Krishnakumar Ra Surugucchi

6687765, Feb 3, 2004

Jan 16, 2001

09/764614

Krishnakumar Rao Surugucchi - Fremont CA
Bruce M. Cassidy - Pleasanton CA

International Business Machines Corporation - Armonk NY

G06F 1314

710 15

Structure, method, and computer program for an explicitly tunable device controller. Method supports high-performance I/O without imposing additional overhead during normal input/output operations. Tuning is performed during explicit pre-I/O operation phase. In one embodiment, invention provides a method for tuning device controller operating characteristics to suit attributes of a data stream in which the method comprises: monitoring a data stream and collecting attributes of the monitored data stream; generating performance metrics of the data stream based on the collected attributes and a plurality of different assumed device controller configurations; comparing expected performance of the plurality of different device controller configurations for effectiveness with a future data stream having similar data stream type attributes to the monitored data stream; and selecting device controller characteristics to provide an effective match between the data stream type and the device controller configuration. In one embodiment, the controller configuration is adjusted automatically and dynamically during normal I/O operations to suit the particular input/output operation encountered. Configuration information may be selected for example, from such parameters as data redundancy level, RAID level, number of drives in a RAID array, memory module size, cache line size, direct I/O or cached I/O mode, read-ahead cache enable or read-ahead cache disable, cache line aging, cache size, or combinations thereof.

6851005, Feb 1, 2005

Mar 3, 2000

09/518551

Govindaraju Gnanasivam - Milpitas CA, US
Krishnakumar Rao Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F 1322

710220, 710200, 710240, 710107, 710112, 710113, 710 36

Apparatus and methods are provided for efficiently implementing logical-device reservations in a cluster computer system. The apparatus includes cooperating controllers programmed in firmware around a distributed reservation table. The apparatus manages access to a logical device, with first and second nodes with respective bus controllers communicatively coupled to each other and to a logical device by means of a bus. The first controller receives a request to reserve the logical device and, in response, communicates a reservation request for the logical device over the bus to the second controller for execution by the second controller. In response to the communicated reservation request, the second controller reserves the logical device for the first node.

6928509, Aug 9, 2005

Aug 1, 2002

10/211109

Krishnakumar Rao Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F013/00

710316, 710312, 710314, 710315, 710317, 710305, 711111

A method, system and apparatus for providing inter-connective access of a plurality of controllers to a plurality of serial storage devices are provided. Serial storage devices are provided with a serial operative connection to a data communication bridge. The bridge is operatively coupled to a plurality of controllers. The plurality of controllers is provided concurrent targeted connections to the set of serial storage devices. In one embodiment, InfiniBand technology further increases the scalability and enhances the reliability of a data communication system provided with a plurality of (S-ATA) storage devices. The reliability of the data communication system is enhanced because if one controller should fail, another controller may still achieve and maintain access to the plurality of serial storage devices.

7073022, Jul 4, 2006

May 23, 2002

10/155315

Mohamad H. El-Batal - Westminster CO, US
Yoshihiro Fujie - Fujisawa, JP
Thomas Sing-Klat Liong - San Jose CA, US
Krishnakumar Rao Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F 12/00

711114

The present invention describes a method and system for interfacing a plurality of device controllers to an array of data storage devices by serial connection. The device controllers are coupled to a serial interface by a bus and the devices of the storage array are coupled to the serial interface by a serial connection. The serial interface receives controller signals through the bus and multiplexes the signals onto the serial connections of the storage array. Arbitration between the various device controllers seeking access to the storage array is resolved through bus protocol and through drive based reserve/release registers in the serial interface processor.

7216207, May 8, 2007

Mar 8, 2004

10/795648

Lawrence Yiumchee Chiu - Cupertino CA, US
Patrick John Cozzi - Drexel Hill PA, US
Patrick Randolph Eaton - Lafayette CA, US
Prasenjit Sarkar - San Jose CA, US
Krishnakumar Surugucchi - Fremont CA, US
Kaladhar Voruganti - San Jose CA, US

International Business Machines Corporation - Armonk NY

G06F 12/12
G06F 12/14

711159, 711163, 726 26

A system, program storage device, and method of optimizing data placement on a storage device, the method comprising establishing a specified time constraint for which the storage device is to delete data stored thereon; dividing a data object into a plurality of data bits; programming a block of data and the data bits with a logic operand if the storage device is incapable of deleting the data within the specified time constraint; creating an encoded block of data from the programmed block of data and the data bits; organizing the encoded block of data and the data bits in the storage device according to data deletion requirements; and removing the data bits from the storage device if the data bits are organized within a specified data deletion requirement, wherein the data bits are removed using a data shredding process, and wherein the logic operand comprises an exclusive-or (XOR) operator.

7536588, May 19, 2009

Feb 8, 2006

11/350247

James Lee Hafner - San Jose CA, US
Krishnakumar Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F 11/00

714 6, 714 5, 714 20

A bladed storage servicing system comprising a RAID layout facilitates the removal of one or more functional storage devices in a tray of the bladed storage subsystem in order to service a failed storage device. Upon removal of the tray from the bladed storage subsystem, the system designates the functional storage devices on the tray as temporarily removed. The system logs writes to the removed functional storage devices so that the functional storage devices can be returned to consistency when replaced in the bladed storage subsystem. The system manages reads to the removed functional storage devices through reconstruction from active storage devices in the bladed storage subsystem. Upon reinsertion of the tray in the bladed storage subsystem, the system verifies the identities of the functional storage devices and resynchronizes the functional storage devices to ensure data consistency.

7793167, Sep 7, 2010

Aug 23, 2007

11/843791

James L. Hafner - San Jose CA, US
Carl E. Jones - Tucson AZ, US
David R. Kahler - Tucson AZ, US
Robert A. Kubo - Tucson AZ, US
David F. Mannenbach - Tucson AZ, US
Karl A. Nielsen - Tucson AZ, US
James A. O'Conner - Ulster Park NY, US
Krishnakumar R. Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F 11/00

714 54, 714 42, 714805, 370248

Methods are provided for detecting and correcting dropped writes in a storage system. Data and a checksum are written to a storage device, such as a RAID array. The state of the data is classified as being in a “new data, unconfirmed” state. The state of written data is periodically checked, such as with a timer. If the data is in the “new data, unconfirmed” state, it is checked for a dropped write. If a dropped write has occurred, the state of the data is changed to a “single dropped write confirmed” state and the dropped write error is preferably corrected. If no dropped write is detected, the state is changed to a “confirmed good” state. If the data was updated through a read-modified-write prior to being checked for a dropped write event, its state is changed to an “unquantifiable” state.

7853823, Dec 14, 2010

Aug 6, 2009

12/537067

Veera W. Deenadhayalan - Fremont CA, US
James Lee Hafner - San Jose CA, US
Krishnakumar Rao Surugucchi - Fremont CA, US

International Business Machines Corporation - Armonk NY

G06F 11/00

714 6, 714 2, 714770

A system (and method) for determining reconstruction formulas for partial strip reconstruction in a storage system in which a plurality of lost strips have been detected, includes using a combination of a direct reconstruction method and a sequential reconstruction method.