Moss L.T., Atre S. Business intelligence roadmap: The complete project lifecycle for decision-support applications
Подождите немного. Документ загружается.
[ Team LiB ]
Chapter Two. Step 2: Enterprise
Infrastructure Evaluation
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
81/631
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
82/631
Chapter Overview
An enterprise infrastructure is to BI applications what a transportation infrastructure
is to automobile owners. In order to safely and comfortably travel with an
automobile, there must be a physical infrastructure, such as roads, bridges, traffic
lights, and traffic signs, as well as nonphysical infrastructure, such as standardized
traffic rules and their interpretation. For example, without the universal
interpretation of the rule that "Green means go, red means stop,"traffic lights would
be of no use. Similarly, an enterprise infrastructure consists of two major
components:
Technical infrastructure, such as hardware, middleware, and database
management systems (DBMSs)
1.
Nontechnical infrastructure, such as standards, meta data, business rules, and
policies
2.
Accordingly, this chapter is divided into two sections—Step 2, Section A, Technical
Infrastructure Evaluation, and Step 2, Section B, Nontechnical Infrastructure
Evaluation. The first section covers the following topics:
Things to consider about technical infrastructure
The importance of scalability for the hardware platform
Middleware, with emphasis on DBMS gateways since they are one of the most
important middleware components for BI applications
DBMS requirements for the specific functionality needed to support BI
applications
Brief descriptions of the technical infrastructure activities, the deliverables
resulting from those activities, and the roles involved
The risks of not performing Step 2, Section A
The second section, on nontechnical infrastructure, covers the following topics:
Things to consider about nontechnical infrastructure
Bad practices and old habits that lead to stovepipe development (automation
silos)
The need for a nontechnical infrastructure to enable an integrated BI decision-
support environment
The enterprise architecture components: business function model, business
process model, business data model, application inventory, and meta data
repository
Enterprise standards for such things as data naming, data quality, and testing
Brief descriptions of the nontechnical infrastructure activities, the deliverables
resulting from those activities, and the roles involved
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
83/631
The risks of not performing Step 2, Section B
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
84/631
[ Team LiB ]
Step 2, Section A: Technical Infrastructure
Evaluation
Things to Consider
Hardware
What hardware platforms do we already have or use?
On which platform should we implement the BI application?
Do we need new hardware? What will it cost?
Will we need more staff to maintain the new hardware?
Will the new hardware integrate with our existing platforms?
How will the new hardware scale to accommodate ever-increasing
loads of processing and volumes of data?
Network
What type of local area network (LAN) are we using?
What type of wide area network (WAN) are we using?
Is the bandwidth of our WAN sufficient to grow?
Middleware
What type of middleware do we already have or use?
Do we have the necessary middleware to retrieve the source data
from heterogeneous platforms and transfer it to the BI decision-
support environment?
What is the operational source architecture? (e.g., enterprise
resource planning [ERP], legacy files)
Do we need new middleware? What will it cost?
Will the connection be permanent between the source files (or
source databases) and the BI target databases?
Which of our hardware, software, and middleware is proprietary?
Have we purchased it? Or are we leasing it?
Database Management Systems
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
85/631
What DBMSs do we already have?
Will we need to buy a new DBMS? What will it cost?
Will the new DBMS be compatible with our operating system(s)?
What software tools can run with it?
Does our staff have the skills to use and administer the new DBMS?
Will we have to hire more database administrators?
Tools and Standards
How are the business analysts currently analyzing the data? What
reporting and querying tools do they use?
What additional tools and utilities do we need?
What other software do these tools need to interact with?
Do we know of any major problems with our technical
infrastructure?
What are our technical standards for compatibility and access?
The development efforts of early BI applications, such as the early data warehouses,
were relatively slow, labor-intensive, risky, and expensive. Extraction and
transformation of operational data into a data warehouse frequently involved
creating new, custom-written application code. The target databases were either
based on proprietary DBMSs or were using proprietary hardware platforms. There
was also a shortage of tools to administer, control, and expand the new decision-
support environment. The lesson learned from the early BI days was that in order to
reach the best performance results for data access and retrieval, a comprehensive
application platform must be chosen. Therefore, it is important to select the
appropriate hardware, middleware, and DBMS and to ensure that these components
are implemented properly.
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
86/631
[ Team LiB ]
The Hardware Platform
For adequate report and query performance, it is very important to have sufficient
"horsepower" with the hardware platform. Scalability is of utmost importance.
Controlled Chaos
Do not despair if your computer environment looks like the one in Figure 2.1. This is
more often the case than not in organizations of any size. What exists can at best be
described as controlled chaos!
Figure 2.1. Controlled Hardware Chaos
Accompanying the hardware chaos are usually a huge portfolio of disparate software
and a large staff with only enough skills to support the existing systems. In order to
minimize the chaos, most organizations implementing a BI decision-support
environment have to consider at least four imperatives in hardware platform
selection.
New hardware platforms have to fit into the existing hardware configuration.1.
The DBMS on the selected hardware platform must perform well as database
access and usage grow. Scalability is therefore one of the major issues to be
addressed.
2.
Platform selection is restricted by the need for interoperability between various
hardware platforms (if required).
3.
Cost and return on investment (ROI) for the previous three qualifiers are4.
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
87/631
3.
controlling factors.
4.
Hardware Platform Requirements
The hardware must have sufficient power to handle complex access and analysis
requirements against large volumes of data. It has to support not only predefined,
simple queries on summary data but also ad hoc complex queries on detailed data. It
must also be scalable because rapid changes will occur in:
Data volumes
Updating frequencies
Data access patterns
Number of reports and queries
Number of people accessing the BI target databases
Number of tools running against the BI target databases
Number of operational systems feeding the BI target databases
It is useful to think of a BI decision-support environment in terms of a three-tier
computing architecture (Figure 2.2). First, the extract/transform/load (ETL) engine
extracts, cleanses, and transforms operational data. Then, using middleware, the BI
target databases are populated. Finally, when data is requested, it is mapped into
suitable representations for the business community at the interface level for running
queries, reports, and online analytical processing (OLAP) applications. The interface
level can be a customized graphical user interface (GUI) application, an enterprise
portal, or Extensible Markup Language (XML) Web services.
Figure 2.2. Three-Tier Computing Architecture
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
88/631
[ Team LiB ]
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
89/631
[ Team LiB ]
The Middleware Platform
The term middleware refers to runtime system software, which is layered between
the application programs and the operating system. It acts as a bridge to integrate
application programs and other software components in an environment with
multiple network nodes, several operating systems, and many software products.
Middleware is needed to run client/server architectures and other complex networked
architectures in a distributed computing environment. Therefore, the middleware
should support directory services, message-passing mechanisms, and database
gateways.
Most middleware falls into two major categories:
Distributed logic middleware supports program-to-program communication
between two pieces of custom-written application code.
1.
Data management middleware connects an application or DBMS on one
platform with a DBMS running on another platform.
2.
Middleware can also be used to enable "reach-through" queries from summaries in
the BI target databases to the underlying detail data held in operational systems. To
keep the cost to a minimum, a number of organizations are already using gateways
to transfer data from multiple heterogeneous sources of server data to client
workstations, as illustrated in Figure 2.3.
Figure 2.3. Gateway Example
北斗成功社区 BeiDouWeb.com 教育音视频/电子书/实用资料文档/励志音乐影视 仅供免费试用/版权原著所有
90/631