Solution | Data Author: Xiahong, vice director of marketing, Sybase Software (Beijing) Co., Ltd.
Content: Data Warehouse concept, Sybase Data Warehouse Solution
--------------------------------------------------------------------------------
The concept of a data warehouse
Any company or enterprise, in order, inventory, Bill list, account liquidation, customer service and
Financial reports and so on have a lot of business applications and technical links. The role of the Data Warehouse is to: from this
Some application systems to get information and switch to a new database, through the new library in the history of information and
Analyze the information of the subject and support the decision. Previous product systems, such as ordering or purchasing systems,
It is difficult to obtain information about the state of business development.
Data warehousing is a part of enterprise decision support. Before making the next decision, the lines in each business organization
Many key business issues need to be clearly understood by political and analytical personnel, for example: Which products are most advantageous to
Which customers will bring us the best benefits? Which ones are going to cost a lot of money? which campaigns
Run best, why? What customers are we likely to lose, and why?
These are data warehouses to
The answer to the "million profit" problem is also one of the biggest markets. According to Gartner estimates, 60%
The Close
The system of database management is used as the application development of decision support system.
Comparison of data Warehouse and data mart
In the mid 1980s, Bill Inmon first proposed the term "Data Warehouse". It was initially
Designed as a business database, with stability (the main ingredient unchanged), historical (including historical information) and
To the theme (information by the customer, product and market, etc.) characteristics. These initial "data warehouses" are based on
Analysis of customer, product, sales and financial situation, get the whole understanding of enterprise activities.
To build a data warehouse, there are generally four steps:
The first step: database design, that is, the design of a database containing business data and information for business entities
Used;
The second step: the development of data extraction and conversion program, from the product system to put the data into the Data Warehouse;
Step three: Develop data loading and updating technology so that when the product data changes, the Data warehouse gets
Dynamic and real-time updates;
Fourth step: Acquisition of query and report generation tools, so that users through the enterprise intranet and personal computer very side
Access to information.
Over the years, customers have found that although enterprise-class data warehouses are attractive, they are difficult to operate.
Degree. The 1996 IDC Research survey showed that, despite the average investment of three years to build a data warehouse
Time and nearly 3.2 million dollars, 50% did not achieve their due effect. From the beginning of the project, three years later, most
Several traders found that the business problems they faced were no longer what they were when they began to build, and a lot of changes took place.
In addition, although development progress has been extended year after year, it is still not done to get all interested customers to
See what information gives a clear definition of requirements. Thus the establishment of an "enterprise data Model" is like practice.
It's been a year and a half.
In the last 18-24 months, a new solution has emerged, the data mart. Data
The bazaar is also a data warehouse, but it is more concise and more topic-oriented. Since its inception, Sybase Company
has established a technical leadership position in the data mart. Currently, among more than 20,000 customers using Sybase products
Most of the data marts that are running on SQL Server have been established, although they are often referred to as data warehouses,
But almost none of them are enterprise-class.
The advantage of data mart is that the construction cycle is shortened and the cost is greatly reduced. Where the cycle replaces the year with the month,
The cost dropped from millions of to 1 million. Because of the sheer size of the enterprise's data, it's really focused on a
is almost impossible in the database. There's a lot of big data warehouses that are actually not data marts that produce
Doubt. After the data mart is used. Design, extraction, conversion, loading and querying are easier to
For a subset of the customers to know more precisely what information they need.
However, if there are many data marts that cannot keep them synchronized, the Data mart solution will encounter
Difficult. Once a unit creates two or more data marts, the biggest problem is how to make it
are coordinated, how to make them operate in real time, and how to maintain all data extraction and transformation. Other
, when an organization wants to create two or more data marts, it is found that each one is going through a
Re-design, extract, load, and query steps. So, in the face of the development of multiple data marts, how to
Sharing design and architecture has become a realistic and challenging issue.
Operational data storage and merging Data Warehouse
One solution for the above problem is to adopt a new Data Warehouse concept---"Operational data storage
Storage (Operational Data store,ods) ". In ODS mode, data is copied from the business database to
A central location, which is extracted from here to multiple data marts. ODS is from customers, products and other business
A "real-time snapshot" of the business situation that is organized in an industry perspective. It does not contain historical information, but can be very
Easily meet the needs of a historical database or a set of theme-oriented data marts.
We generally call it a "merged Data Warehouse" because it is a letter before entering the decision support database
The combination point of interest. Although the ODS is small, it can be modified frequently, so it is very suitable for the establishment of adaptive
Server
Enterprise and Replication server.
Multidimensional or OLAP (Online analytical Processing) market
As part of the Data Warehouse application process, the market share has been growing rapidly and becoming larger and bigger.
In simple terms, OLAP is an information organization from a business perspective, unlike the usual rows, columns, and tables.
For example, in an OLAP data similar to Arbor or Oracle Express, information is passed through the customer, production
Products, dates, sales departments and geographical attributes to access, which for data understanding and information acquisition are
Seems very intuitive.
After the OLAP product gets the relational data, it is placed in a very simple table that makes it easy to analyze.
Databases and an OLAP product can be viewed as a multidimensional table. This market is quite popular, Arbor,
Oracle's Express and MicroStrategy each occupy a place in this field, while Sybase's
Power
Dimentions (formerly Whitelight), impromptu of Cognos and Powerplay,brio Technology
Brioquery is in a dominant position.
Competitors and partners at a glance
RDBMS Company: Sybase,oracle,ibm,teradata/ncr,informix,microsoft
Hardware Company: IBM,TERADATA,SUN,DIGITAL/COMPAQ,HP
Conversion tools: Vmark,infomatica,carleton/apertus,etz,prism Solutions
Olap:sybase/powerdimentions,arbor,oracle/express,microstrategy,
Information Advantage.
Sybase's solution and its composition
Sybase has a unique and powerful point-to-point solution for designing, building, and managing data warehouses and
According to the fair. Each department interacts through centralized metadata, which has the integrity, concentration, and
Flexibility and other characteristics. Our tools also have a lot of superior performance.
The following table lists the various components:
(1) PowerDesigner Warehouse Architect
PowerDesigner is not only a well-known database design tool in the industry, but also a data Warehouse model design tool. Its
The Warehouse Architect module supports a variety of data warehouse models, including star mode, snowflake mode,
and snowstorm mode. This is the best and most flexible development tool in the industry that can be used to design a relationship or
An OLAP software warehouse. PowerDesigner occupies the largest share of the data Warehouse design tools market. It can be from
The existing database is reverse engineering, from the operating system in the current data structure extracted to form a data model
Type, making the design simple.
(2) Powerstage
Powerful data extraction and data transformation products. It is the leading market for client/server conversion methods that enable the Data Warehouse
The library model is easier and more intuitive to implement with PowerDesigner. Powerstage is really safe and based on
The engine. It has a simple processing-oriented graphical user interface that enables users to quickly start and repeat
Use of the past
