5.3 Fibre Channel RAID

Redundant arrays of independent disks, or RAIDs, are a well-established technology in traditional storage applications. RAID standards define methods for storing data to multiple disks and imply intelligence in the form of a RAID controller. The controller can be implemented in software for example, as an application running on a file server but typically is a dedicated card installed in a RAID storage enclosure. RAID storage configurations may include eight to ten disks for a departmental-level array, or many more for enterprise applications.

When a server is connected to a single disk drive, reads or writes of multiple data blocks are limited by the buffering capability and rotation speed of the disk. While the disk is busy processing one or more blocks, the host must wait for acknowledgment before sending or receiving more data. You can increase throughput by dispersing data blocks across several disks in a RAID, a technique called striping. In a write operation, for example, the host can send fewer data blocks to multiple targets consecutively and thereby avoid swamping the capacity of any individual drive. This simplified RAID is called level 0. Although it boosts performance, it does not provide data security. If a single disk fails, data cannot be reconstructed from the survivors.

Other RAID levels introduce data integrity by either writing parity data to a dedicated drive or writing parity information on each drive in an array. RAID level 3 writes byte-level parity to a dedicated drive. RAID level 4 writes block-level parity to a parity drive. Because a dedicated drive contains the information required to reconstruct data, levels 3 and 4 pose a data integrity problem if one of the data drives and the parity drive itself fail. RAID level 5 addresses this problem by striping block-level parity information across each drive. If an individual drive fails, it can be reconstructed from the parity information contained on the other drives. Another technique, level 1, achieves full data security by sacrificing the performance gain of striping in favor of simple disk mirroring. In disk mirroring, every write operation to a primary disk is duplicated on a secondary, or mirrored, disk. If the primary disk fails, the server can switch over to the backup disk. Some RAID techniques may be combined, as in level 0+1, to provide both striping throughput and data redundancy via mirroring, as shown in Figure 5-5.

Figure 5-5. RAID 0+1 striping plus mirroring of data to two arrays

graphics/05fig05.gif

The various techniques that RAID provides for redundancy and speed can be implemented by a dedicated RAID controller housed in the same enclosure as the disks, or by a RAID controller provisioned in the host system or file server. Data is passed from the operating system to the RAID controller, which then manages the striping or mirroring task. If the operating system itself manages data striping or mirroring, the host CPU must devote cycles to the task. The latter technique is called software RAID. In the next section, you will see how software RAID is implemented on JBODs.

Fibre Channel-attached RAID enclosures use an integrated RAID controller that sits behind the Fibre Channel interface. The RAID controller receives SCSI-3 requests via the loop or fabric and stores or retrieves data from the array using the appropriate RAID levels. From the standpoint of the Fibre Channel interconnect, the physical interface between the RAID controller and its disks is irrelevant. The RAID controller appears as an N_Port or NL_Port to the outside world but can use a proprietary bus, a SCSI bus, or other architecture to talk to its drives, as shown in Figure 5-6. For performance and reliability, however, RAID manufacturers can also incorporate Fibre Channel technology behind the RAID controller. This allows the use of Fibre Channel drives and the redundant topologies provided by, for example, dual-loop configurations. The use of arbitrated loop between the RAID controller and its Fibre Channel disks is invisible to the SAN, because the RAID controller still appears as a single N_Port or NL_Port to the rest of the topology.

Figure 5-6. RAID with internal SCSI bus and internal arbitrated loop RAID

graphics/05fig06.gif

Because a Fibre Channel RAID controller must support the functionality of an HBA and independent intelligence to manage striping and mirroring functions, it brings additional cost to a storage solution. Fibre Channel RAID products justify the additional expense by including high-availability and storage management features required for enterprise networking. RAID enclosures offer redundant, hot-swappable power supplies, redundant fans, and self-diagnostics. Most of these products accommodate on-line spare drives for automatic reconstruction of data if a disk failure occurs. Large systems can be scaled to storage ranging from hundreds of gigabytes to tens of terabytes, and that implies an enterprise-level budget.

For mission-critical SANs, Fibre Channel RAIDs offer a number of advantages. The advanced diagnostics support, management features, reliability, and scalability typical of such high-end systems fill in most of the checkboxes for enterprise storage selection criteria. Server processing resources benefit by offloading data redundancy tasks to the RAID controller. Additionally, when the server is freed from the overhead of software RAID, it is also liberated from exclusive ownership of the storage device. No single server is responsible for managing the striping of data to multiple disks, so any server may own part or all of the RAID controller's data. Fibre Channel RAID thus enables server clustering and other applications that are predicated on co-ownership of storage resources.

From a SAN design standpoint, Fibre Channel RAIDs simplify configurations because the RAID controller appears as a single port address to the loop or fabric. Adding drives to a RAID enclosure does not add to the population of the SAN transport, even if the RAID enclosure internally uses arbitrated loop to access its disks. For arbitrated loops, you can attach more than a terabyte of storage without consuming all available AL_PAs. For direct fabric attachment, the RAID controller and all its storage appear as a single N_Port on its own 100MBps segment, whereas a single JBOD enclosure will appear as multiple NL_Ports sharing a common bandwidth. However, JBODs do have one distinct advantage over RAIDs: cost.



Designing Storage Area Networks(c) A Practical Reference for Implementing Fibre Channel and IP SANs
Designing Storage Area Networks: A Practical Reference for Implementing Fibre Channel and IP SANs (2nd Edition)
ISBN: 0321136500
EAN: 2147483647
Year: 2003
Pages: 171
Authors: Tom Clark

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net