Storage
*** Netezza FPGA reduces the network cost and boosts performance
*** Now both Teradata & Netezza supports Map-Reduce to re-distribute the data
Key Technology Requirements
4
Criteria
Data Volume
Parallel Query
Simplification of Physical
Schema
Parallel Data Load
Incremental Additions
Abstract Complexity in
Query Execution and
Query Data Volume
*** Netezza dose not have any Primary key concept which simplifies the Physical Schema
*** Both Netezza & TD provides inbuilt utilities for Parallel, Incremental load
Key Physical Architecture Principles
5
Criteria
Each block within the
Server has Dedicated
Hardware and Process
Common Process
Controler with high speed
interconnectivity across
the blocks
Distribution of Data across
the Data blocks controlled
by Common Processes
Scalable for Cloud BI
In-memory Analytics
Parameter for Comparison
6
Criteria
Performance and Linear
Scalability
More Scalable Scalable
Reliability and Availability Mirror imaging and backup
SPU
Data Movement Capacity In Built Utility
Maintenance Overhead Less Medium
Replacement Price Less Medium
Data Integration Tool
Support (ETL, EAI, Web
services)
Depends on environment to
be evaluated
Depends on environment to
be evaluated
Real Time Integration Proven Proven
Data Architecture Support
and Transaction Hub
Capabilities
Depends on environment to
be evaluated
Depends on environment to
be evaluated
TD Vs NZ Architecture
7
Netezza Vs Oracle Vs IBM
8*** From Netezza site, and in reality I have experienced the same performance
9
Comparison
Netezza adopted a 2 Tier Architecture within Hardware rather a Node based architecture with all
high level functions handled using SMP and Low level functions are handled using the SPUs. The
SMP host connects to hundred of MPP drives called Snippet Processing Units (SPUs)
Netezza Database is designed to spread in a vertically integrated hardware matrix along with
horizontally integrated components.
The Scalability is addressed by Netezza’s Asymmetric Massively Parallel Architecture (AMPP).
The AMPP uses modular component at two architectural levels. At the higher level the components
are rack mounted standard SMP units and at the lower level series of intelligent controllers and
redundant components that makes a small package easily manageable and extensible compared to
Teradata.
The High degree of reliability is achieved in Netezza by full redundancy at every level of the 2
tiered AMPP Architecture.
In Teradata the parsing engine decomposes the query passed and then executes in parallel on all
SMP Nodes. The parsing engine in a peer relationship to the execution engines all on the same SMP
architecture.
In Netezza, the Intelligent Query Streaming Technology leverages the 2 dimension architecture.
The physical database level processing are moved at the hardware level . This allows processing
decomposition into many more processors than any other product. This increases the magnitude of
performance compared to Teradata or any other product.
10
Comparison (Conn….)
Netezza uses the Ethernet Switch Technology Compared to Teradata’s bynet Technology. The
bynet is based on TCP/IP, Rather the Gigabyte Ethernet is based on Switch technology . This
technology can share vast amount of data across many processors. This leads to optimization of
resources across the units and supports high degree data transfers. This addresses the high
degree of bandwidth requirement of Netezza compared to Teradata.
Netezza implemented a new disk controller architecture in the Performance Server. This disk
controller device partners with the SMP platform to disperse the disk controller functions that are
processed normally only by host alone. This is a key breakthrough of Netezza compared to other
databases
Netezza does not have much database administration tasks such as Index, Spacing and Other
Objects. This reduces the Operational Admin tasks and eases the maintenance of the same.
Teradata related Adapters from ETL ,EAI ,Messaging and Business Automation Tools are tested
across multiple business/Technical scenarios compared to Netezza.
Scenario (Existing Organization has 100 TB and in next 5 Years it will reach 500 TB)
If you are buying 100 TB Box, upfront you have to pay for the 100 TB Box in Teradata
If you are buying 100 TB Box, you just have to pay per usage (Only for 100 TB) in Netezza
Netezza automatically re-distribute the data by looking the access Path for better query
performance.
11
Evaluation Consideration
Scalable : Make sure proposed solution scales in a near-linear fashion and behaves consistently
with growth in database size, number of concurrent users and complexity of queries. Understand
additional hardware and software required for each of the incremental uses.
Powerful : Designed for complex decision support activity in a multi-user mixed workload
environment.
Manageable: Minimal support tasks requiring DBA/System Administrator intervention. It
should provide a single point of control to simplify system administration. We should be able to
create and implement new tables and indexes at will.
Extensible: Provides flexible database design and system architecture that keeps pace with
evolving
business requirements and leverages existing investment in hardware and applications.
Available: Supports mission critical business applications with minimal down time. These can
include batch load times, software/hardware upgrades, severe system performance issues and
system maintenance outages.
Interoperable: Integrated access to the web, internal networks, and corporate mainframes.
Affordable : Proposed solution (hardware, software, services, required customer support)
providing a low total cost of ownership (TCO) over multi-year period.
Proven: Platform must be proven.
Flexible: Provides optimal performance across the full range of normalized, star and hybrid data
schemas with large numbers of tables. Proven ability to support multiple applications from
different business units, leveraging data that is integrated across business functions and subject
areas
Cost: Over all the onetime installation cost, maintenance/Support cost
12
Conclusion
Netezza Scores well from the subjective Architectural Analysis standpoint. Netezza
addresses the VLDB (Very Large Data Base) needs of organization with New
Generation Technologies rather trying to address the bottleneck in the existing
technology in an cost effective manner.
Retailer DW/BI Platform
13
Retailer DW/BI Platforms
WalMart Teradata, Netezza
Pepsi Co Teradata
Sears Teradata, Netezza (Migrating to NZ)
Safeway Inc. Teradata, Netezza (Migrating to NZ)
Sysco Foods Netezza
PetSmart Inc. Netezza (Migrated to NZ from Oracle)
Procter & Gamble Teradata
Kroger Teradata
The Growing Netezza Community…
15
About Author
Asis Mohanty has more than 11 Years of Industry experience on Data Warehousing and
Business Intelligence field. Asis has wide Industry experience in Retail/CPG (Procter &
Gamble, Pepsi Co., Sears Holdings, PertSmart Inc, Sysco Foods, Safeway Inc), Telecom
(AT&T), Manufacturing (Apple Inc), Insurance & HealthCare (BlueCross Blue Shield, North
Western Mutual, LTC Partners).
He has executed various DW re-platform programs such as Oracle platform to Netezza
platform, Teradata to Netezza, UDB to Teradata.
He is a Certified Business Intelligence Professional from www.tdwi.org and Certified Data
Management Professional from www.dama.org. He is also Netezza Silver Certified
Professional, MicroStrategy Certified Professional.
Thank You!