This content has been marked as final. Show 4 replies
1a. Getting the system administration group to change their standard so you can order the hardware you want and need.
1b. Getting the system administration group to change their standard so you can order the hardware you want and need.
1c. Getting the system administration group to change their standard so you can order the hardware you want and need.
1d. Getting development/DBA team to drop indexes and only rebuild those that are actually required going forward.
2. From whose standpoint?
3b. System Admins can't buy incompatible hardware which is one of their favorite hobbies.
3c. Network Admins don't try to stuff the cache fusion interconnect into a poorly configured and totally inadequate VLAN.
3d. DBAs get a more stable platform.
If you want perfect answers ignore what I wrote and contact dan.norris @ oracle.com (remove spaces).
I'll offer some thoughts on this...
*1. What are the top 4 challenges can be faced during project execution in Exadata?*
Are you referring to technical or political? :D
The biggest challenge I see in the data warehouse domain, is simply to shake off old and bad habits. The performance is in the platform and leveraging it best may mean some refactor & redesign is required. This should not be looked at as a bad thing. Many try the lift-and-shift model, which I despise because it brings the "dirty" past into what is now a "clean" future.
Few areas that come to mine:
- set based processing leveraging parallel execution
- moving away from bulk data modification (bulk deletes and updates) and to an append only/mostly approach and nologging techniques
- just big bulk data processing best practices in general
*2. How different is development or environment in Exadata environment in comparison to not having Exadata.*
Functionally there is really none -- Oracle is Oracle. From a performance perspective there is obvious differences, but that's generally what staging/pre-prod labs are for.
*3. What are the top 4 business benefits/KRAs we improved due to this Exadata implementation*
I don't have specifics here, but many of the customers I've worked with in the DW domain can do more with less and are able to do things that were previously impossible. They spend less time trying to prop up the system with "unnatural acts of engineering" and focus on delivering data to the business and are able look ahead to new projects that enable the business even more.
Given the perspective with which your question was framed, I'll assume you already have the answers.
A few others
1. What are the top 4 challenges can be faced during project execution in Exadata?
a. Getting the business to adapt to a different division of responsibilities than they are comfortable with. For example the storage team loses a lot of the control they had.
There is a lot more sharing of perfromance issues between os and Dba teams.
b. Convincing people that Exadata is not a whole new skill set.
c. Convincing people that on the whole Exadata is a comparatively stable platform.
d. As was mentioned before , getting the OS admin team to consider a new technology stack.
2. How different is development or environment in Exadata environment in comparison to not having Exadata.
From a developer perspective not too different. I think the major issues are optimizing for performance. Supposedly the way Exadata is implemented you could retrofit performance to an existing Application by making indexes invisible and then dropping them, Pinning lookup tables in flash cache and other non code related changes.
3. What are the top 4 business benefits/KRAs we improved due to this Exadata implementation
"improved " would imply you've already done it. Performance is one,single vendor is another benefit ,ability to consolidate and get rid of a lot of legacy hardware and associated support costs could be another depending on your situation. Often Exadata also forces a move from older database version to the latest 11Gr2 which again adds to your supportability.