Posted by arlene

The data warehouse is there to be accessed by users. The whole purpose of building the data warehouse is to give an identified set of users access to a particular set of data.

The whole area of access to the data is a minefield. There are a host of tools that can be used, each with its own advantages and disadvantages. There is a battle raging in the industry over relational versus multidimensional tools. New terms such as OLAP, MOLAP and ROLAP are being bandied around. This makes the choice of tools an interesting challenge if not an impossible task. The aim of this appendix is to aid you with this task by discussing all of the major issues.

Some Tips on Data Access Tool

There are a couple of golden rules relating to tools that need to be adhered to when designing the data warehouse:

1 Never design the data warehouse to suit a specific tool.

2 Make no assumptions about the type of tool that will be used.

Rule 1 is the cardinal rule. Never make the mistake of designing the data warehouse around a particular tool. Even if a specific tool has been chosen as a company standard, there is no guarantee that some users with specific needs may not use different tools, or that the standard will not change. The tools marketplace is constantly changing, and any choices made today may be invalid in a few months’ time.

Remember that the data warehouse will evolve over time; new users added later may have different requirements, and different tools. It is an extremely costly and painful task to reorganize a data warehouse after the fact.

Unfortunately, designing the data warehouse around a specific tool is one of the common mistakes made. It is impossible to stress strongly enough how costly this can prove. If the data warehouse is structured around a specific tool and that tool is later found to be inadequate, or other tools are required for specialist operations, it is unlikely that these tools will work efficiently. The data may be structured in just the wrong order, or too few or too many tables or views may be involved. Experience shows that the effect of the data warehouse’s being designed around one tool on the performance of a second tool can be anywhere between 30 and 300%.

Once rule 1 has been accepted, rule 2 follows logically. If you cannot predict the tools or types of tools that will be used, it makes no sense to design the data warehouse for a specific type of tool. The data warehouse should be designed for efficiency of query access, and ease of management. Clearly the design has to be centred on some basis. Primarily it will be based on the understanding of the business requirements, and the types of query that are likely to be asked.

Possibly related posts: (automatically generated)
Data Warehouse, how far we have gone?

• Ten Scamps of a Data Warehousing Failure
• Database Tool, Efficient Enhance
• Data Project Plan Note part 2
• Data Project Plan Note part 1
• Data Warehouse, Databases Easy Access Top ten Considerations part 1
• Internet Commerce Virtual Warehouse
• Hand in hand Database Design and Data Backup, Recovery
• Web Server Identifying your Customers by their Email Addresses
• Database Warehousing Skills
• E-newsletter Design