Timeline for Data center from scratch which way to go
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 2, 2013 at 18:15 | comment | added | David Spillett | (BTW: you should add that extra sizing information to your question, then other people who might post a different answer are more likely to see the extra data and use it to make their more relevant) | |
| Sep 2, 2013 at 18:11 | comment | added | David Spillett | Unless you need to for specific performance reasons, try to avoid denormalising your data. Improving your base table+index design is usually the better way to go. If you need to denormalise for reporting purposes, try to avoid doing this on the live database and do it on copy specifically for such reporting copy instead (you may want to look into transforming the data as you copy from live to reporting copies, instead of just maintaining a direct replica: search around the term ETL (Extract/transform/load), and perhaps OLAP (Online analytical processing) too, for more pointers regarding that. | |
| Sep 2, 2013 at 18:05 | history | edited | David Spillett | CC BY-SA 3.0 |
added 76 characters in body
|
| Sep 2, 2013 at 18:04 | comment | added | adopilot | My main database is 80GB but It is not normalized, and I have a lot missing index, I except that it will grow up at leas 0,5TB when I will have space to demoralize it, Like extincting sells data per days,weeks etc. | |
| Sep 2, 2013 at 17:51 | history | answered | David Spillett | CC BY-SA 3.0 |