System for I/O path load balancing and failure which can be...

Details System for I/O path load balancing and failure which can be... System for I/O path load balancing and failure which can be...

1999-03-25

2002-01-22

Ray, Gopal C. (Department: 2181)

Error detection/correction and fault detection/recovery

Data processing system error or fault handling

Reliability and availability

C714S043000, C709S241000, C711S203000, C711S114000

Reexamination Certificate

active

06341356

ABSTRACT:

FIELD OF INVENTION
The present invention relates generally to the use of peripheral devices in a computer system and more particularly to a system and method for balancing the loads on channel paths in a multiple host operating system environment.
BACKGROUND OF THE INVENTION
Many, if not most, computer systems provide for linking a host computer operating system to one or more peripheral devices which are used as data storage media. For example, the AIX and OS/2 operating systems made by International Business Machines Corporation of Armonk, N.Y. can interface with several different types of storage devices that have a so-called Small Computer System Interface (SCSI) interface format.
Each SCSI contains input/output (I/O) paths to each of its attached data storage devices. Included among SCSI devices are magnetic tape data storage devices, magnetic disk data storage devices, and optical disk data storage devices. Also included among SCSI devices are medium changer library devices, each of which contains several data storage devices. In medium changer devices, one of the data storage disks at a time can be selected and then engaged with a transparent element within the device for accessing data of a data storage device.
In a typical file server, the storage needs typically exceed the capacity of current hard disks, and thus many file servers use multiple disks to provide the necessary storage capacity. A typical disk drive storage configuration uses a redundant array of inexpensive disks, referred to as a RAID configuration, whereby the drives (data storage devices) are linked together through hardware to form a drive array.
FIG. 1
depicts a conventional data transfer system, generally designated
10
. As shown, the conventional system
10
includes a computer
12
with associated operating system
14
. The conventional system
10
includes a plurality of computer peripheral device drivers
16
,
18
,
20
for controlling data storage devices
22
,
24
,
26
, respectively. More particularly, the system includes first, second, and third device drivers
16
,
18
,
20
for respectively controlling first, second, and third data storage devices
22
,
24
,
26
. It is to be understood that the conventional system
10
can include greater or fewer device drivers. The data storage devices
22
,
24
,
26
in the array are coordinated with each other and information is allocated between them. In this manner, a number of individual hard disks are combined to create a massive virtual system.
In accordance with principles well-known in the art, each device driver
16
,
18
,
20
receives data I/O requests from an application running on the operating system
14
. Further, each device driver
16
,
18
,
20
executes each I/O request by issuing an appropriate command sequence to its associated data storage device
22
,
24
,
26
to cause the device
22
,
24
,
26
to transfer data in accordance with the I/O request. Thus, each device driver
16
,
18
,
20
is an intermediary between the operating system
14
and the associated data storage device
22
,
24
,
26
, respectively. Stated differently, each device driver
16
,
18
,
20
functions as a data transfer controller between the operating system
14
and the associated data storage device
22
,
24
,
26
.
One aspect of a RAID storage system is its ability to account for failures using redundancy. To decrease the possibility of losing data stored in a particular drive, each disk drive in the RAID can have multiple paths connected to it.
FIG. 2
depicts a conventional operating system
100
with multiple paths
108
connecting the device driver
102
to the respective data storage devices in the RAID
110
. However, the conventional operating system
100
cannot recognize multiple paths to a single device. In this configuration, the conventional operating system
100
operates as though each path is connected to a different device, as opposed to multiple paths connected to a single device. Although conventional solutions to these problems exist, these conventional solutions are typically incorporated directly into the operating system
100
. These conventional solutions are platform specific and thus not portable from one operational environment to another. Therefore, these conventional solutions can only be utilized by a single operating system
100
.
Today, many enterprises have a multiplicity of host computer systems that either function independently or are connected through a network. It is desirable that each system in the multiple host system environment provide the redundancy associated with multiple paths connected to each disk drive. Each system must therefore be capable of managing the multiple paths for each drive. This capability should be easily provided to each operating system in the multiple system environment.
SUMMARY OF THE INVENTION
The present invention provides a method and system for path management of data processing in an environment which includes a plurality of hosts running on multiple operating systems. The method and system comprises providing a portable path management code and providing a plurality of control modules. Each of the control modules has an area which can only be accessed by the portable path management code. The method and system also comprises providing an interface to at least one device driver from the portable management code to allow the portable management code to control access to plurality of paths to a plurality of storage subsystems by the at least one device driver.
The method and system in accordance with the present invention can provide path load balancing and fail over for a plurality of paths to a plurality of storage subsystems. The path management code is portable and platform independent so that the functionality can be ported easily to each operating system in the multiple system environment.


REFERENCES:
patent: 4589063 (1986-05-01), Shah et al.
patent: 5239649 (1993-08-01), McBride et al.
patent: 5442789 (1995-08-01), Baker et al.
patent: 5557740 (1996-09-01), Johnson et al.
patent: 5586268 (1996-12-01), Chen et al.
patent: 5623696 (1997-04-01), Johnson et al.
patent: 5680539 (1997-10-01), Jones
patent: 5768623 (1998-06-01), Judd et al.
patent: 5790775 (1998-08-01), Marks et al.
patent: WO9826553 (1998-06-01), None
patent: WO9828686 (1998-07-01), None
IBM Technical Disclosure Bulletin: vol. 38, No. 7, Jul. 1995, “Dynamic Load Sharing for Distributed Computer Environment,” pp. 511-515.

Affiliated with

Also associated with

Rating

System for I/O path load balancing and failure which can be... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with System for I/O path load balancing and failure which can be..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System for I/O path load balancing and failure which can be... will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2826550