公开(公告)号：US07370103B2

公开(公告)日：2008-05-06

申请号：US10987404

申请日：2004-11-12

申请人： Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  David S. Stutz ,  Bassam Tabbara ,  Robert V. Welland

发明人： Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  David S. Stutz ,  Bassam Tabbara ,  Robert V. Welland

IPC分类号： G06F15/173

CPC分类号： H04L41/042 ,  H04L41/0659 ,  H04L41/0672 ,  H04L63/06 ,  H04L69/40

摘要： A multi-tiered server management architecture is employed including an application development tier, an application operations tier, and a cluster operations tier. In the application development tier, applications are developed for execution on one or more server computers. In the application operations tier, execution of the applications is managed and sub-boundaries within a cluster of servers can be established. In the cluster operations tier, operation of the server computers is managed without concern for what applications are executing on the one or more server computers and boundaries between clusters of servers can be established. The multi-tiered server management architecture can also be employed in co-location facilities where clusters of servers are leased to tenants, with the tenants implementing the application operations tier and the facility owner (or operator) implementing the cluster operations tier.

摘要翻译： 采用多层次的服务器管理架构，包括应用程序开发层，应用程序操作层和集群操作层。 在应用程序开发层中，开发应用程序用于在一个或多个服务器计算机上执行。 在应用程序操作层中，管理应用程序的执行，并且可以建立服务器集群内的子边界。 在集群操作层中，管理服务器计算机的操作，而不用担心在一个或多个服务器计算机上执行哪些应用程序，并且可以建立服务器集群之间的边界。 多层次的服务器管理架构也可以用于将服务器集群租赁给租户的共同位置设施中，承租人实施应用程序操作层和实施集群操作层的设施所有者（或运营商）。

公开(公告)号：US07334235B2

公开(公告)日：2008-02-19

申请号：US10385381

申请日：2003-03-10

申请人： Galen C. Hunt ,  Gerald Cermak ,  Robert J. Stets, Jr.

发明人： Galen C. Hunt ,  Gerald Cermak ,  Robert J. Stets, Jr.

IPC分类号： G06F9/46

CPC分类号： G06F9/465 ,  G06F2209/463

摘要： Operating system functions are defined as objects that are collections of data and methods. The objects represent operating system resources. The resource objects can be instantiated and used across process and machine boundaries. Each object has an associated handle that is stored in its private state. When an application requests a resource, it is given a second handle or pseudo handle that corresponds with the handle in the object's private state. The second handle is valid across process and machine boundaries and all access to the object takes place through the second handle. This greatly facilitates remote computing. In preferred embodiments, the objects are COM objects and remote computing is facilitated through the use of Distributed COM (DCOM) techniques. Other embodiments of the invention provide legacy and versioning support by identifying each resource, rather than the overall operating system, with a unique identifier that can specified by an application. Different versions of the same resource have different identifiers. This ensures that applications that need a specific version of a resource can receive that version. This also ensures that an application can specifically request a particular version of a resource by using its unique identifier, and be assured of receiving that resource. Other embodiments of the invention provide legacy support by intercepting calls for operating system functions and transforming those calls into object calls that can be understood by the resource objects. This is accomplished in preferred embodiments by injecting a level of indirection between an application and an operating system.

摘要翻译： 操作系统功能被定义为数据和方法集合的对象。 对象代表操作系统资源。 资源对象可以在进程和机器边界之间实例化和使用。 每个对象具有存储在其私有状态的关联句柄。 当一个应用程序请求一个资源时，它被赋予与该对象的私有状态中的句柄相对应的第二个句柄或伪句柄。 第二个句柄在进程和机器边界之间是有效的，对对象的所有访问都通过第二个句柄进行。 这极大地促进了远程计算。 在优选实施例中，对象是COM对象，并且通过使用分布式COM（DCOM）技术来促进远程计算。 本发明的其它实施例通过使用可由应用指定的唯一标识符来识别每个资源而不是整个操作系统来提供遗留和版本控制支持。 同一资源的不同版本具有不同的标识符。 这样可确保需要特定版本资源的应用程序可以接收该版本。 这还可以确保应用程序可以通过使用其唯一标识符来特定请求特定版本的资源，并确保接收该资源。 本发明的其他实施例通过拦截对操作系统功能的调用并将这些调用变换成资源对象可以理解的对象调用来提供传统支持。 这在优选实施例中通过在应用程序和操作系统之间注入一定程度的间接来实现。

公开(公告)号：US20080005750A1

公开(公告)日：2008-01-03

申请号：US11428037

申请日：2006-06-30

申请人： Galen C. Hunt ,  James R. Larus ,  Manuel A. Fahndrich ,  Bjarne Steensgaard ,  David R. Tarditi ,  Brian Zill

发明人： Galen C. Hunt ,  James R. Larus ,  Manuel A. Fahndrich ,  Bjarne Steensgaard ,  David R. Tarditi ,  Brian Zill

IPC分类号： G06F9/46

CPC分类号： G06F9/545 ,  G06F9/468

摘要： Described herein are one or more implementations that separate kernel interfaces functions into those that act on kernel objects owned by a process and accessed exclusively by that process—described herein as local kernel objects—from access to kernel objects owned by a process and accessible_by other active processes.

摘要翻译： 这里描述的是一个或多个实现，其将内核接口功能分离成对由进程所拥有的内核对象进行操作且仅由此处描述的本地描述的本地内核对象进行访问的实现 - 从进程所拥有的内核对象和可访问的其他活动 过程。

公开(公告)号：US07162509B2

公开(公告)日：2007-01-09

申请号：US11324062

申请日：2005-12-30

申请人： Mathilde C. Brown ,  Charlie D. Chase ,  Kevin Grealish ,  Galen C. Hunt ,  Aamer Hydrie ,  David Noble ,  Geoffrey Outhred ,  Glenn R. Peterson ,  Bassam Tabbara ,  Alexander Torone ,  Robert V. Welland

发明人： Mathilde C. Brown ,  Charlie D. Chase ,  Kevin Grealish ,  Galen C. Hunt ,  Aamer Hydrie ,  David Noble ,  Geoffrey Outhred ,  Glenn R. Peterson ,  Bassam Tabbara ,  Alexander Torone ,  Robert V. Welland

IPC分类号： G06F15/16

CPC分类号： G06F8/20

摘要： A design tool includes a service definition model to enable abstract description of distributed computing systems and distributed applications. The design tool also includes a schema to dictate how functional operations within the service definition model are to be specified. The functional operations include design of distributed applications, deployment of distributed applications, and management of distributed applications.

公开(公告)号：US07127514B2

公开(公告)日：2006-10-24

申请号：US09752114

申请日：2000-12-28

申请人： Galen C. Hunt

发明人： Galen C. Hunt

IPC分类号： G06F15/16

CPC分类号： G06F9/546 ,  G06F9/547 ,  H04L67/2842 ,  H04L67/306 ,  H04L67/42

摘要： A stateless distributed computer architecture allows state-caching objects, which hold server state information, to be maintained on a client or network rather than on a server. In one implementation, the computer architecture implements object-oriented program modules according to a distributed component object model (DCOM). Using an object-oriented network protocol (e.g., remote procedure call), a client-side application calls through an application program interface (API) to a program object residing at a server computer. The program object, responsive to the call, creates a state caching object that contains state information pertaining to the client connection. The server inserts the state-caching object into a local thread context and processes the request to generate a reply. The server subsequently attaches the state-caching object to the reply and returns them both to the client. The client stores the state-caching object for later communication with the server. When the client subsequently calls the program object at the server, the client submits the state-caching object along with the request packet. The server uses the state information in the state-caching object to quickly restore state for the client reconnection. In this manner, the server can offload its state information to other computing devices in the distributed architecture, thereby improving scalability. In another implementation, the network itself caches the server-oriented state-caching object.

公开(公告)号：US07043545B2

公开(公告)日：2006-05-09

申请号：US11007828

申请日：2004-12-08

申请人： Bassam Tabbara ,  Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  David S. Stutz ,  Robert V. Welland

发明人： Bassam Tabbara ,  Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  David S. Stutz ,  Robert V. Welland

IPC分类号： G06F15/173

CPC分类号： H04L67/34 ,  H04L69/329

摘要： A controller, referred to as the “BMonitor”, is situated on a computer. The BMonitor includes a plurality of filters that identify where data can be sent to and/or received from, such as another node in a co-location facility or a client computer coupled to the computer via the Internet. The BMonitor further receives and implements requests from external sources regarding the management of software components executing on the computer, allowing such external sources to initiate, terminate, debug, etc. software components on the computer. Additionally, the BMonitor operates as a trusted third party mediating interaction among multiple external sources managing the computer.

公开(公告)号：US07039919B1

公开(公告)日：2006-05-02

申请号：US09196836

申请日：1998-11-20

申请人： Galen C. Hunt

发明人： Galen C. Hunt

IPC分类号： G06F9/00

CPC分类号： G06F9/5066 ,  G06F9/449 ,  G06F9/465 ,  H04L43/0864 ,  H04L43/0894 ,  H04L67/10

摘要： An automatic distributed partitioning system (ADPS) determines which unit exposes an interface. The ADPS detects a reference to an interface. For the interface, the ADPS assures that the unit that exposes the interface is identified. In some embodiments, a data structure such as a hash table associates interfaces with identified units. An interface wrapper for the interface stores the identity of the unit that exposes the interface, as well as information about the interface and a reference to instrumentation. When a client unit calls a member function of an interface, the interface wrapper intercepts the call and invokes the instrumentation. The instrumentation performs an operation such as profiling the application or classifying a unit. The ADPS detects when an interface is undocumented, and handles undocumented interfaces without invoking the full instrumentation capabilities of the ADPS.

摘要翻译： 自动分配分区系统（ADPS）确定哪个单元暴露了一个接口。 ADPS检测到对接口的引用。 对于接口，ADPS确保识别出暴露接口的单元。 在一些实施例中，诸如散列表的数据结构将接口与已标识的单元相关联。 接口的接口包装器存储暴露接口的单元的标识，以及有关接口的信息和对仪器的参考。 当客户端单元调用接口的成员函数时，接口包装器拦截该调用并调用该检测。 仪器执行诸如对应用进行分析或分类单元的操作。 ADPS检测何时未记录接口，并处理未记录的接口，而无需调用ADPS的完整仪器功能。

公开(公告)号：US06915338B1

公开(公告)日：2005-07-05

申请号：US09696752

申请日：2000-10-24

申请人： Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  Bassam Tabbara ,  Mark D. Van Antwerp ,  Robert V. Welland

发明人： Galen C. Hunt ,  Aamer Hydrie ,  Steven P. Levi ,  Bassam Tabbara ,  Mark D. Van Antwerp ,  Robert V. Welland

IPC分类号： G06F9/50 ,  G06F15/177

CPC分类号： G06F9/5066

摘要： A system and procedure to automatically enforce policy in distributed multi-computer service applications. Such service applications include multiple software modules that execute on multiple computers. The computers have access to communications media that allow data communications between the computers. Logical ports are configured on different modules according to a logical model of the multi-computer service application. Each logical port is defined by port software. Logical data connections between the logical ports are configured in accordance with the logical model. Each port is configured to communicate through different numbers of logical data connections without modifying the port software.In response to the occurrence of a condition, a module sends an event notification to a policy module. The policy module responds to the notification by formulating a request for one or more destination modules. The policy module provides the request to an output port of the policy module. The output port forwards the request to input ports of multiple modules according to the configured logical data connections.

摘要翻译： 在分布式多计算机服务应用程序中自动执行策略的系统和过程。 这样的服务应用程序包括在多台计算机上执行的多个软件模块。 计算机可以访问允许计算机之间进行数据通信的通信媒体。 逻辑端口根据多计算机服务应用的逻辑模型配置在不同的模块上。 每个逻辑端口由端口软件定义。 根据逻辑模型配置逻辑端口之间的逻辑数据连接。 每个端口配置为通过不同数量的逻辑数据连接进行通信，而无需修改端口软件。 响应于条件的发生，模块向策略模块发送事件通知。 策略模块通过为一个或多个目标模块制定请求来响应通知。 策略模块向策略模块的输出端口提供请求。 输出端口根据配置的逻辑数据连接将请求转发到多个模块的输入端口。

公开(公告)号：US06381628B1

公开(公告)日：2002-04-30

申请号：US09197211

申请日：1998-11-20

申请人： Galen C. Hunt

发明人： Galen C. Hunt

IPC分类号： G06F1300

CPC分类号： G06F11/3612 ,  G06F9/465 ,  H04L41/0233 ,  H04L43/0864 ,  H04L43/0894 ,  H04L67/10

摘要： An automatic distributed partitioning system (ADPS) profiles a computer network using short term sampling to estimate minimum network latency and maximum network bandwidth. The ADPS profiles an application program to produce an application profile that summarizes inter-unit communication within the application program. Summarization of the application profile reduces overhead during profiling of the application program. The application profile is network independent, but by combining it with information in the network profile, the ADPS produces a model of the behavior of the application program on the computer network. Quick estimation of network characteristics facilitates the ADPS repartitioning the application to adjust to a different computer network or a changed computer network.

摘要翻译： 自动分布式分区系统（ADPS）使用短期采样来配置计算机网络，以估计最小网络延迟和最大网络带宽。 ADPS描述应用程序，以生成应用程序配置文件，其总结应用程序内的单元间通信。 应用程序配置文件的汇总减少了应用程序分析过程中的开销。 应用程序配置文件与网络无关，但通过将其与网络配置文件中的信息相结合，ADPS会在计算机网络上生成应用程序行为的模型。 网络特征的快速估计有助于ADPS重新分配应用程序以适应不同的计算机网络或改变的计算机网络。

公开(公告)号：US06263491B1

公开(公告)日：2001-07-17

申请号：US09197009

申请日：1998-11-20

申请人： Galen C. Hunt

发明人： Galen C. Hunt

IPC分类号： G06F9445

CPC分类号： G06F11/3612 ,  G06F9/465 ,  H04L41/0233 ,  H04L67/10

摘要： An instrumentation system performs operations such as profiling an application and partitioning and distributing units of the application using different versions of metadata describing the application. Performing an operation on an executing application generates overhead. Detailed metadata used in operations such as profiling create unnecessary overhead during other operations. By removing metadata detail unnecessary for a particular operation, an instrumentation system using reduced metadata generates less overhead for that particular operation. Different instrumentation packages include different versions of metadata for performing operations on the application.

摘要翻译： 仪器系统执行诸如使用不同版本的描述应用程序的元数据来分析应用程序并分配和分发应用程序单元的操作。 对执行的应用程序执行操作会产生开销。 在诸如分析等操作中使用的详细元数据在其他操作期间产生不必要的开销。 通过删除特定操作不必要的元数据细节，使用减少的元数据的仪器系统会为该特定操作生成较少的开销。 不同的仪器包包括用于对应用程序执行操作的不同版本的元数据。