Reliable enclosure services in a dual loop enclosure
Original Publication Date: 2000-Jul-01
Included in the Prior Art Database: 2003-Jun-12
Peripheral devices, such as disk drives, can be connected to a host computer using a uni-directional loop. One example of such a connection is Fibre Channel Arbitrated Loop (FC-AL). In a uni-directional loop, a message between any two nodes and its reply will use the entire loop, and thus a single break in the loop will prevent all useful communication. Thus to continue operation if a link between nodes fails, it is necessary to provide two independent loops to all nodes. FC-AL disk drives are made with dual interfaces, to allow attachment to both loops, and will present the same loop address on both loops. For reliable access to data, disk drives can be used in a redundant array of independent disks (RAID), which will allow continued access all data if one disk fails. Data access is thus preserved after any single failure, of disk or loop. When disk drives are packaged in an enclosure, it is common to provide an additional node on the loop to provide enclosure services setting and querying the state of the enclosure. It is desirable to make this also continue operation in the event of a single failure. The host computer software expects to find one enclosure services node on the loop, but if a single node is provided with dual loop interfaces, like a disk drive, the failure of this node will cause the enclosure services to be unavailable. This invention provides two physical enclosure services nodes, one on each loop. They are connected together by a communications route separate from the two main loops. This enables them to work together to provide the image to the host computer of a single node, at the same address on each loop. Commands and queries can be sent to either, with exactly the same effect. If one node (or one loop) fails, full function operation continues by using the other.