Data Warehouse Appliances
Data Warehouse Appliances [DWA] were designed
specifically to support streaming business intelligence [BI] workloads.
They consist of architecturally integrated hardware, DBMS and storage
into one device.
The aim of the DWA architecture is to combine the best elements
of SMP and massively parallel processing [MPP] into one solution
to optimize queries, and thereby removing bottlenecks to data flow.
By supporting a streaming data-flow, through standard interfaces,
a data warehouse appliance is fully compatible with existing BI
applications, tools and data. The only constraint to performance
is the disk speed.
It has an extremely low total cost of ownership, requiring minimal
system integration and administration. Data warehouse appliances
are far simpler to install and maintain than a typical database
and storage infrastructure.
To many data warehousing appliance and appliance-like systems are
seen as 'disruptive innovations' in that they divert significantly
from the norm. This has largely been fostered by certain claims
that the promised super performance did not stack up when tested
in a real life environment of complex, mixed workloads and multiple
concurrent queries and updates.
This could well be the reason that Teradata has refused to adopt
the 'appliance' term in reference to it massively parallel processing,
shared nothing Active Data Warehouse. The ADW is a much greater
innovation than simple DWA.
However, other claims demonstrate that processing 1.4TB of data
[table - 5.7 billion rows], a data warehouse appliance will perform
20 times faster for less than half the total cost of a patchwork
Typcially, the need for data warehouse appliances was justified
in terms of a hierarchy of technology power needs:
DWA and Data Marts
Most data warehouse appliances support large data marts, integrated
to analytic applications, focused on a single subject such as
call-level detail, customers, shopping baskets etc.
Some CIO's use data marts as a strategic shield to introduce
a data warehouse appliance into their data center with minimal
risk. The appliance satisfies the analytic application’s
requirements and also serves as a 'proof of concept' data warehouse.
DWA and EDW
Deploying an enterprise data warehouse [EDW] on an appliance
is gaining support, where previously EDW's were more commonly
deployed on data warehouse bundles.
Large data marts usually manage between 1 TB and 10 TB of live,
query-able data. This range is moving up as vendors introduce
models of greater capacity. Appliances up to 20 TB are currently
Regardless as to whether you view the data warehouse appliance
market as pretty much over, the database innovation that it stirred
will beneft standard relational technology going forward. IN the
future, enterprises will need only one kind of database to perform
both transactional and business intelligence processes, albeit many
IT departments will continue to implement separate instances for
reasons of operational efficiency.
The DWA integrates with BI solutions and ETL tools using standardized
interfaces, protocols and functionality. This includes interfaces
with MicroStrategy, Business Objects, Cognos, SAS and SPSS.
ETL tool interfaces include Ab Initio, Ascential and Informatica.
The data warehouse appliance works seamlessly with these tools
and applications as well as other in-house applications.
Key DWA Vendors
The key vendors for data warehouse appliances include:
When to Use A Data Warehouse Appliance
A data warehouse appliance should be considered when a terabyte-size
data mart supports an autonomous analytic application.
If the application involves intense ad hoc queries that place greater
load on the organization’s enterprise data warehouse, then
the DWA is even more appropriate.
Implementing a data warehouse appliance for an isolated analytic
application is a low-risk way of proving the value of a data appliance
in a data center.
They are also deployed in projects that require short time frames,
low price per terabyte, and minimal system integration and administration.
Other Uses for Appliances
The use of data warehouse appliances is not unique to data warehousing.
Appliances are also used in:
- Network Appliance [storage]
- Google Search Appliance
- Thunderstone Search Appliance
- Blade servers
- Network storage
- Other rack-mounted systems that resemble appliances.
Performance Server [NPS]
Back To Top
The World's Leading Guide To BI Strategy, Program & Technology
DW Solutions | Teradata
ADW | Netezza NPS | DATAllegro
| HP NeoView | Teradata
12.0 | Green Data Centers