Of course you can perform you own test, however creating big datafiles is quite conclusive because every sungle block has to be "formatted" .
Measure IO performance and u will see that you can pump much more data per sec on SAN versus same server using NAS.
BTW, what is your evidence that NAS is same or better than SAN?
One, who thinks that the other one who thinks that know and does not know, does not know either!
Of course you can perform you own test, however creating big datafiles is quite conclusive because every sungle block has to be "formatted" .
Measure IO performance and u will see that you can pump much more data per sec on SAN versus same server using NAS.
BTW, what is your evidence that NAS is same or better than SAN?
That only tests Sequential Write performance. Which is hardly representative of a DB. Except for the online redo logs and maybe a few special cases.
My thought would be to test Sequential Read, Sequential write, Random read and Random Writes. Then make a determination as to the best storage medium for your applicaiton.
Just a thought:
I really don't see an edge either by NAS or SAN storage device without considering or understanding first the other entities of an Infrastructure.
I mean.. if your infra involved with low or high network performance. Or your servers maybe an entry-level, midrange, or high-end, this clearly will influence the performance of your storage device.
If you have at least a midrange server which has a fibre channel bus architecture(IO comm channels/network/host adapters) then it will best fit to use SAN. Because SAN fits or maximized with this. On the otherhand NAS will be most likely be suitable in an infra where communication/network high performance become a standard(using Gigabyte Ethernet network systems/fibre optics).
Reply With Quote
Originally Posted by BV1963
Bookmarks