Intel AS/400 RISC Server, 170 Servers, 7xx Servers manual Generic Configuration Concepts

14.5.1.2 Generic Configuration Concepts

There are several important principles to keep track of in terms of getting good performance. Most of the following are issues when the disks are configured. A great many problems can be eliminated (or, created) when the drives are originally configured. The exact nature of some of these difficulties might not be easily predicted. But, much of what follows will simply avoid trouble at no other cost.

1.Ensure that RAID protection is performed as close to the physical device as possible . This is typically done out at an I/O adapter level or on the external disk array product. This means that either the external disk's configuration tools or (for internal disks assigned to VIOS) VIOS' tools will be used to create RAID configurations (RAID5, RAID10, or RAID1). When this is done, as far as IBM i operating system disk status displays are concerned, the resulting virtual drives appear to be "unprotected." It might be superficially reassuring to have IBM i operating system do the protection (if IBM i operating system even permits it). WRKDSKSTS would then show the protection on that path. DST/SST disk configuration functions would show the protection, too. However, it is better to put up with what appears to IBM i operating system's disk status routines to be unprotected devices (which are, after all, actually protected) than to take on the performance problems of doing this under IBM i operating system. RAID recovery procedures will have to be pursued outside of IBM i operating system in any event, so the protection may as well go where the true physicality is understood (either in VIOS or the external disk array product).

Note also that you also want to configure things so that the outboard devices, rather than VIOS, do the RAID protection whenever possible. This enables I/O to flow directly from the device to IBM i operating system as directed by VIOS.

High Availability scenarios also need to be considered. In some cases, to enable appropriate redundancy, it may be necessary to do the protection a little farther away from the device (e.g. spread over a couple of adapters) so as to enable the proper duplexing for high availability. If this applies to you, consult the documentation. Some external storage devices have extensive duplexing within themselves, for instance, which could allow one to keep the protection close to the device after all.

2.Recognize that Internal Disks remain the "gold standard" for performance. We have consistently measured external disks as having less performance than 520 byte, internally attached disks. However, the loss of throughput, with proper configuration, is not a major concern. What is harder to control is response time. If you have sensitivity to response time, consider internal disks more strongly.

IBM i 6.1 Performance Capabilities Reference - January/April/October 2008
© Copyright IBM Corp. 2008	Chapter 14 DASD Performance	217