|
Lines 53-59
Link Here
|
| 53 |
addresses these three problems. Let us look at them in more detail. Various |
53 |
addresses these three problems. Let us look at them in more detail. Various |
| 54 |
solutions to these problems have been proposed and implemented:</para> |
54 |
solutions to these problems have been proposed and implemented:</para> |
| 55 |
|
55 |
|
| 56 |
|
|
|
| 57 |
<para>Disks are getting bigger, but so are data storage requirements. |
56 |
<para>Disks are getting bigger, but so are data storage requirements. |
| 58 |
Often you will find you want a file system that is bigger than the disks |
57 |
Often you will find you want a file system that is bigger than the disks |
| 59 |
you have available. Admittedly, this problem is not as acute as it was |
58 |
you have available. Admittedly, this problem is not as acute as it was |
|
Lines 89-95
Link Here
|
| 89 |
fastest drives spin at 15,000 rpm, so the average rotational latency |
88 |
fastest drives spin at 15,000 rpm, so the average rotational latency |
| 90 |
(half a revolution) is 1.75 ms. At 30 MB/s, the transfer itself takes about |
89 |
(half a revolution) is 1.75 ms. At 30 MB/s, the transfer itself takes about |
| 91 |
150 μs, almost nothing compared to the positioning time. In such a |
90 |
150 μs, almost nothing compared to the positioning time. In such a |
| 92 |
case, the effective transfer rate drops to a little over 1 MB/s and is |
91 |
case, the effective transfer rate drops to a little over 1 MB/s and is |
| 93 |
clearly highly dependent on the transfer size.</para> |
92 |
clearly highly dependent on the transfer size.</para> |
| 94 |
|
93 |
|
| 95 |
<para>The traditional and obvious solution to this bottleneck is |
94 |
<para>The traditional and obvious solution to this bottleneck is |
|
Lines 233-239
Link Here
|
| 233 |
<para><indexterm><primary>RAID-5</primary></indexterm>An alternative |
232 |
<para><indexterm><primary>RAID-5</primary></indexterm>An alternative |
| 234 |
solution is <emphasis>parity</emphasis>, implemented in the |
233 |
solution is <emphasis>parity</emphasis>, implemented in the |
| 235 |
<acronym>RAID</acronym> levels 2, 3, 4 and 5. Of these, |
234 |
<acronym>RAID</acronym> levels 2, 3, 4 and 5. Of these, |
| 236 |
<acronym>RAID-5</acronym> is the most interesting. As implemented |
235 |
<acronym>RAID-5</acronym> is the most interesting. As implemented |
| 237 |
in Vinum, it is a variant on a striped organization which dedicates |
236 |
in Vinum, it is a variant on a striped organization which dedicates |
| 238 |
one block of each stripe to parity of the other blocks: As implemented |
237 |
one block of each stripe to parity of the other blocks: As implemented |
| 239 |
by Vinum, a <acronym>RAID-5</acronym> plex is similar to a |
238 |
by Vinum, a <acronym>RAID-5</acronym> plex is similar to a |