Hi,
I have implemented a test case where I send batches of 300, 500, 600, 700, 900, and 2000
demands. I want to measure size of each demand, so I decided to check the size of jboss
application server folder before and after sending demand, and by dividing the difference
by number of demands find the size of each demand. I did this test case for both
persistent and non-persistent demand migrations. As you see the results, you would notice
that size of each demand remains quite fixed as number of demands increases. However, this
does not apply to the persistent situation, where size of each demand increases as number
of demands does.
Our demand is nothing but just sending a request to compute something. There is no
difference between these two modes, so that's what it rose following questions.
Here are my questions:
1- Is my method of measuring the size of each demand correct? It is for sure not accurate,
but is there any other way to measure the demand inside the jboss demand storage?
2- Why size of demand changes in persistent scenario ?
3- Each persistent demand is roughly 20.41, whereas in non-persistent it's about 7.10.
Is the difference because the overhead in terms of keeping additional information for the
sake of being persistent ? Can you name some of those overheads or reasons?
And finally, here are my result, so you can see the difference ;
Non-Persistent (K.B) Persistent (K.B)
300 7.10 18.33
300 7.09 17.67
300 7.09 19.00
Average 7.09 18.33
500 7.14 24.00
500 7.07 26.00
500 7.06 22.00
Average 7.09 24.00
600 7.06 26.67
600 7.23 25.00
600 7.11 24.17
Average 7.13 25.28
700 7.13 25.71
700 7.12 24.29
700 7.13 27.14
Average 7.13 25.71
900 7.33 26.11
900 7.12 25.00
900 7.11 26.67
Average 7.19 25.93
2000 7.12 26.00
2000 7.13 27.00
2000 7.14 32.50
Average 7.13 28.50
[/img]
View the original post :
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4149278#...
Reply to the post :
http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&a...