Data Warehouse Appliances – Sensible Decision Taking Exercise
I have started this blog thread by not giving what are the advantages of Data warehouse Appliance (DWA)? Comparisons over DWA vendors – which are the better fit and which is the best? RATHER how customers has to evaluate DWA before they invest big time. If I would like to category this blog then this would be bucketed under sensible decision taking exercise rather strategic or tactical decisions exercise.
- A data warehouse appliance should be easy to set up, configure and get it working. Ease of use is probably one of the main reasons why customer considering buying an appliance. Please look into this factor validated during a demo. The demo sessions would be lengthy but it is worth conducting this acid test during the demo sessions with the vendors.
- It is suggested that customer should not be buying a DWA because it’s MPP but it provides the functionality and scalability which is also primarily needed to evaluate. Get the biggest data set which has functionality to grow large in size. Automate / simulate the growth in the DWA and see how scalable the DWA is all about.
- Please conduct a full and complete demo on the DWA. It is important that more than PPT demo it is important to do the demo over the box from initial setup till analytics. Use your needy checklist to make comparisons with different vendors though the vendors would provide a complete list of comparison favoring their product to be sold.
- A pre-built demonstration of analysis which is usually packed in the DWA demos to me is a worthless a watch. Do not allow the sales team to demo with pre-built analytical analysis which they brought that to demo. It is important to collect data from the internal power users and bring existing analytical package to be tested in the DWA. It is very important to keep focused on the problems you want it to solve rather than on the flashy demo.
- The right step to evaluate DWA would be building the data set which is of similar size of the existing EDW. Steps have to be taken to mask the data and have full proof data erasure process before even start evaluating the DWH.
- It is important to quantify that DWA should have inbuilt data cleansing utility. It is important that you mask a data which is not clean / not of good quality in nature – Data warehouse appliances is expected to bring this package. Most data in source systems isn’t clean, so it is informative to ask how the data warehouse appliance is going to address this problem
- DWA are built on commodity components, the main cost of the appliance is unlikely to be the hardware; it will be those somewhat more intangible elements, the software and the support. Intangibles are always up for negotiation, and in a recession you have more scope.
- Please commit with the vendor with the turn-around time to deliver the box. It is important to get it clearly defined in Contract. Please negotiate aggressively on the delivery date.
To conclude, please do a complete diligence with what you need from the DWA rather what it could offer. Also foresee how scalable DWA could be to accommodate your needs in a larger picture.
The writer is the Practice Director – Business Intelligence & Analytics Europe , Hexaware Technologies.