Selection of DW and analysis software? 1

[lcrans]

Hello, all....

We are an educational organization, several school districts joined together to share costs and development efforts related to curriculum and student performance analysis. We are currently looking for a DW/OLAP system, a way to pull together all our data and study it. And there is a question that has come up:

For roughly the same annual costs we have a choice between the following:

1) An existing education oriented DW/OLAP/Reporting system developed in SQL Server and .NET from a company with about 40 employees;
2) A to-be-developed system (in MySQL, PHP, HTML, and Javascript) to do what we say we want from a company of 8 employees that has developed other smaller application packages for some member districts.

What are your thoughts about selecting one of these options, and why? Your ideas will really be of value in out selection process....

Thanks....lcrans

 

[blom0344]

Well this is certainly a difficult one to comment on.........
We have both DB2,ORACLE and SQL server environments around and our experience is that for relative small amounts of data SQL server seems to be the best bet. (I have no experience with MySQL though) Compared to other big RDBMS systems SQL server (with its T-SQL) is probably slightly less sophisticated, because it does not have all the newest OLAP type of SQL functions, but it certainly beats the other ones in terms of ease of maintenance,backing up data and other DBA kind of tasks. The interface is also much slicker (partly because SQL server has fewer years of history to drag with it. )
I think you should concentrate on these issues issues:

1. How will you consolidate data from various platforms to a staging area? Do you need third-party tools to get the job done? Are all extractions going to be hard-coded consultant jobs which for ever will require outside assistance for every modification?

2. How advanced do you want your reporting to be. Can you use straightforward database views to fetch data into spreadsheets or do you want a sophisticated third party tools that gives you probably more than you'll ever need (at quite hefty prices)

Just by chance I witnessed a demo from one of the big business intelligence vendors this morning. One of their cases was almost exactly a pilot for the situation you are in and the output was absolutely Sci-Fi, Interactive dashboards, data-drilling you name it. You can really get a sophisticated front-end if you have the money........

T. Blom
Information analyst
tbl@shimano-eu.com

 

[Rob999]

I find it unbelievable that the costs of an EXISTING Microsoft based based system would be as expensive to implement and maintain as the "to-be-developed system (in MySQL, PHP, HTML, and Javascript)". Especially the maintenance part. The alternative system sounds web based, which is certainly going to limit the likely user interfaces. I agree pretty much with T. Blom, but - without starting any holy wars - I think SQL Server 2000 is easier to use and just as scalable as any alternatives - particularly something like MySQL, which I've only used briefly.

Perhaps the most important thing for you however, is to nail down what your getting for how much money. Ask for a specs listing on exactly what functionality you'll be getting - Know what you want!

 

[BillDHS]

Don't be short sighted in your selection. Who is going to support and maintain the application after it is built? All too often, a consultant will come in and start to built the system, the scope will creep, the contract will expire, and the application will never be completed.

Find the tools that you can support over the long hall. If you currently have SQL Server support staff, don't bring in a new achitecture that you will be under continuous development.

 

[MarleneD]

Universities and some school districts publish details of their Data warehouses online.

This would be a good research avenue to proceed.