Technical information

About This DIGITAL HiTest Suite
HiTest Notes for Lotus Domino Windows NT Prioris MX
2–3
− When three drives are used for data and high availability is required, DIGITAL
recommends the use of RAID 5 providing approximate storage capacity of 8.6 GB.
Understanding Availability Features
This section provides background information on the availability features included in this
HiTest Suite.
Backup and Restore
Backup and restore ensures that data remains available from one day to the next. It is
generally identified as a data protection technique because the stored information can also be
removed to a remote, protected environment. DIGITAL offers a range of backup and restore
capabilities from individual tape systems to automated tape libraries.
Disk Storage Technologies
This section describes the disk storage technologies used to provide availability for DIGITAL
HiTest configurations.
Just a Bunch of Disks (JBOD)
Just a bunch of disks (JBOD) refers to a multiple disk drive configuration, internal or external
to a host computer, in which there is no storage controller. Disk drives are managed by the
host system. To increase availability, JBOD storage systems are often configured with
hardware such as redundant power supplies and fans, or multiple SCSI buses.
Redundant Array of Independent Disks (RAID)
A Redundant Array of Independent Disks (RAID) is a collection of disks managed by
specialized array management software. When using RAID, all disks in the RAIDset should
be the same type. Array management software may be host-based (execute in the host
computer) or subsystem-based (execute in an intelligent disk controller).
1
Disk striping (RAID Level 0), is technically not RAID because it does not offer redundancy.
RAID Levels 2 and 3 are parallel access arrays (members are accessed concurrently). To
ensure that all disks participate in every I/O request, the minimum chunk size is kept small
(for example, a byte).
RAID Levels 4 and 5 are independent access arrays (members are not required to be accessed
concurrently). By keeping the minimum chunk size at least as large as a disk sector (block),
not all members have to participate in each I/O request.
RAID Levels 2 and 4 are not in general use.
Figure 2-1 summarizes the RAID levels. The shaded areas in Figure 2-1 refer to space used
for redundancy features.
1
The RAID technique was described by D. A. Patterson, G. Gibson, and R. H. Katz “A Case for
Redundant Arrays of Inexpensive Disks (RAID),” Report No. UCB/CSD 87/391, University of
California, Berkeley CA 1987.