Abstract: A scheme is provided that permits the use of a selectable depacketization module to depacketize data streams. An RTP session manager is responsible for receiving RTP packets from a network and parsing/processing them. A depacketizer module is located using the type of data received on the stream. Thus a specific depacketizer is located at runtime depending on the coding decoding scheme (“codec”) used to compress the incoming data stream. A naming convention is followed in order for a specific depacketizer to be located. The depacketizer receives data that has already been parsed and is in a readable form. The depacketizer outputs this data using a well defined interface. This interface has been designed such that it is generic across a number of codecs. The interface passes all relevant information to the decoder where the actual depacketized data stream will be decompressed. The session manager need not know of any codec details since the depacketizer handles all codec specific issues.

Abstract: A computer program or a computer process is provided by replacing a native computer instruction with a trapping computer instruction which is the size of the native computer instruction and which, when executed, causes a trap to the kernel. A trap handler in the kernel determines that the inserted trapping computer instruction caused the trap and transfers control to a user trap handler. The user trap handler maps the trap site to a patch of computer instructions. When the trapping computer instruction is executed, the trap handler transfers control from the kernel to the user trap handler which in turn transfers control to the patch. Native computer instructions in sufficient proximity to corresponding patches of computer instructions may be replaced with branching computer instructions of the size of the native computer instruction and which transfer control to those corresponding patches.

Abstract: A method and system for transmitting objects in an object-oriented distributed system allows the reconstruction of objects and downloading of code associated with such object reconstruction to be deferred until it is needed by a program. The deferred reconstruction is accomplished by using a Java object called a “marshalled object,” which is a self-describing stream and which is a first-class entity in the system, meaning that it exists within a type system of a programming language and can be accessed and manipulated by instructions written in that language. Such an object can typically be produced by any object that can be passed from one address space to another, and it can be stored, passed to other objects, or used to reconstruct an object of the original type on demand. The reconstruction of an object is deferred until the code holding a marshalled object directly calls the creation of the object using the marshalled object.

Abstract: According to the present invention, each processor in a multi-processor system separates locally generated processor requests and remote processor requests from the snoop queue into two categories. In the first category, are all coherence transactions, both read and write, generated by the local processor, as well as all coherence transactions generated by a remote processor which are write accesses. Each of the transactions in the first category maintain a strict FIFO structure wherein accesses to the cache are performed and retired. In the second category are all coherence transactions generated by a remote processor which are read accesses.

Abstract: Information links, such as hypertext links, found in network based documents, such as those found on the World Wide Web, sometimes break, i.e. fail to result in retrieval the desired document. Techniques are described to record information about such broken links and to provide relatively current information to a user whether a link was broken or, if previously broken, has been restored. A centralized broken link service is also described by which information about the status of links on the network can be gathered. A spider can be utilized to periodically check on whether links reported as broken have been restored.

Abstract: A capacitor, including a substrate, a first interconnect layer disposed upon the substrate and a first insulating layer disposed on the first interconnect layer. A first metal layer is disposed on the first insulating layer and formed as at least two regions, the at least two regions of the first metal layer connected to the first interconnect layer through vias. A second insulating layer is disposed on the first metal layer. A second metal layer is disposed on the second insulating layer and is formed as at least two regions. The capacitor further includes a third insulating layer disposed on the second metal layer, a second interconnect layer disposed on the third insulating layer and connecting to the at least two regions of the second metal layer through vias. Finally, a first terminal is connected to the first interconnect layer and a second terminal is connected to the second interconnect layer.

Abstract: An SNMP trap daemon in a distributed network management system is enhanced by improving the flexibility of the filtering functions to allow more generality in matching trap information for CMIP conversion and more specificity in determining how traps are forwarded to other managers and mapped to CMIP events. Filtering is added to the trap forwarding mechanism so that specific traps can be routed to specific managers. The format of a mapping file is also modified to allow type conversion and string concatenation of values in the incoming traps. This modification allows improved matching capabilities in CMIP conversion. The format is further modified to allow “wildcard” matching so that a separate entry is not needed for each incoming trap.

Abstract: An apparatus for reducing the magnitude of an external clock signal is provided wherein the external clock signal is provided on the motherboard of a computer, the signal is provided onto a plug-in CCA, and the signal must pass through a resistive voltage divider prior to being provided to circuits requiring the reduced magnitude signal.

Abstract: Methods, systems, and articles of manufacture consistent with the present invention provide a program component including a set of instructions native to the system, include in the set of native instructions an instruction to maintain information on use of a particular object, and permit reuse of memory resources corresponding to the particular object based on an indication from a source that the particular object is no longer being used, the source being different from any source used to provide information on use of objects associated with non-native instructions of the program component. Additionally, garbage collection is not permitted during native code operations to read or write data in object fields because during such operations an indication exists that such collection may be inaccurate and could possibly reclaim or relocate objects referenced by native code though not specified as such in the native code stack and global variables.

Abstract: A system for implementing flow control in an information network such as a local area network (LAN) utilizing a Carrier Sense Multiple Access with Collision Detection (CSMA/CD) as specified by the IEEE standard 802.03. The information network transmits computer information such as computer data between various computer systems coupled to the information network. The information network includes information network switches capable of routing information packets received via its input ports to its output ports. The received packets are held in buffers of the output ports before being transmitted via the output ports. A portion of each output port buffer is allocated to each input port. The information network switch provides a control packet such as a PAUSE frame to an upstream source operably coupled to the input port in response to the level occupancy of the portion of the buffer allocated to the input port exceeding a first level.

Abstract: A method enables access of a plurality of data sources by a single access operation where each data source in the plurality of data sources requires a different driver to access the data source so that there is a plurality of different drivers. The method includes using an application programming interface (API) for each driver in the plurality of different drivers. The API is substantially identical for each of the drivers in the plurality of different drivers. A merging driver receives the single access operation, and in response to the single access operation, the merging driver accesses the plurality of different drivers through the API.

Abstract: A graphic user interface for entering a new file directory on a computer uses a single file navigation dialogue window that displays a user selectable list of predetermined default directories, user-defined directories, and previously used directories. The user uses a single action to navigate among the directories listed in the single file navigation window.

Abstract: A method and mechanism for controlling network bandwidth by dynamically determing a current window size for a destination buffer. A first window size is determined for the destination buffer at a first time. Then, a current window size is determined a second time as a function of the first window size, a desired bandwidth value, and a bandwidth used between the first time and the second time. A desired bandwidth for a particular connection is achieved by an adaptive or delta-based approach rather than by use of round trip time (RTT) computations. Bandwidth control can be achieved by a software or a hardware bandwidth control mechanism. Possible applications of the bandwidth control is in a server or a quality of service unit.

Abstract: A system for mostly concurrent compaction in a garbage collection system is disclosed. Objects that have been allocated to a program are relocated by first identifying those variables pointing to a selected set of objects that are in use within a subset of memory. As these pointers are identified, they are added to a data structure. The selection of the objects, identification of the pointers, and addition of the pointers to the data structure may all be performed concurrently with execution of the program. At the same time, a write barrier marks as “dirty” those memory regions in which one or more pointers are modified by the program. A number of locations outside the subset of memory are reserved to be used to store the selected objects. Execution of the program is then suspended. The memory regions marked as “dirty” are examined to identify any further variables pointing to the selected objects. Any such identified variables are added to the data structure.

Abstract: A heat sink assembly includes a heat sink, a circuit board and an integrated circuit package. The heat sink includes torque bars having locking features. The package is located between the heat sink and the circuit board and attached to a die attach area of the circuit board. A heat sink retainer has ends secured in the locking features of the torque bars. Tension in the retainer imparts torque on the heat sink which causes the heat sink to impart a downward force on the package. The retainer presses the heat sink down uniformly on the package. Further, the downward force applied by the heat sink to the package is countered by an equal upward force applied by a base section of the retainer on the circuit board. Since the base section is located directly opposite the die attach area, the upward force is transferred from the base section directly back to the heat sink without imparting any bending force on the circuit board.

Abstract: A flow control nozzle for hot gases includes an input end to receive a hot gas. An output end of the flow control nozzle delivers the hot gas to a surface mount device attached with solder to a printed circuit board. A gas distribution mechanism is positioned between the input end and the output end. A gas flow control mechanism is positioned in the gas distribution mechanism to selectively alter the flow of the hot gas through the gas distribution mechanism such that the gas distribution mechanism and the gas flow control mechanism operate to deliver the hot gas to the output end in a substantially uniform manner that facilitates substantially uniform melting of the solder.

Abstract: A conductive body contact or layer is embedded in the bulk region of a partially depleted SOI device. The contact or layer is connected to the output of a bias voltage generator which generates a substrate bias voltage. The substrate bias voltage is routed to the bulk by way of the contact or layer to bias the threshold voltage of the partially depleted SOI device.

Abstract: The present invention provides an instruction fetch unit aligner. In one embodiment, an apparatus for an instruction fetch unit aligner includes selection logic for selecting a non-power of two size instruction from power of two size instruction data, and control logic for controlling the selection logic.

Abstract: One or more filters may be included in each object implementation in a CORBA distributed object system. Each CORBA server object maintains a registry of filters containing unique identifiers and specifications for each of the filters and the order in which the filters must be applied. The filters execute selected code either before or after the conventional marshaling and unmarshaling which take place during a method invocation in the system. The CORBA client object builds a filter registry, from information that it received from the server. Filters may also be present in the client side of the ORB in order to execute code before and after the marshaling and unmarshaling that takes place in the client side of the ORB and these latter filters are also included in the client filter registry. The client then uses its filter registry to invoke the filters during a subsequent method invocation. The client also receives a time stamp from the server to identify the current filter composition.

Abstract: A network element maintains a local representation of a variable from a network directory service without needing to poll the variable in the directory service. The network element is arranged to accept replication messages from the directory service for updating the local record. By using a replication mechanism, it is possible to maintain the current value of a variable without repeated polling of the variable in the directory service.