1 |
On Montag 27 April 2009, Andrei Susnea wrote: |
2 |
> Neil Bothwick wrote: |
3 |
> > On Mon, 27 Apr 2009 16:18:39 +0400, Yahya Mohammad wrote: |
4 |
> >> I'm setting up a new desktop machine with RAID 0. The motherboard I |
5 |
> >> bought supports the so-called "Fake" RAID, which offloads most of the |
6 |
> >> processing to the system CPU. What are the pros and cons of using this |
7 |
> >> as opposed to pure software RAID? |
8 |
> > |
9 |
> > The advantage of FakeRAID is that you get to depend on a Windows-only |
10 |
> > driver that only works with your motherboard and will prevent the RAID |
11 |
> > working if the motherboard files and you try to connect the drives to a |
12 |
> > different system. For some reason,this gives Windows users a warm, fuzzy |
13 |
> > feeling. |
14 |
> |
15 |
> Maybe you want pure hardware raid that's a pci(-e) slot that has a chip |
16 |
> doing all the parity calculations plus a battery keeping data that |
17 |
> didn't manage to write in case of a power failure. That's the best option. |
18 |
> For a mb in order to have fake raid capabilities from my understanding |
19 |
> it has to have a raid chip itself but still most of the calculation is |
20 |
> done by the CPU. |
21 |
> Software raid can be done OS dependant... and you don't need any |
22 |
> hardware or chips for that, it's a form of OS fooling itself instead of |
23 |
> the MB fooling the OS (fake one). |
24 |
> |
25 |
> From my experience with fake raid 0 on 2 hdd's the speeds would be very |
26 |
> nice when talking about small files, but for files with 1gb or more the |
27 |
> writing speed would decrease as the parity calculations get more |
28 |
> complicated. |
29 |
|
30 |
there is no parity with raid 0. |