AHCI and RAID on HP Compaq Elite 8000, 8100, 8200, and 8300 Business PCs - Technical White Paper

RAID 1 with two hard drives (Mirror))

Because it is a very cost-effective way to increase system storage reliability and a great value proposition, RAID 1 is the

only RAID configuration that HP pre-configures for HP Elite 8000-series Business PCs. RAID 1 provides high availability

with minimal performance impact, as well as greater reliability compared to a single hard drive configuration.

RAID 1 has redundancy and hence is a true RAID. It more than doubles reliability because the MTBF of RAID 1 is equal to

the MTBF of an individual drive multiplied by the number of drives (2). In other words, the probability of one hard drive

failure on a given day is the square root of that same probability. Hypothetically, if the chance for a single hard drive

failure is 1:2000, then the chance that both hard drives failing in RAID 1 is 1:4,000,000.

Mirroring, segmentation, and striping have no real meaning in RAID 1. In the table and graphic, the data is arranged in

rows for representation of different pieces of data.

Table 4: RAID 0 with two hard drives (Mirror)

First disk Second disk

Data Segment 1 Data Segment 1

Data Segment 2 Data Segment 2

Data Segment 3 Data Segment 3

Data Segment 4 Data Segment 4

Data Segment 5 Data Segment 5

Data Segment 6 Data Segment 6

Data Segment 7 Data Segment 7

Data Segment 8 Data Segment 8

In the previous table, each “Data Segment n” represents a group of data, known as a strip. In this case, each rows

represent a stripe. This table shows how information is duplicated in both hard drives. The size of the strips is mostly

irrelevant and not a configurable option.

To better illustrate the concept of RAID 1 and mirroring, Figure 3 shows how a sequence of data “ABCD...” is stored in a

RAID 1 volume. In this example, each letter represents a data segment. The graphic shows how the various pieces are

replicated for both of the hard drives; hence, if any one member of the RAID 1 volume fails, the information is kept in the

surviving members. After a hard drive failure, the user interface sends a notification so the failed hard drive can be

replaced. No user information is lost in this scenario.

Figure 3: Reliability - RAID 1 mirroring