Abstract: A plurality of requests are received for computing processing. At least some of the plurality of requests are replicated. The requests are replicated based on a fractional replication factor. Each received request and each replicated request are transmitted to a computer resource for processing. At least some embodiments provide the capability for meeting tail latency targets with improved performance and reduced cost.

Abstract: Embodiments include pre-processing work items to be processed by computerized processing elements. Aspects include accessing a performance index, which relates to (dynamic) processing performances of work items as processed by the computerized processing elements. Aspects also include determining a time interval (during which the receiver may group queued work items into a block, according to the accessed performance index. Aspects further includes setting a timer to the determined time interval, to allow the receiver to group work items being queued until that time interval has elapsed, according to the timer set. As a result, a block of grouped work items will be obtained, which can then be passed to a scheduler for subsequent processing by computerized processing elements.

Abstract: A plurality of requests are received for computing processing. At least some of the plurality of requests are replicated. The requests are replicated based on a fractional replication factor. Each received request and each replicated request are transmitted to a computer resource for processing. At least some embodiments provide the capability for meeting tail latency targets with improved performance and reduced cost.

Abstract: Embodiments include pre-processing work items to be processed by computerized processing elements. Aspects include accessing a performance index, which relates to (dynamic) processing performances of work items as processed by the computerized processing elements. Aspects also include determining a time interval (during which the receiver may group queued work items into a block, according to the accessed performance index. Aspects further includes setting a timer to the determined time interval, to allow the receiver to group work items being queued until that time interval has elapsed, according to the timer set. As a result, a block of grouped work items will be obtained, which can then be passed to a scheduler for subsequent processing by computerized processing elements.

Abstract: Embodiments include pre-processing work items to be processed by computerized processing elements. Aspects include accessing a performance index, which relates to (dynamic) processing performances of work items as processed by the computerized processing elements. Aspects also include determining a time interval (during which the receiver may group queued work items into a block), according to the accessed performance index. Aspects further includes setting a timer to the determined time interval, to allow the receiver to group work items being queued until that time interval has elapsed, according to the timer set. As a result, a block of grouped work items will be obtained, which can then be passed to a scheduler for subsequent processing by computerized processing elements.

Abstract: A plurality of requests are received for computing processing. At least some of the plurality of requests are replicated. The requests are replicated based on a fractional replication factor. Each received request and each replicated request are transmitted to a computer resource for processing. At least some embodiments provide the capability for meeting tail latency targets with improved performance and reduced cost.

Abstract: A computer-implemented method for preventing the occurrence of performance tickets in a computing system comprising a plurality of virtual machines hosted by a host computing system. The method comprising the steps of: monitoring a resource usage parameter associated with the workload of a virtual machine in order to obtain historic resource usage information; determining information regarding certain dependencies within the historic resource usage information; selecting partial information from the historic resource usage information based on said determined information regarding certain dependencies; estimating the future workload of the virtual machines based on said selected partial information; and resizing the computational resources of one or more virtual machines based on the estimated future workload such that a usage threshold of a computational resource is below a given threshold in order to avoid the generation of a performance ticket.

Abstract: A computing system comprises one or more multicore processor(s) comprising a set of multiple processing units each operable at a variable frequency, and a main memory operable at a variable frequency. A feedback controller is configured to control the frequency of each processing unit of the set and the frequency of the main memory dependent on a measure representative of a current performance of an application running on one or more of the multiple processing units of the set.

Abstract: The present invention is notably directed to systems and methods for detecting resource contention on a computerized system collocation of computer processes. Most basically, such methods comprise: monitoring data produced by each computer process of a set of computer processes co-located on the computerized system; detecting in the monitored data a change in a behavior of data produced by a given computer process of the set of computer processes; and throttling other computer processes of the set than said given computer process to detect potential resource contention at said given computer process. Such systems and methods advantageously apply to co-located virtual machines.

Abstract: A persistent caching system is provided. The persistent caching system includes a storage system having a caching server for storing data, and a client for accessing the data through a network. The caching server is configured to store the data in a number of virtual memory blocks. The virtual memory blocks refer to an associated memory-mapped file in a file system of the caching server. The caching server is configured to export addresses of the virtual memory blocks to the client. The client is configured to access at least some of the virtual memory blocks through RDMA using the exported addresses. The caching server is configured to page virtual memory blocks being accessed by one or more clients through RDMA to and/or from the memory-mapped files associated with the accessed virtual memory blocks.

Abstract: A plurality of requests are received for computing processing. At least some of the plurality of requests are replicated. The requests are replicated based on a fractional replication factor. Each received request and each replicated request are transmitted to a computer resource for processing. At least some embodiments provide the capability for meeting tail latency targets with improved performance and reduced cost.

Abstract: A device is described. The device includes a daughterboard having a first distinct side and a second distinct side. A first connector on the first distinct side is attached to a first baseboard. A second connector on the second distinct side is attached to a second baseboard. A method for attaching a daughterboard to a second device is described. Connectors are included on opposing distinct sides of the daughterboard. The second device includes baseboards mechanically secured in parallel by lead screw actuators. As part of the method, the lead screw actuators are actuated so as to move the baseboards away from each other. A first connector is attached to one of the baseboards. The lead screw actuators are again actuated so as to move the baseboards toward each other until a second connector is brought into attachment with the other of the baseboards.

Abstract: A computer-implemented method for preventing the occurrence of performance tickets in a computing system comprising a plurality of virtual machines hosted by a host computing system. The method comprising the steps of: monitoring a resource usage parameter associated with the workload of a virtual machine in order to obtain historic resource usage information; determining information regarding certain dependencies within the historic resource usage information; selecting partial information from the historic resource usage information based on said determined information regarding certain dependencies; estimating the future workload of the virtual machines based on said selected partial information; and resizing the computational resources of one or more virtual machines based on the estimated future workload such that a usage threshold of a computational resource is below a given threshold in order to avoid the generation of a performance ticket.

Abstract: A printed circuit board (PCB) assembly includes a first PCB and a second PCB disposed substantially parallel and opposite to each other, such that a second side of the first PCB is opposite to a first side of the second PCB; wherein the second PCB has a first set of side connectors on its first side and a second set of side connectors on its second side, configured for both electrical power supply to and signal communication with the second PCB; the second PCB both electrically and mechanically connected to the second side of the first PCB via a first elastomeric connector; and the second PCB electrically connected to the first PCB via its second set of side connectors and a flexible electrical connector that is electrically connected to the second set of side connectors and the first PCB.

Abstract: Embodiments include methods, computer systems and computer program products for performing superscalar out-of-order processing in software in a computer system. Aspects include: loading opcodes into an analysis thread of the computer system, analyzing opcodes to identify certain non-independent opcode snippets, distributing non-independent opcode snippets to separate threads of computer system, instructing each of separate threads to execute each of non-independent opcode snippets, respectively, and collecting results of executions of each of separate threads by a consolidation thread.

Abstract: Embodiments include methods, computer systems and computer program products for performing superscalar out-of-order processing in software in a computer system. Aspects include: loading opcodes into an analysis thread of the computer system, analyzing opcodes to identify certain non-independent opcode snippets, distributing non-independent opcode snippets to separate threads of computer system, instructing each of separate threads to execute each of non-independent opcode snippets, respectively, and collecting results of executions of each of separate threads by a consolidation thread.

Abstract: Embodiments include pre-processing work items to be processed by computerized processing elements. Aspects include accessing a performance index, which relates to (dynamic) processing performances of work items as processed by the computerized processing elements. Aspects also include determining a time interval (during which the receiver may group queued work items into a block, according to the accessed performance index. Aspects further includes setting a timer to the determined time interval, to allow the receiver to group work items being queued until that time interval has elapsed, according to the timer set. As a result, a block of grouped work items will be obtained, which can then be passed to a scheduler for subsequent processing by computerized processing elements.

Abstract: The present invention is notably directed to systems and methods for detecting resource contention on a computerized system collocation of computer processes. Most basically, such methods comprise: monitoring data produced by each computer process of a set of computer processes co-located on the computerized system; detecting in the monitored data a change in a behavior of data produced by a given computer process of the set of computer processes; and throttling other computer processes of the set than said given computer process to detect potential resource contention at said given computer process. Such systems and methods advantageously apply to co-located virtual machines.

Abstract: The sizing of virtual machines is optimized based on projected performance metrics. All virtual machine configuration resources are normalized by a processing device. The normalized resources for the virtual machine configurations are then stored in a catalogue. An application is then profiled to obtain resource demand estimates for each virtual machine configuration and a base performance is calculated for the application. The base performance is used to predict performance estimates on all virtual machine configurations in the catalogue. Accordingly, a virtual machine configuration having a lowest response time is selected.

Abstract: A computing system comprises one or more multicore processor(s) comprising a set of multiple processing units each operable at a variable frequency, and a main memory operable at a variable frequency. A feedback controller is configured to control the frequency of each processing unit of the set and the frequency of the main memory dependent on a measure representative of a current performance of an application running on one or more of the multiple processing units of the set.