Raid system with fibre channel arbitrated loop

• US 5,812,754 A
• Filed: 09/18/1996
• Issued: 09/22/1998
• Est. Priority Date: 09/18/1996
• Status: Expired due to Term

• Current Assignee: Hewlett Packard Enterprise Development LP (Hewlett-Packard Enterprise Company)
• Original Assignee: Silicon Graphics Incorporated (Graphics Properties Holdings, Inc.)
• Inventors: Oliver, James Wesley, Naminski, Ronald John, Lui, Albert S., Wood, Neill Preston, Aster, Radek
• Primary Examiner(s): Canney, Vincent P.
• Application Number:
  US08/718,743
• Time in Patent Office: 734 Days
• Field of Search: 395/182.04, 395/182.03, 395/182.05, 395/182.06, 395/441, 371/21.1, 371/21.6, 371/40.15
• US Class Current: 714/6.21
• CPC Class Codes: G06F 11/2007 using redundant communication media
• CPC Class Codes: G06F 11/2089 Redundant storage control functionality

Abstract

A modular and highly available RAID system has a fiber channel arbitrated loop (FC-AL) interface coupled with a disk array. Fault-tolerant operation is assured. The system provides dual and isolated arbitrated host and storage device loop circuits for redundant, independent input/output (I/O) paths to local and/or remote host computers. Each loop includes bypass circuits which prevent the failure of any device (host computer or storage device) from affecting the operation of loop. Orthogonal data striping may be used to further assure data integrity.

Claims

1. A controller for controlling communications between a plurality of host computer systems and a data storage subsystem, comprising:

a first host loop; and

a first plurality of port bypass circuits, each port bypass circuit connecting one of said plurality of host computer systems to said host loop, each port bypass circuit bypassing its associated host computer system if said associated host computer system becomes unavailable, wherein communication with remaining host computer systems is uninterrupted by said unavailable host computer system.

2. The controller of claim 1, further comprising:

a second host loop operatively isolated from said first host loop; and

a second plurality of port bypass circuits, each port bypass circuit connecting one of said plurality of host computer systems to said second host loop, each port bypass circuit bypassing its associated host computer system if said associated host computer system becomes unavailable, wherein communication with remaining host computer systems is uninterrupted by said unavailable host computer system,

wherein said second host loop provides a redundant communication path to said host computer systems.

3. The controller of claim 2, where each of said first and second host loops comprises:

a fiber channel arbitrated host loop.

4. A controller for controlling communications between a plurality of host computer systems and a plurality of storage devices, comprising:

a first drive loop; and

a first plurality of port bypass circuits, each port bypass circuit connecting one of said plurality of storage devices to said drive loop, each port bypass circuit bypassing its associated storage device if said associated storage device becomes unavailable, wherein communication with remaining storage devices is uninterrupted by said unavailable storage device.

5. The controller of claim 4, further comprising:

a second drive loop operatively isolated from said first drive loop; and

a second plurality of port bypass circuits, each port bypass circuit connecting one of said plurality of storage devices to said second drive loop, each port bypass circuit bypassing its associated storage device if said associated storage device becomes unavailable, wherein communication with remaining storage devices is uninterrupted by said unavailable storage device,

wherein said second drive loop provides a redundant communication path to said plurality of storage devices.

6. The controller of claim 5, where each of said first and second drive loops comprises:

a fiber channel arbitrated drive loop.

7. The controller of claim 6, further comprising:

a first environmental control bus for monitoring the status of said plurality of storage devices;

a first environmental bus control circuit for controlling said first plurality of port bypass circuits in response to input from said first environmental control bus.

8. The controller of claim 7, further comprising:

a second environmental control bus for monitoring the status of said plurality of storage devices, said second environment control bus being operably isolated from said first environmental control bus to provide redundancy;

a second environmental bus control circuit for controlling said second plurality of port bypass circuits in response to input from said second environmental control bus.

9. A modular and highly available fibre channel arbitrated loop system for ensuring fault-tolerant operation between a plurality of host computer systems and a data storage array having a plurality of storage devices, the system comprising:

a first controller having a first arbitrated host loop coupling said first controller to a first host computer system;

a first arbitrated drive loop coupling said first controller to the plurality of storage devices;

a second controller having a second arbitrated host loop coupling said second controller to a second host computer system;

a second arbitrated drive loop coupling said second controller to the plurality of storage devices, said second arbitrated host and second arbitrated drive loops being operatively isolated from said first arbitrated host and first arbitrated drive loops;

host bypass means, operably connected between each of said arbitrated host loops and each host computer system, for bypassing an associated host computer system upon failure of said associated computer system, thereby ensuring continued operation of each of said arbitrated host loops; and

drive bypass means, operably connected between each of said arbitrated drive loops and each of said storage devices, for bypassing an associated storage device upon failure of said associated device, thereby ensuring continued operation of each of said arbitrated drive loops.

10. A RAID subsystem comprising:

a plurality of disk modules, each disk module having a plurality of disk assemblies;

a first controller having a first drive loop for communicating with said plurality of disk modules;

a second controller having a second drive loop for communicating with said plurality of disk modules, wherein said second drive loop is operably isolated from said first drive loop;

a plurality of first bypass circuits, each first bypass circuit connecting said first drive loop to one of said disk assemblies; and

a plurality of second bypass circuits, each second bypass circuit connecting said second drive loop to one of said disk assemblies.

11. The RAID subsystem of claim 10, wherein each controller further comprises:

a plurality of third bypass circuits interposed in said first and second drive loops for bypassing a failed disk module.

12. The RAID subsystem of claim 11, further comprising:

a plurality of orthogonal logical units for storing data, each logical unit being formed from not more than one disk assembly of each disk module, wherein each of said orthogonal logical units can assure data integrity despite a failure of one of said disk modules.

13. A method for operating an array of storage devices to store data, the method comprising the steps of:

organizing the plurality of storage devices into a plurality of storage modules so that each of said storage modules includes a plurality of storage devices, wherein the storage devices of each storage module share common power and cooling resources;

forming logical units orthogonally across said storage modules, each logical unit including not more than one storage device from each of said storage modules; and

reading and writing blocks of data to/from said logical units, wherein data integrity is assured in the event of failure of any one storage module.