公开(公告)号：AT343277T

公开(公告)日：2006-11-15

申请号：AT04742012

申请日：2004-08-04

Applicant: IBM

Inventor： ELKO DAVID ARLEN ,  LEPORE DANIEL ,  MEHTA CHETAN ,  PFISTER GREGORY FRANCIS ,  SUGRUE PATRICK JOHN

IPC: H04L12/24

Abstract: A method, system, and product in a data processing system are disclosed for providing centralized management of an INFINIBAND distributed system-area network that includes multiple end nodes. A manager application is established in one of the end nodes. An agent application is established in one or more end nodes. Each agent application is independent from the manager application. The manager application maintains a current list of active agent applications and uses the list to manage the agent applications in the end nodes.

公开(公告)号：BRPI0413589A

公开(公告)日：2006-10-17

申请号：BRPI0413589

申请日：2004-08-04

Applicant: IBM

Inventor： ELKO DAVID ARLEN ,  LEPORE DANIEL ,  MEHTA CHETAN ,  PFISTER GREGORY FRANCIS ,  SUGRUE PATRICK JOHN

IPC: H04L12/24

Abstract: A method, system, and product in a data processing system are disclosed for providing centralized management of an INFINIBAND distributed system-area network that includes multiple end nodes. A manager application is established in one of the end nodes. An agent application is established in one or more end nodes. Each agent application is independent from the manager application. The manager application maintains a current list of active agent applications and uses the list to manage the agent applications in the end nodes.

公开(公告)号：CA2576267A1

公开(公告)日：2005-11-03

申请号：CA2576267

申请日：2005-03-17

Applicant: IBM

Inventor： MEHTA CHETAN ,  GRAHAM CHARLES SCOTT ,  SRIKRISHNAN JAYA ,  KIEL HARVEY GENE ,  BILLAU RONALD LEROY

IPC: G06F9/50 ,  G06F9/40 ,  G06F9/46

Abstract: At least one input/output (I/O) firmware partition is provided in a partitioned environment to facilitate access to I/O resources owned by the at least one I/O firmware partition. The I/O resources of an I/O firmware partition are shared by one or more other partitions of the environment, referred to as consumer partitions. The consumer partitions use the I/O firmware partition to access the I/O resources. Since the I/O firmware partitions are responsible for providing access to the I/O resources owned by those partitions, the consumer partitions are relieved of this task, reducing complexity and costs in the consumer partitions.

公开(公告)号：CA2283046C

公开(公告)日：2003-07-15

申请号：CA2283046

申请日：1999-09-23

Applicant: IBM

Inventor： FORSMAN STEPHANIE M ,  PATEL MAULIN I ,  HAMILTON RICK A II ,  MEHTA CHETAN

IPC: G06F9/455 ,  G06F9/48 ,  G06F9/46

Abstract: To emulate multi-threaded processing in an operating system supporting only single-threaded processes and single-level interrupts, the processor timer is started with a selected time-out period during execution of a master code thread. Processing of the master code thread proceeds until the timer interrupt, at which time the operating system timer interrupt service routine (ISR) transfers execution control to a slave code thread or slave code thread component. The slave code thread or component is executed in its entirety, at which time the timer is reset and execution control is returned to the master code thread, where processing resumes at the point during which the timer interrupt was asserted. To minimize disruption of the master code thread execution, a maximum latency should be enforced on the slave code thread, which may be accomplished by breaking the slave code thread into multiple components. The timer ISR maintains an index of the predetermined starting points within the slave code thread(s) with a pointer identifying the next slave code thread component to be selected when the timer interrupt is asserted. Processing thus alternates between the master code thread and the slave code thread or components, with different slave code thread components being selected in round-robin fashion. The duty cycle between the master code thread and the slave code thread or components may be varied by selection of the time- out period and the maximum latency allowed to slave code thread processing.