Abstract: A performance management system and method for cluster-based web services comprising a gateway for receiving a user request, assigning the user request to a class, queuing the user request based on said class, and dispatching the user request to one of a plurality of server resources based on the assigned class and control parameters. The control parameters are continuously updated by a global resource manager which tracks and evaluates system performance.

Abstract: A synchronous collaboration environment that supports real-time collaboration of multiple participants, each having shared and independent views of the shared workspace. Multiple views per participant are provided. Some of the views seen by a participant can be shared views with the usual common cursor and annotation tools. The shared views need not be homogeneous, which means that for a given view, each participant can see more than just some common data in his or window for the view. What the participant sees separately from the common data can make his or her shared view different from that of other participants. The view can be different due to different data being exposed in the view. Some of the views seen by a participant can be independent views. This allows to the participant synchronous working with the shared workspace alone on his or her own. The collaboration system includes a user interface and support for aligning views including goto and overlaying.

Abstract: In an ASP server farm, requests to use an application are directed to a particular executing instance of the application (or an appropriate component thereof) that is identified as being the least loaded of the available such instances of the application or its component. The number of such instances is dynamically increased or decreased in response to the number of requests for the application or components thereof. Requests may be directed (in accordance with the first aspect) or the instances adjusted (in accordance with a second aspect) on a per client-basis, in which instances of the application and/or components thereof are reserved for the use of a user or a particular group of users. Operation in this manner facilitates compliance with service agreements with respective users or groups of users.

Abstract: A peer-to-peer protocol is based on the use of global timestamps and client priorities in serializing modifications to a shared workspace of real-time collaboration. The method caters to dynamic clients wherein a client can leave or join an ongoing collaboration session as long as there is always at least one client present/remaining in the collaboration session. The method can support multiple definitions of a modification, including partitioning-based definitions, wherein the method provides full support for locking of partitions, and a full treatment of inter-partition synchronisation via a modification definition over multiple partitions. The method is capable of utilizing the many standard methods of creating a global, distributed, synchronized clock for the global timestamps utilized by it. The method is rollback-based for correcting tentative but incorrect serializations, and provides additional backup in terms of checkpoints for additional safety and for the support of lightweight, pervasive clients.

Abstract: A static, double-edge-triggered flip-flop has an upper data path and a lower data path connected between a data input node and an output terminal. The upper path includes a switch connected to a first data loop, and the lower path includes a switch connected to a second data loop. The first and second data loops share a forward path having a data-inverting circuit. In addition, each loop has a feedback path which contains only one element in the form of a switch. However, no data-inverting circuit is included in either of the feedback paths. Advantageously, all the elements of the flip-flop may be constructed using MOSFET transistors implemented according to any one of a variety of semiconductor technologies. In more than one particularly advantageous embodiments, the flip-flop is constructed using a total of twelve transistors.

Abstract: Effective real-time collaboration across remote sites in which any type of data can be shared in a common work space in a consistent manner is made possible by an annotation layer having multiple distinct modes of operation during a collaborative session with two or more people sharing the same work space. One mode is a discussion mode in which one or more users simply share a common view of the shared data and manipulate the view independent of the shared data. During the discussion mode, all user input is handled by the annotation layer which interprets user inputs to move common cursors, create, move or delete markings and text which, since the annotation layer is transparent, appear over the application. Another mode is an edit mode in which one or more users actually edit the shared data. The applications and the data are synchronized among all clients to display the same view. Manipulating the view includes moving a common cursor and placing markings on the common view using text and/or drawing tools.

Abstract: A method for building a locking, migration, dynamic clients, and dynamic partitions capable distributed server for a real-time collaboration session supports the synchronous creation and deletion of partitions by clients as well as the addition and withdrawal of clients during a current collaboration session. The method is based on history servers for providing a history of modifications so that a newly-added client can compute the current state of a shared workspace. The history servers cache and granularize intermediate modification sequences so that computation space and time are reduced. The method supports migrating partition server(s), history server(s), a creation/deletion server, and a collaboration server to different machines. Partition(s) can be dynamically locked and unlocked and, in an extension of this procedure, creation and deletion of partition(s) can be pre-announced and supported.

Abstract: A distributed server for real-time collaboration is substituted for a centralized server to address the problem of the development of unacceptable communication and computation bottlenecks resulting from the use of a one-software-process-based centralized server running somewhere on the available network. The substitute distributed server improves scaleability of real-time collaboration by being based on multiple, independently-communicating, asynchronous, independent (i.e., no shared memory, data, variables, etc.) software processes. The processes can be distributed to multiple machines throughout the network and run simultaneously in order to avoid the centralized server's bottlenecks. To be used, a distributed server requires a disjoint, fully covering partitioning of a work space, wherein it can handle partition hierarchies and groups comprehensively. The distributed server solution is general because of the ability of distributed servers to work with different definitions of a modification.

Abstract: A method for resolving race conflicts in a loop circuit having a forward path and a feedback path includes enabling and disabling the feedback path in accordance with a phase waveform. The phase waveform may be a system clock, in which case one of two approaches may be used to ensure that data from the feedback loop arrives later in time than data from an input signal line. During the first approach, only the rising edge of a clock signal used to control data flow in the feedback loop is delayed relative to the rising edge of a clock signal that controls data flow in the forward path. During the second approach, both the rising and falling edges of the clock signal are delayed. Through these approaches, the method of the present invention achieves improved performance in terms of power consumption, frequency response, area, and switching capacitance.

Abstract: Efficient support of synchronization in parallel processing is supported by methods for building two data structures that are optimal in that they permit simultaneous access to multiple readers and one writer without using any synchronization constructs such as locks and/or any special instructions. The structures rely on atomic read and/or write of some simple integral values to an underlying, actual or emulated, shared memory. No copying is required of any reader or writer. Each structure comprises a first-in-first-out (FIFO), constant-space, circular queue in which each reader and writer tracks shared bookkeeping data such as queue pointers. Readers are impure since they too write bookkeeping data. If a position of interest for access is beyond a boundary defining legitimate access, say trying to read an empty queue or write to a full queue, then access to the position is prevented until legitimization occurs. One of the structures can use the space in its queue to full capacity.