Error detection/correction and fault detection/recovery – Data processing system error or fault handling – Reliability and availability
Reexamination Certificate
1999-03-29
2002-04-23
Iqbal, Nadeem (Department: 2184)
Error detection/correction and fault detection/recovery
Data processing system error or fault handling
Reliability and availability
C711S112000
Reexamination Certificate
active
06378084
ABSTRACT:
TECHNICAL FIELD
The invention relates to computers and computer peripherals. More particularly, the invention relates to a computer enclosure processor with failover capabilities.
BACKGROUND ART
SCSI (small computer systems interface) is a commonly used parallel I/O (input/output) system for computers. Computers can be interconnected with peripherals or modules along a SCSI bus or SCSI channel. A common interconnection arrangement is a daisy chain, where the bus extends out of one module into the next peripheral. The number of modules on a SCSI channel is bound by a fixed upper limit (e.g., 8 or 16, including the host computer). Although ANSI (American National Standards Institute) standards documents exist for SCSI, many variations are possible and proprietary adaptations are common. For example, a SCSI-
1
bus is a 50 conductor, single ended, parallel bus; a SCSI-
2
bus is an extension of SCSI-
1
to 68 conductors; and SCSI-
3
is a faster, low voltage differential version of SCSI-
2
. SCSI requires termination at the ends of the bus, and the terminations may be either active or passive and may be either internal or external to the peripheral. Finally, a variety of SCSI connectors can be used.
FIG. 1
illustrates a computer system and various peripheral devices interconnected by a single SCSI I/O bus. A computer
100
includes two internal SCSI devices
102
and
104
. The internal devices
102
and
104
might be, for example, a disk drive and a backup tape drive. Two external devices
106
and
108
are also connected to the SCSI bus. The external SCSI devices might be, for example, a printer and a scanner. In general, the SCSI bus system may have more or fewer devices. In the SCSI system illustrated in
FIG. 1
, the internal device
102
must provide a bus termination impedance. The internal devices
102
and
104
are typically connected by a ribbon cable with a single connector (for example,
110
) for each device. The external devices
106
and
108
are typically connected by a series of double ended cables. A first cable connects a connector
112
on the computer
100
to the external device
106
. A second cable connects the external device
106
to the external device
108
. The external device
108
has an open connector
114
(no cable attached) that may be terminated with an external terminator plug
116
(mandatory for Plug and Play SCSI) or may be terminated internally to the device
108
. The total length of a SCSI bus to a final termination must be less than a predetermined limit so as to ensure signal integrity along the entire bus.
Associated with each device on a SCSI bus is an adapter. An adapter is an interface between the SCSI channel and the device. For example, the computer
100
illustrated in
FIG. 1
includes a host adapter
118
. Peripheral devices
102
,
104
,
106
, and
108
each contain a peripheral adapter (not shown). An adapter may be physically packaged as a circuit card, board, or any other suitable form. In addition, one or more of the various peripherals may include SCSI controllers and other processors, which may be located on the same card or board as the adapter.
The host or host adapter is the initiator of the SCSI bus, and the other modules or peripherals are targets. The initiator commands the targets on the bus to read or write data from/to the bus. Each device on the bus has a unique address that the host initiator uses to direct commands and facilitate data transfer.
One or more SCSI peripheral devices may form a single SCSI enclosure. For example, a SCSI enclosure may be a collection of storage units in the same physical housing, sharing a common power supply and cooling system. The degree of collocation of devices within an enclosure can vary depending upon the needs of the overall computer system. For example, all peripherals in a room may be treated as an enclosure. On the other extreme, a single circuit board or IC (integrated circuit) chip may be logically partitioned into several SCSI devices and the board or chip may be treated as an enclosure. A single SCSI device by itself may be an enclosure. Typically, an enclosure defines a common environment in which there is high correlation of environmental conditions from device to device within the enclosure. However, an enclosure need not be so constrained, and one skilled in the art would appreciate that other considerations may warrant combining and treating as an enclosure one or more peripherals that do not share a common physical environment.
SCSI devices within a SCSI enclosure may retain their logical identity as separate, devices, each individually accessible by address by the host computer. On the other hand, the enclosure itself may manage and hide the underlying devices and present a single composite device (single address) to the host computer.
A SCSI enclosure may have associated with it an enclosure processor. An enclosure processor is a processor that performs monitoring and control of the enclosure. An enclosure processor typically monitors power supplies, cooling fans, doors, temperatures, and individual device locks. An enclosure processor may also control displays and a keypad or other switches on a front panel of the enclosure. U.S. Pat. No. 5,586,250, issued to Carbonneau et al. on Dec., 17, 1996, and hereby incorporated by reference, describes an enclosure processor using the terminology “SMARt (status monitoring and reporting) means” and “CMAC (cabinet monitor and control) board”.
An enclosure processor is typically a device on the SCSI bus, just like a peripheral device, with its own unique address. In this way, a host adapter can command and control an enclosure processor. To facilitate communications between a host adapter and an enclosure processor, a communications protocol may be implemented on top of the SCSI protocol. Two common protocols for use with an enclosure processor on a SCSI bus are the SAF-TE (SCSI accessible fault-tolerant enclosure) and SES (SCSI enclosure specification) protocols.
FIGS. 2-4
will next be described.
FIGS. 2-4
depict various configurations of a SCSI peripheral enclosure
200
with enclosure processors. These configurations illustrate problems that the present invention solves. In
FIGS. 2-4
, the peripherals are disk drives and the enclosure includes an array of disk drives, such as might be employed in a RAID (redundant array of inexpensive disks) or JBOD (just a bunch of disks) system.
FIGS. 2-4
depict a remote disk storage enclosure
200
containing four SCSI disk drives
210
-
213
on a first SCSI bus
220
and four SCSI disk drives
230
-
233
on a second SCSI bus
240
. Both buses
220
and
240
are attachable to the outside world through an adapter board
250
. The adapter board
250
contains a SCSI enclosure processor (SEP) on each SCSI bus—SEP
252
on the first bus
220
and SEP
254
on the second bus
240
. The adapter board
250
also contains one external host connection repeater on each SCSI bus—repeater
256
near the external connections to the first bus
220
and repeater
258
near the external connections to the second bus
240
. Depending upon how external connections are made to the adapter board
250
, the two SCSI bus systems of enclosure
200
can be utilized in a simplex, duplex, or cluster configuration, as will next be described.
FIG. 2
illustrates a simplex configuration, in which the first and second SCSI buses
220
and
240
are joined to form a single bus and a host computer
300
is attached to the single joined bus. The first SCSI bus
220
is joined to the second SCSI bus
240
by an external jumper cable
302
, which contributes additional length to the overall bus. The other external connection on the second SCSI bus
240
is capped by an external terminator
304
. In this simplex arrangement, the host computer
300
is able to access all eight disk drives
210
—
213
and
230
—
233
of the enclosure
200
. In this mode, the host computer is also able to access both SEPs
252
and
254
, because each has its own SCSI address. Typically, control of the enclosure
200
is
Erickson Michael J.
Strunk Glenn W.
Zilavy Daniel V.
Hewlett--Packard Company
Iqbal Nadeem
LandOfFree
Enclosure processor with failover capability does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Enclosure processor with failover capability, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Enclosure processor with failover capability will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2895478