-
公开(公告)号：US07302450B2
公开(公告)日：2007-11-27
申请号：US10876021
申请日：2004-06-24
申请人： Fabio Benedetti , Pietro Iannucci , Scot MacLellan
发明人： Fabio Benedetti , Pietro Iannucci , Scot MacLellan
IPC分类号： G06F7/00
CPC分类号： G06F9/5083 , G06Q10/06 , Y10S707/99945 , Y10S707/99948
摘要： A workload scheduler supporting an efficient distribution and balancing of the workload is proposed. The scheduler maintains (383-386) a profile for each job; the profile (build using statistics of previous executions of the job) defines an estimated usage of different resources of the system by the job. The scheduler tends to select (318-342) the jobs with complementary resource requirements (according to a combination of their attributes); this process can be carried out using either a heuristic approach (318-334) or an optimization approach (335-342). As a result, the jobs that will be submitted are very likely to consume different resources of the system; in this way, any contention for the different resources is strongly reduced.
摘要翻译： 提出了一种支持工作负载有效分配和平衡的工作负载调度器。 调度程序维护（383 - 386）每个作业的配置文件; 配置文件（使用以前执行作业的统计信息构建）定义了作业对系统的不同资源的估计使用情况。 调度器倾向于选择具有补充资源需求（根据其属性的组合））的作业（318-342）; 该过程可以使用启发式方法（318 - 334）或优化方法（335 - 342）进行。 因此，提交的工作很可能会消耗不同的系统资源; 以这种方式，对于不同资源的任何争用都大大降低。
-
公开(公告)号：US20050076043A1
公开(公告)日：2005-04-07
申请号：US10876021
申请日：2004-06-24
申请人： Fabio Benedetti , Pietro Iannucci , Scot MacLellan
发明人： Fabio Benedetti , Pietro Iannucci , Scot MacLellan
CPC分类号： G06F9/5083 , G06Q10/06 , Y10S707/99945 , Y10S707/99948
摘要： A workload scheduler supporting an efficient distribution and balancing of the workload is proposed. The scheduler maintains (383-386) a profile for each job; the profile (build using statistics of previous executions of the job) defines an estimated usage of different resources of the system by the job. The scheduler tends to select (318-342) the jobs with complementary resource requirements (according to a combination of their attributes); this process can be carried out using either a heuristic approach (318-334) or an optimization approach (335-342). As a result, the jobs that will be submitted are very likely to consume different resources of the system; in this way, any contention for the different resources is strongly reduced.
摘要翻译： 提出了一种支持工作负载有效分配和平衡的工作负载调度器。 调度程序维护（383-386）每个作业的配置文件; 配置文件（使用以前执行作业的统计信息构建）定义了作业对系统的不同资源的估计使用情况。 调度程序倾向于选择（318-342）具有补充资源需求的作业（根据其属性的组合）; 该过程可以使用启发式方法（318-334）或优化方法（335-342）进行。 因此，提交的工作很可能会消耗不同的系统资源; 以这种方式，对于不同资源的任何争用都大大降低。
-
公开(公告)号：US20060288049A1
公开(公告)日：2006-12-21
申请号：US11425230
申请日：2006-06-20
申请人： Fabio Benedetti , Pietro Iannucci , Franco Mossoto
发明人： Fabio Benedetti , Pietro Iannucci , Franco Mossoto
IPC分类号： G06F17/30
CPC分类号： G06F9/526
摘要： A solution is proposed in a scheduler for updating a control file (220) storing a workload plan (defining a sequence of jobs to be executed during a production day). A handler (225) has exclusive write access to the control file (for updating information relating to the state of the jobs during execution of the plan). A planner (215) that needs to write a new plan into the control file (for the next production day) notifies the handler. In response thereto, the handler makes a copy (320) of the control file for use by the planner and starts logging (335) any update that it performs on the (original) control file. Once the planner completes its work, the handler applies the logged updates to the copy of the control file (as updated by the planner). The handler can now restart working on the copy of the control file, which then replaces its original version.
摘要翻译： 在用于更新存储工作负载计划（定义要在生产当天执行的作业序列）的控制文件（220）的调度器中提出了一种解决方案。 处理程序（225）具有对控制文件的独占写入访问（用于在执行计划期间更新与作业的状态有关的信息）。 需要将新计划写入控制文件（在下一个生产日期）的计划员（215）会通知处理程序。 响应于此，处理程序制作控制文件的副本（320）供规划者使用，并开始记录（335）对（原始）控制文件执行的任何更新。 一旦规划人员完成工作，处理程序将记录的更新应用到控制文件的副本（由计划员更新）。 处理程序现在可以重新启动控制文件的副本，然后替换其原始版本。
-
公开(公告)号：US20120204180A1
公开(公告)日：2012-08-09
申请号：US13500345
申请日：2010-08-24
IPC分类号： G06F9/46
摘要： A method, system or computer usable program product for managing jobs scheduled for execution on a target system in which some jobs may spawn additional jobs scheduled for execution on the target system including intercepting jobs scheduled for execution in the target system, determining whether there is resource sufficiency in the target system for executing jobs, responsive to an affirmative determination of resource sufficiency, releasing previously intercepted jobs for execution in the target system, computing a limit of a number of jobs which can be concurrently scheduled by an external system to the target system, and transmitting the computed limit to the external system.
摘要翻译： 一种用于管理计划执行在目标系统上的作业的方法，系统或计算机可用程序产品，其中一些作业可以产生计划在目标系统上执行的附加作业，包括计划执行在目标系统中的拦截作业，确定是否存在资源 响应于资源充足的肯定的确定，释放在目标系统中执行的先前截获的作业，执行作业的目标系统的充分性，计算可以由外部系统同时调度到目标系统的多个作业的限制 并将所计算的限制传送到外部系统。
-
公开(公告)号：US08799911B2
公开(公告)日：2014-08-05
申请号：US13500345
申请日：2010-08-24
IPC分类号： G06F9/46
摘要： A system or computer usable program product for managing jobs scheduled for execution on a target system in which some jobs may spawn additional jobs scheduled for execution on the target system including intercepting jobs scheduled for execution in the target system, determining whether there is resource sufficiency in the target system for executing jobs, responsive to an affirmative determination of resource sufficiency, releasing previously intercepted jobs for execution in the target system, computing a limit of a number of jobs which can be concurrently scheduled by an external system to the target system, and transmitting the computed limit to the external system.
摘要翻译： 一种用于管理计划在目标系统上执行的作业的系统或计算机可用程序产品，其中一些作业可以产生计划在目标系统上执行的附加作业，包括计划在目标系统中执行的拦截作业，确定是否存在资源足够 用于执行作业的目标系统，响应于资源足够的肯定的确定，释放在目标系统中执行的先前拦截的作业，计算可以由外部系统同时调度到目标系统的作业数量的限制;以及 将计算的限制传输到外部系统。
-
6.发明申请Method and apparatus to create a reservation against a future scheduling object instantiation 有权针对未来调度对象实例化创建预留的方法和装置公开(公告)号：US20060132828A1
公开(公告)日：2006-06-22
申请号：US11014064
申请日：2004-12-16
申请人： Fabio Benedetti , Maria Cagnetta , David Kumhyr , Pietro Iannucci
发明人： Fabio Benedetti , Maria Cagnetta , David Kumhyr , Pietro Iannucci
IPC分类号： G06F3/12
CPC分类号： G06F3/1222 , G06F3/1238 , G06F3/126 , G06F3/1288
摘要： A method in a data processing system for managing jobs. A change to a job receiving that is un-instantiated. A reservation for the job is created. Instantiated jobs for an instance of the job are monitored corresponding to the reservation, and applying the change to the instance of the job responsive to identifying the instantiated job.
摘要翻译： 用于管理作业的数据处理系统中的方法。 接收未实例化的作业的更改。 创建作业的预约。 对作业的实例的实例化作业是根据预留进行监视的，并根据识别实例的作业将更改应用于作业的实例。
-
7.发明申请公开(公告)号：US20050235289A1
公开(公告)日：2005-10-20
申请号：US11044604
申请日：2005-01-27
IPC分类号： G06F9/46
CPC分类号： G06F9/50 , G06F9/526 , G06F2209/522
摘要： A method (300) and a corresponding system for allocating resources in a hierarchical data processing system are proposed. The system includes a plurality of nodes that are arranged in a tree. Each leaf-node, or client, controls corresponding local resources, while each intermediate node, or domain manager, controls global resources that are visible to all its dependent nodes; the global resources defined on a root node, or master domain manager, are then visible to all the nodes of the system. Whenever a job, being submitted (306) on a generic client, requests a set of resources for its execution, the request is sent (309,318,327) through the tree to the master domain manager. The master domain manager allocates (339) the corresponding global resources as soon as they are all available; this event enables (345) its child node (in a path to the client) to perform the same operations, and so on down to the client itself. Once the client has allocated (357) the corresponding local resources, the execution of the job can be launched (360).
摘要翻译： 提出了一种方法（300）和用于在分层数据处理系统中分配资源的对应系统。 该系统包括布置在树中的多个节点。 每个叶节点或客户端控制对应的本地资源，而每个中间节点或域管理器控制对其所有依赖节点可见的全局资源; 然后，在根节点或主域管理器上定义的全局资源对系统的所有节点都可见。 每当在通用客户端上提交（306）的作业请求一组用于其执行的资源时，通过树将请求发送（309,318,327）到主域管理器。 主域名管理员尽快分配（339）相应的全球资源; 此事件使（345）其子节点（在客户端的路径中）执行相同的操作，依此类推，直到客户端本身。 一旦客户端分配（357）相应的本地资源，就可以启动作业的执行（360）。
-
8.发明授权Allocating resources across multiple nodes in a hierarchical data processing system according to a decentralized policy 有权根据分散策略在分层数据处理系统中的多个节点之间分配资源公开(公告)号：US07810098B2
公开(公告)日：2010-10-05
申请号：US11044604
申请日：2005-01-27
IPC分类号： G06F9/46 , G06F15/173
CPC分类号： G06F9/50 , G06F9/526 , G06F2209/522
摘要： A method (300) and a corresponding system for allocating resources in a hierarchical data processing system are proposed. The system includes a plurality of nodes that are arranged in a tree. Each leaf-node, or client, controls corresponding local resources, while each intermediate node, or domain manager, controls global resources that are visible to all its dependent nodes; the global resources defined on a root node, or master domain manager, are then visible to all the nodes of the system. Whenever a job, being submitted (306) on a generic client, requests a set of resources for its execution, the request is sent (309,318,327) through the tree to the master domain manager. The master domain manager allocates (339) the corresponding global resources as soon as they are all available; this event enables (345) its child node (in a path to the client) to perform the same operations, and so on down to the client itself. Once the client has allocated (357) the corresponding local resources, the execution of the job can be launched (360).
摘要翻译： 提出了一种方法（300）和用于在分层数据处理系统中分配资源的对应系统。 该系统包括布置在树中的多个节点。 每个叶节点或客户端控制对应的本地资源，而每个中间节点或域管理器控制对其所有依赖节点可见的全局资源; 然后，在根节点或主域管理器上定义的全局资源对系统的所有节点都可见。 每当在通用客户端上提交（306）的作业请求一组用于其执行的资源时，通过树将请求发送（309,318,327）到主域管理器。 主域名管理员尽快分配（339）相应的全球资源; 此事件使（345）其子节点（在客户端的路径中）执行相同的操作，依此类推，直到客户端本身。 一旦客户端分配（357）相应的本地资源，就可以启动作业的执行（360）。
-
9.发明授权Mechanism to create a reservation against a future scheduling object instantiation 有权针对未来调度对象实例化创建预留的机制公开(公告)号：US08054487B2
公开(公告)日：2011-11-08
申请号：US11014064
申请日：2004-12-16
IPC分类号： G06F3/12
CPC分类号： G06F3/1222 , G06F3/1238 , G06F3/126 , G06F3/1288
摘要： A method in a data processing system for managing jobs. A change to a job receiving that is un-instantiated. A reservation for the job is created. Instantiated jobs for an instance of the job are monitored corresponding to the reservation, and applying the change to the instance of the job responsive to identifying the instantiated job.
摘要翻译： 用于管理作业的数据处理系统中的方法。 接收未实例化的作业的更改。 创建作业的预约。 对作业的实例的实例化作业是根据预留进行监视的，并根据识别实例的作业将更改应用于作业的实例。
-
公开(公告)号：US08713578B2
公开(公告)日：2014-04-29
申请号：US12415452
申请日：2009-03-31
CPC分类号： G06F9/4887 , G06F9/5038 , G06F2209/506
摘要： This disclosure describes monitoring the execution of jobs in a work plan. In an embodiment, a system maintains a risk level associated with the critical job to represent whether the execution of a job preceding the critical job has a problem, and it maintains the list associated with the critical job so as to quickly identify the preceding job which may cause a delay to the critical job execution.
摘要翻译： 本公开描述了监视工作计划中作业的执行情况。 在一个实施例中，系统维护与关键作业相关联的风险级别，以表示关键作业之前的作业的执行是否具有问题，并且维护与关键作业相关联的列表，以便快速识别先前的作业， 可能导致关键作业执行的延迟。
-
-
-
-
-
-
-
-
-
搜索结果
15 条记录 (耗时 0.021 秒)
