System and method for workload balancing by resetting an average queue depth upon the start of the server instance
    1.
    发明授权
    System and method for workload balancing by resetting an average queue depth upon the start of the server instance 有权
    通过在服务器实例启动时重置平均队列深度来实现工作负载平衡的系统和方法

    公开(公告)号:US07409688B2

    公开(公告)日:2008-08-05

    申请号:US10713959

    申请日:2003-11-13

    IPC分类号: G06F13/38 G06F9/46

    CPC分类号: G06F9/546 G06F9/5083

    摘要: The present invention relates to an apparatus, method and computer program for workload balancing in an asynchronous messaging system. The number of server instances, which process work items from a queue of messages, is controlled based upon that queue's average queue depth and one or more predetermined thresholds.

    摘要翻译: 本发明涉及一种用于异步消息系统中的工作负载平衡的装置,方法和计算机程序。 基于队列的平均队列深度和一个或多个预定阈值来控制从消息队列处理工作项的服务器实例的数量。

    Publish-subscribe looping detection and correction
    2.
    发明授权
    Publish-subscribe looping detection and correction 失效
    发布订阅循环检测和纠正

    公开(公告)号:US07558210B1

    公开(公告)日:2009-07-07

    申请号:US12250536

    申请日:2008-10-13

    CPC分类号: H04L12/1854 H04L12/1895

    摘要: This system detects and corrects looping problems in a publish-subscribe messaging network. This system requires a token which uniquely identifies a node in this network or universally unique in this messaging network. This system maintains a list of Universally Unique Identifiers (UUID) as a metadata attached to each publish-subscribe message (“publication”). As a node forwards a publication to another node, it is required to append its own UUID to this list or discard the message if its UUID already is in the attached list. Several nodes could form a broker cooperative aggregate node and have a unique identifier for all the nodes in the cooperative if they have an internal mechanism for preventing publication looping.

    摘要翻译: 该系统检测并纠正发布订阅消息传递网络中的循环问题。 该系统需要一个唯一标识该网络中的节点或在该消息传递网络中普遍唯一的令牌。 该系统将通用唯一标识符(UUID)的列表作为附加到每个发布订阅消息(“发布”)的元数据。 当节点将发布转发到另一个节点时,需要将其自己的UUID附加到此列表中,或者如果其UUID已在附件列表中,则丢弃该消息。 如果一个节点具有用于防止发布循环的内部机制,那么多个节点可以形成一个代理协作聚合节点,并具有协作中所有节点的唯一标识符。

    Data processing arrangement and method
    3.
    发明授权
    Data processing arrangement and method 有权
    资料处理安排及方法

    公开(公告)号:US06948093B2

    公开(公告)日:2005-09-20

    申请号:US10228636

    申请日:2002-08-27

    IPC分类号: G06F11/00 H04L1/22

    CPC分类号: G06F11/1471

    摘要: A data processing arrangement (100) and method for fuzzy backup and restore of MQSeries shared queues held in coupling facility list structure (106). Periodically backups are taken of the data held in the CF list structure by backing up messages in the sequence uncommitted puts (220), committed puts (230), uncommitted gets (240) and expired messages (250), whereby all messages may be backed up in the backup means except for exceptions which are recorded in the record log. In the event of media failure of the coupling facility, data can be restored by restoring messages in the sequence restoring messages from the backup means (320), replaying messages from the record log means (340) and determining final disposition of messages from replay of the record log means (350). This provides the advantage of allowing the number of log records written to be reduced.

    摘要翻译: 一种用于在耦合设施列表结构(106)中保存的MQSeries共享队列的模糊备份和恢复的数据处理装置(100)和方法。 通过在未提交的投放(220),提交投放(230),未提交的收益(240)和过期消息(250)的顺序中备份消息来对CF列表结构中保存的数据进行定期备份,从而所有消息都可以被备份 在记录日志中记录的异常除外。 在耦合设备的媒体故障的情况下,可以通过在从备份装置(320)恢复消息的顺序中恢复消息来恢复数据来恢复数据,从记录日志装置(340)重放消息并确定来自重播的消息的最终处置 记录日志意味着(350)。 这提供了允许减少写入的日志记录的数量的优点。

    Method for detecting and correcting publish-subscribe looping in a messaging network
    4.
    发明授权
    Method for detecting and correcting publish-subscribe looping in a messaging network 失效
    在消息传递网络中检测和纠正发布订阅循环的方法

    公开(公告)号:US07453812B1

    公开(公告)日:2008-11-18

    申请号:US12020554

    申请日:2008-01-27

    CPC分类号: H04L12/1854 H04L12/1895

    摘要: A method for detecting and correcting looping problems in a publish-subscribe messaging network is disclosed. The messaging network requires a token which uniquely identifies a node in this messaging network or universally unique in this messaging network. The messaging network maintains a list of Universally Unique Identifiers (UUID) as a metadata attached to each publish-subscribe message (“publication”). As a node forwards a publication to another node, it is required to append its own UUID to this list or discard the message if its UUID already is in the attached list. Several nodes could form a broker cooperative aggregate node and have a unique identifier for all the nodes in the cooperative if they have an internal mechanism for preventing publication looping.

    摘要翻译: 公开了一种用于检测和纠正发布订阅消息传递网络中的循环问题的方法。 消息传递网络需要一个唯一标识此消息传递网络中的节点或在此消息传递网络中普遍唯一的令牌。 消息传递网络将常规唯一标识符(UUID)的列表作为附加到每个发布 - 订阅消息(“发布”)的元数据。 当节点将发布转发到另一个节点时,需要将其自己的UUID附加到此列表中,或者如果其UUID已在附件列表中,则丢弃该消息。 如果一个节点具有用于防止发布循环的内部机制,那么多个节点可以形成一个代理协作聚合节点,并具有协作中所有节点的唯一标识符。