International Business Machines Corporation

David B Emmes

Jeffrey David Aman

Donald W Schmidt

Steven Jay Greenspan

Joshua W Knight

Alan Michael Webb

John L Czukkermann

Roger E Bonsteel

Donna N Dillenberger

Juliet C Candee

Curt Lee Cotner

Donald Fred Ault

Ernest Scott Bender

Michael Gary Spiegel

Michael Wayne Brown

Elpida Tzortzatos

David Bruce Lindquist

Rajesh Ramkrishna Bordawekar

Geoffrey Owen Blandy

Casper A Scalzi

Kenneth E Plambeck

Carl Edward Clark

Christopher Meyer

Danny R Sutherland

Richard J Schmalz

David Walsh Palmieri

Tri M Hoang

Mariama Ndoye

Mark Alvin Sehorne

Ronald F Hill

Tracy D Butler

Steven Edward Rosengren

Frank S Samuel, Jr

Donna N T E Dillenberger

Thomas J Szczgielski

Satwinder S Kahlon

Juilet C Candee

Donna Ngar-Ting Dillenberger

A system for porting code from a native platform to a non-native platform is provided which utilizes a non-native, platform-specific layer of code and a platform-neutral layer of code. The platform-neutral layer substantially emulates the native platform of ported code and provides a platform-neutral interface to the ported code. Together, the platform-specific layer and the platform-neutral layer provide an execution environment in which the ported code operates as a kernel extension of the non-native platform. The platform-neutral layer of the execution environment is portable to other non-native platforms so that code can be ported to another non-native platform by replacing the platform-specific layer with a customized platform-specific layer, which is adapted to the other non-native platform.

System and program product for porting code utilizing a portable execution environment

A system and program product for facilitating computing-environment control of a resource to be accessed by existing program code is provided which includes transparently replacing an uncontrolled resource referenced by existing program code with a resource controlled by a computing environment. The protocol also includes transparently interposing at least one resource management service of the computing environment between the existing program code and the resource controlled by the computing environment, wherein the at least one resource management service provides computing-environment control of the resource to be accessed by the existing program code. The replacing of the uncontrolled resource prevents the existing program code from directly accessing the resource.

System and program product for providing computing-environment control of a resource to be accessed by program code

A method of facilitating computing-environment control of a resource to be accessed by existing program code is provided which includes transparently replacing an uncontrolled resource referenced by existing program code with a resource controlled by a computing environment. The method also includes transparently interposing at least one resource management service of the computing environment between the existing program code and the resource controlled by the computing environment, wherein the at least one resource management service provides computing-environment control of the resource to be accessed by the existing program code. The replacing of the uncontrolled resource prevents the existing program code from directly accessing the resource.

Method for providing computing-environment control of a resource to be accessed by program code

A method, system, and storage medium for the InfiniBand&#x2122; Poll verb to support a multi-threaded environment without the use of kernel services to provide serialization for mainline Poll logic. Poll is the verb, which allows a consumer to determine which of its work requests have completed, and provides ending status. In addition to multiple concurrent threads using Poll against a single Completion Queue, Poll is serialized with Destroy Queue Pair and Destroy Completion Queue. Completion Queues are used to maintain completion status for work requests. Queue Pairs are used to submit work requests and are related to a Completion Queue at the time they are created.

Method, system and storage medium for lockless InfiniBand&#x2122; poll for I/O completion

A method of porting code from a native platform to a non-native platform is provided which utilizes a non-native, platform-specific layer of code and a platform-neutral layer of code. The platform-neutral layer substantially emulates the native platform of ported code and provides a platform-neutral interface to the ported code. Together, the platform-specific layer and the platform-neutral layer provide an execution environment in which the ported code operates as a kernel extension of the non-native platform. The platform-neutral layer of the execution environment is portable to other non-native platforms so that code can be ported to another non-native platform by replacing the platform-specific layer with a customized platform-specific layer, which is adapted to the other non-native platform.

Method, system and program product for porting code utilizing a portable execution environment

A method and apparatus for managing shared virtual storage in an information handling system in which each of a plurality of processes managed by an operating system has a virtual address space comprising a range of virtual addresses that are mapped to a corresponding set of real addresses representing addresses in real storage. The virtual address spaces are 64-bit address spaces requiring up to five levels of dynamic address translation (DAT) tables to map their virtual addresses to real addresses. One or more shared ranges of virtual addresses are defined that are mapped for each of a plurality of virtual address spaces to a common set of real addresses. The operating system manages these shared ranges using a system-level DAT table that reference a shared set of DAT tables used by the sharing address spaces for address translation, but is not attached to the hardware address translation facilities or used for address translation. The shared range of virtual addresses straddles the 2-byte boundary between ranges served by different third-level DAT tables and is situated between a lower private range and an upper private range so that an individual address space can map both a lower private range and a shared range using only three levels of DAT tables. Each shared address range may be shared with either global access rights, in which each participating process has the same access rights, or local access rights in which each participant may have different access rights to the given range. Access rights for each participant may be changed over the lifetime of the process.

Method and apparatus for managing shared virtual storage in an information handling system

A sectioned ordered queue in an information handling system comprises a plurality of queue sections arranged in order from a first queue section to a last queue section. Each queue section contains one or more queue entries that correspond to available ranges of real storage locations and are arranged in order from a first queue entry to a last queue entry. Each queue section and each queue entry in the queue sections having a weight factor defined for it. Each queue entry has an effective weight factor formed by combining the weight factor defined for the queue section with the weight factor defined for the queue entry. A new entry is added to the last queue section to indicate a newly available corresponding storage location, and one or more queue entries are deleted from the first section of the queue to indicate that the corresponding storage locations are no longer available. Various processes associated with a real storage manager are responsible for keeping the queue sections in balance, updating the queue sections, and removing entries from the queue to &ldquo;steal&rdquo; real storage frames.

Method and apparatus for maintaining order in a queue by combining entry weights and queue weights

In a virtual machine environment, the invention enables creation of a long running, reusable, virtual machine are disclosed. The environment includes a shared heap where requisite runtime code to bring the virtual machine into a 'ready' mode are loaded, linked, verified, initialized and compiled. Subsequent virtual machines are started and jointly use the shared heap. Applications create their objects in 'private heaps' that are exclusively reserved for the respective applications. At the end of execution of an application, each private heap is reinitialized. Static initializers are run in a persistent area of each private heap. This persistent area is reset to its initial values in between execution of applications. This obviates the need to terminate the virtual machine.

Long running, reusable, extendible, virtual machine

A method and apparatus for selectively using input/output (I/O) buffers as a retransmit vehicle in a client/server system. The decision whether to use an I/O buffer as a retransmit vehicle is based on a number of factors, including the packet size, the expected round-trip time (RTT) for an acknowledgment of the transmission, the number of I/O buffers currently allocated, and the number of I/O buffers remaining. If the decision is made not to use the I/O buffer as a retransmit vehicle, then the data is copied into a send buffer that is maintained by the system for the particular requester. Initially three threshold values, the round-trip time (RTT) threshold, the critical threshold, and the tight buffer threshold, are set. Connections having a longer round-trip time than a set round-trip time threshold or connections made when the number of I/O buffers remaining is below the critical threshold are not allowed to keep the I/O buffer as a retransmission vehicle. If the number of I/O buffers remaining falls below the critical threshold, a critical stabilization interval is started. During a critical stabilization interval, the I/O buffers may not be used as a retransmit vehicle if the number of I/O buffers already allocated exceeds the tight buffer threshold, even if the number of I/O buffers remaining is above the critical threshold. For each I/O buffer, a use count is maintained of the number of packets in the buffer awaiting acknowledgment. The use count is decremented each time an acknowledgment is received for one of the packets in the I/O buffer. When the use count has been decremented to zero, the I/O buffer is freed.

Method and apparatus for selectively using input/output buffers as a retransmit vehicle in an information handling system

Apparatus and accompanying methods for use preferably in a multi-system shared data (sysplex) environment (,), wherein each system (,) provides one or more servers (,), for dynamically and adaptively assigning and balancing new work and for new session requests, among the servers in the sysplex, in view of attendant user-defined business importance of these requests and available sysplex resource capacity so as to meet overall business goals. Specifically, systems and servers are categorized into two classes: eligible, i.e., goal-oriented servers running under a policy and for which capacity information is currently available, and candidate, i.e., servers which lack capacity information. Work requests for a client application are assigned first to various eligible systems and eligible servers thereon based on their current capacity to accept new work and in a manner that meets business goals inherent in a sysplex policy; followed, if additional servers are requested by that application, to candidate systems and candidate servers thereon. As to session placement, first those system(s) are selected that have lowest utilization, at a target importance level, but with sufficient available capacity at that level. Competing servers on the selected system(s) are then evaluated based on their corresponding session count data to yield a single resulting server. Thereafter, identification of multiple servers and their corresponding weights are returned to, e.g., a client application, for eventual routing of work requests to those servers, or the identification of a single server is returned to that application for establishing a new session therewith.

Apparatus and accompanying method for assigning session requests in a multi-server sysplex environment

In a client/server system, a method and apparatus for a listening daemon to distribute work and client connectivity across multiple processes each supporting multiple threads. Upon receiving a request from a client, a listening daemonaccepts the conversation, thus creating a socket. The listening daemon then does a putwork( ) call which places a work request on a system managed work queue. The work request behaves in a manner very similar to spawn in that it provides for parameters to be passed as well as transfer of ownership of the socket. In the meantime, server processes that are created either manually or by the operating system create multiple threads, which invoke a getwork( ) service to retrieve work requests created by the listening daemon. The getwork( ) service transfers the socket to the calling process and notifies the thread which file descriptors it has inherited. The thread is then able to process the client request and respond directly to the client without requiring an intermediate process to provide the communication link. The present invention allows a server in a client/server system to spread the workload across multiple threads in multiple processes, avoiding the creation of too many processes or too many threads in a single process. The direct transfer of the socket eliminates the unnecessary overhead typically required to rebind a new process to the client application.

Method and apparatus for transferring file descriptors in a multiprocess, multithreaded client/server system

A method and apparatus for handling outgoing communication requests in an information handling system in which outgoing communication packets are accumulated into a block that is written to an input/output (I/O) device. For each I/O device there is generated a blocking factor representing a predetermined number of packets that are accumulated before the block is written to the I/O device, as well as a push interval representing a maximum period of time for which any packet in the block can be stalled. Upon the arrival of a new outgoing packet, the packet is added to the block, and the block is written to the I/O device if either the block now contains the predetermined packets or any packet in the packet has been waiting for more than the push interval. A timer running asynchronously with the arrival of outgoing requests periodically pops to write the block to the I/O device if it has been waiting overlong, even if no new requests have arrived. Both the blocking factor and the push interval are periodically adjusted in accordance with the actual throughput so that the blocking factor corresponds to the exact level of consistent parallelism for a given workload.

Method and apparatus for adaptively blocking outgoing communication

A method and apparatus for controlling the assignment of units of work to workload enclaves in a client/server system in which units of work are organized into enclaves with the units of work in each enclave being managed as a single entity. Incoming work requests are queued for processing by one or more worker threads of one or more server processes. In response to the receipt of an incoming work request, a putwork() service of the operating system kernel defines an workload enclave for the work request and adds the work request to a queue of incoming work requests to enqueue the request. In response to a request from a worker thread for more work, a getwork() service of the operating system kernel removes a work request from the queue to dequeue the request and associates the worker thread with the enclave defined for the request. If a worker thread must create a new unit of work while processing a work request, it calls an appropriate service of the operating system kernel, which creates the new unit of work and associates the new process with the enclave of the worker thread. The worker thread is disassociated from the enclave upon completion of the work request.

Method and apparatus for controlling the assignment of units of work to

Distributed computing environment (DCE) remote procedure calls (RPCs) are integrated with an advisory work load manager (WLM) to provide a way to intelligently dispatch RPC requests among the available application server processes. The routing decisions are made dynamically (for each RPC) based on interactions between the location broker and an advisory work load manager. Furthermore, when the system contains multiple coupled processors (tightly coupled within a single frame, or loosely coupled within a computing complex, a local area network (LAN) configuration, a distributed computing environment (DCE) cell, etc.), the invention extends to balance the processing of RPC requests and the associated client sessions across the coupled systems. Once a session is assigned to a given process, the invention also supports performance monitoring and reporting, dynamic system resource allocation for the RPC requests, and potentially any other benefits that may be available through the specific work load manager (WLM).

Integrating distributed computing environment remote procedure calls

Apparatus and accompanying methods for use preferably in a multi-system shared data (sysplex) environment (100), wherein each system (110) provides one or more servers (115), for dynamically and adaptively assigning and balancing new work and for new session requests, among the servers in the sysplex, in view of attendant user-defined business importance of these requests and available sysplex resource capacity so as to meet overall business goals. Specifically, systems and servers are categorized into two classes: eligible, i.e., goal-oriented servers running under a policy and for which capacity information is currently available, and candidate, i.e., servers which lack capacity information. Work requests for a client application are assigned first to various eligible systems and eligible servers thereon based on their current capacity to accept new work and in a manner that meets business goals inherent in a sysplex policy; followed, if additional servers are requested by that application, to candidate systems and candidate servers thereon. As to session placement, first those system(s) are selected that have lowest utilization, at a target importance level, but with sufficient available capacity at that level. Competing servers on the selected system(s) are then evaluated based on their corresponding session count data to yield a single resulting server. Thereafter, identification of multiple servers and their corresponding weights are returned to, e.g., a client application, for eventual routing of work requests to those servers, or the identification of a single server is returned to that application for establishing a new session therewith.

Growing community of inventors

Wappingers Falls, NY, United States of America

David B Emmes