Bull Sas

Bull S.a.

Commissariat a L'energie Atomique Et Aux Energies Alternatives

Georges Lecourtier

Anne Kaszynski

Alain Spir

Claude Brassac

Laurent Cargemel

Jean-Claude Cassonnet

Jean-Louis Fressineau

A cooling system for the electronic components of a computer rack, including a primary circuit of primary liquid; a primary exchanger intended to transfer heat from first electronic components to the primary liquid; a secondary circuit of refrigerant fluid; a primary-secondary exchanger; a tertiary circuit of tertiary liquid; a secondary-tertiary exchanger; and a tertiary-thermal source exchanger. The cooling system also includes a tertiary-secondary exchanger; and an air-tertiary exchanger intended to transfer, to the tertiary liquid, heat from the air of the computer rack heated by second electronic components, the air-tertiary exchanger being arranged, in the tertiary circuit, downstream of the tertiary-secondary exchanger and upstream of the secondary-tertiary exchanger.

Cooling system, cooled computer system and computer facility

A method for the exchange of data between nodes of a server cluster includes a plurality of nodes interconnected together by a geographic interconnection network including a plurality of transmission segments linking the plurality of nodes together according to a predetermined limited number of several different connection directions respectively associated with several different coordinates of a system of coordinates, each transmission segment of the geographic interconnection network thus belonging to a single one of the different connection directions and the system of coordinates thus being defined such that each coordinate of the system of coordinates is associated with a single one of the different connection directions, the method including sending, by a sending node, data intended for at least one other receiving node; transmitting the data using the geographic interconnection network; and receiving the data by each the receiving node.

Method for the exchange of data between nodes of a server cluster, and server cluster implementing said method

A method and system for synchronizing a server cluster having a plurality of nodes each provided with an internal clock and interconnected with each other by a clock interconnection network comprising a plurality of transmission segments. Time-stamping information is generated by the internal clock of a source chosen from the nodes of the server cluster. The time-stamping information is transmitted to all the nodes in the server cluster from the source; and and the internal clock of each node of the server cluster is adjusted from this time-stamping information The time for transmission of the time-stamping information is adjusted by each transmission segment to a constant value established for each transmission segment. Upon receiving the time-stamping information by any one of the nodes in the server cluster, its internal clock is adjusted from the time-stamping information and information relating to the transmission segments passed through between the source and this node.

Method for synchronising a server cluster and server cluster implementing said method

A server () having a plurality of modules (-), each module including a communication element (), a plurality of processors (CPU) (), a system on a chip (SOC) () executing firmware, and a network of programmable ports (FPGA) (). The modules interconnected by an interconnection () between each communication element, an interconnection () between each system on a chip (SOC). The executed firmware produces two software components: a satellite management controller (SMC) component of the system () and a baseboard management controller (BMC) component ().

Server comprising a plurality of modules

A micro power outages compensation module for at least one server includes one or more capacitive storage element for storing electrical energy which is releasable for compensating the outages, the module further including a controller for the charging and/or the discharge of the capacitive storage element(s), limiting the charging and/or discharge current of the capacitive storage element(s) sufficiently to enable hot plugging and/or unplugging of the module even during operation of the server.

Micro power outage compensating module for a server

A method for the exchange of data between nodes of a server cluster comprising a plurality of nodes interconnected together by a geographic interconnection network comprising a plurality of transmission segments linking the nodes together, including sending by a sending node, data intended for at least one other node, or receiving node, transmitting the data using the geographic interconnection network, receiving the data by each receive node, supplying the data to each receiving node with information relating to the transmission segments of the geographic interconnection network traversed between the sending node and this receiving node on the transmission thereof.

A method for synchronously running an application in a high availability environment including a plurality of calculating modules interconnected by a very high-speed broad band network, includes: configuring the modules into partitions including a primary and a secondary partition and a monitoring partition; running the application on each running partition, inputs-outputs processed by the primary partition transmitted to the secondary running partition via the monitoring partition; synchronizing the runnings via exploiting microprocessor context changes; transmitting a catastrophic error signal to the monitoring partition; continuing the running by switching to a degraded mode, the running continuing on a single partition.

Method and device for synchronously running an application in a high availability environment

A direct current backup power supply device of a data processing system with at least one computer supplied with very low voltage whose direct current power supply is connected to an alternating current power supply system using an AC/DC converter of alternating current into direct current includes means for storing electrical energy. The device also includes means for charging the means for storing electrical energy from a part of the direct current provided by the AC/DC converter to the very low voltage power supply of the data processing system, and means for discharging the stored energy in the means for storing electrical energy to the very low voltage power supply of the data processing system, to a noticeably constant predetermined very low voltage, following the detection of a micro-cut of the supply system.

Direct current uninterruptible power supply device for a data-processing system with at least one processor

A system for supplying an electrical system with direct current, the system including at least two direct current power supply devices each having means of electrical connection to an alternating current power supply source and an AC/DC converter of alternating current into direct current, a module for managing the power supply of the electrical system which, after reception of an information of failure of a first active power supply device, transmits a command to start up a second inactive power supply device and a back up electrical energy storage device electrically by-pass connected between the power supply devices and the electrical system. The management module also receives the information of failure from the back up device and/or from the electrical system and transmits the command to start up the second power supply device during a discharge phase of the back up device.

System and method of supplying an electrical system with direct current

The method of management of the hot insertion of an electronic card () in a system () comprises the successive connection of the card to two supply potentials (U, Ui) available in a connector () so as to obtain transient connection signals () during the hot insertion of the card, the detection of the transient signals for providing a binary signal (Vi), one binary state (Vi=1) of which represents the hot insertion of the card, and the use of said state of the binary signal (Vi) for rendering the card operational in the system. The hot insertion detector () includes a bistable logic circuit fed during the connection of the card to said potentials and provided with biasing means adjustable depending on the presence or otherwise of the transient signals.

Hot insertion of a service processor card in a system

An interconnection interface for transmitting digital data between a sending module and a receiving module connected to one another through point-to-point serial links, forming a global data link. Each of the modules are provided, firstly, with a multiplexing unit associated with each of the serial links disposed in the module's physical layer, so as to distribute data transmitted through the global link at a speed determined by the sending module to a given number of parallel links, with each of these parallel links conveying a part of the transmitted data at a speed higher than the sending module data speed; and secondly, with a demultiplexing unit associated to the parallel links conveying received data parts, so as to synchronously and integrally reconstitute these received data parts at a speed accepted by the receiving module.

Interconnection process and interface using parallelized high-speed serial links

A process and interface for interconnection of multiprocessor modules by point-to-point serial-to-parallel links. Data processing or data communication modules (A and B) are interconnected by means of high-speed point-to-point serial links conveying multiplexed information organized into frames comprising a start-of-frame recognition pattern. The process, on transmission and on reception, performs an analog synchronization of the basic clocks of the modules to a reference clock generated by one of the modules designated as a reference module, called the master module, the other modules being called slave modules, and a digital synchronization of the start-of-frame of each slave module to the start-of-frame sent by the master module.

A videoconferencing center ( ) for data communications and teleprocessing is provided which comprises first routing switches having a plurality of sending ports ( ) and a plurality of receiving ports ( ), each sending port adapted to be operatively connected to a receiving data terminal equipment unit ( ) or to a receiving port of second routing switch ( ), each receiving port adapted to be operatively connected to a sending data terminal equipment unit ( ) or to a sending port of another group switching center ( ). Each receiving port is assigned a receiving stack ( ) that is addressable in a single logical address space and each sending port is assigned a sending stack ( ) for containing addresses of said single logical address space. The single logical address reduces the risk of losing data received, which risk results from the necessary limits on physical addresses of stacks. This group switching center is particularly useful for high-speed links of the type that may be required for a videoconferencing system.

Stand alone routing switch and videoconferencing system using the stand alone routing switch

In a multiprocessor system, events capable of modifying the dispatching of processes to the processors (CPU, CPUi) start the execution of dispatching software charged with defining the new dispatching as a function of events and various data. When the number of processors (CPU, CPUi) is large, the microsoftware must be optimized. To that end, the microsoftware is subdivided into a plurality of specialized microprogram modules for determining the dispatching in response to respective categories of particular events. A central microprogram module takes into account the events other than those processed by the specialized modules.

Growing community of inventors

Versailles, France

Georges Lecourtier