Timeline for How to estimate storage needs using the PERT distribution for filesizes? How to aggregate them without falling into the flaw of extremes?
Current License: CC BY-SA 3.0
16 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 11, 2020 at 14:32 | history | edited | CommunityBot |
Commonmark migration
|
|
| Jul 9, 2012 at 17:50 | vote | accept | Luxspes | ||
| Jul 30, 2012 at 5:26 | |||||
| Jul 9, 2012 at 13:46 | comment | added | whuber♦ | Please take this conversation into chat, guys: chat.stackexchange.com/rooms/4045/how-to-estimate-storage-needs. | |
| Jul 9, 2012 at 13:37 | comment | added | Luxspes | Would you kindly explain why I am getting a different answer with: $Quantile[Table[Fold[Plus,0,RandomVariate[PERTDistribution[.25,5,1],40000]],10000],0.964]$ | |
| Jul 9, 2012 at 13:33 | comment | added | Luxspes | Are you reading the updated question? The Result is 61Gb NOT 128Gb you can do it yourself just write: $Quantile[Table[Fold[Plus,0,RandomVariate[PERTDistribution[.25,5,1],40000]],10000],0.964]$ | |
| Jul 9, 2012 at 10:30 | history | edited | image_doctor | CC BY-SA 3.0 |
Add CDF graph
|
| Jul 9, 2012 at 8:36 | history | edited | image_doctor | CC BY-SA 3.0 |
added 451 characters in body
|
| Jul 9, 2012 at 8:02 | comment | added | Luxspes | That was not your answer, what you wrote was: "If you specify 128GB of storage you will have a 0.964 chance of your storage not being exceeded" But, it turns out, that I only need 61GB to achieve that... ¿care to explain why the simulation gives such a different result? | |
| Jul 9, 2012 at 7:40 | history | edited | image_doctor | CC BY-SA 3.0 |
clarify the function being referred to
|
| Jul 9, 2012 at 3:06 | comment | added | Luxspes | Sorry @image_doctor but it appears that you were wrong, theres is a big difference between generate the 40,000 file sizes, 10,000 times and the current answer...(see the changes in the question) | |
| Jul 8, 2012 at 16:48 | vote | accept | Luxspes | ||
| Jul 9, 2012 at 2:52 | |||||
| Jul 8, 2012 at 9:22 | comment | added | Luxspes | For example, I could generate the 40,000 file sizes, 1000 times, and analize the resulting 1000 total file sizes (using something like $Table[RandomVariate[PERTDistribution[{0, 1}, 0.5], sampleSize], {i, numOfSamples}];$ But... ¿should I? | |
| Jul 8, 2012 at 9:18 | comment | added | Luxspes | I agree that I should err on the side of caution in my spec, but, regardless of the possible error in my input parameters.. would my estimation be a lot more precise if, for example i were to generate a number of samples of data, and then re-compute thestatistics of those samples? | |
| Jul 8, 2012 at 3:32 | comment | added | Luxspes | I am being told (by some books I am reading) that one shouldn't just aggregate estimates... but in this case I am estimating the size of one file and then "aggregating" it 40,000 times to guess the storage capacity... ¿am I not violating this "do not just aggregate estimates" rule? ¿shouldn't I be using some special formula to estimate the full storage capacity instead of just making a multiplication? | |
| Jul 8, 2012 at 3:26 | comment | added | Luxspes | ¿why not specify 1 TB and be done with it? Server resources upgrade cycle, they are offering me a server with 140Gbytes, or, a server with 1TB, if I choose wrong, I have to live with my wrong choice for 18 months before being able to apply for an upgrade (internal multi-national fortune 100 company with this silly impossible-to-change policy). Now the easy thing is to ask for the 1TB but they want a document justifying it BEFORE I even start coding the sytem that is going to run there (I also need to do the same with processor power, RAM usage and network traffic) | |
| Jul 8, 2012 at 1:38 | history | answered | image_doctor | CC BY-SA 3.0 |