Abstract: A system, apparatus, and methods are provided for managing on-demand profiling of one or more instances of a software application executing on a plurality of machines within one or more data centers. During operation, the system executes the one or more instances of the software application on the plurality of machines. Next, the system publishes, to a command channel, a command message that comprises a profiling request, wherein the profiling request specifies a subset of the machines. The system then receives, via a data channel, one or more data messages from the subset of the machines, wherein the data messages comprise data gathered by the subset of the machines in response to receiving the command message. Next, the system then evaluates the performance of the software application by aggregating and processing the data messages. Responsive to detecting an anomaly in the performance, the system then executes one or more remedies.

Abstract: Intelligent adjustment of compression levels of messages being brokered or otherwise handled by message service providers in a cloud-computing environment to strike a balance between the reduction in network usage and storage requirement resources owing to message compression on one hand and the increase in CPU resource usage owing to message compression on the other hand for reducing usage-based charges from the cloud customer's point of view and reducing the equipment and operational cost from the cloud provider's point of view (resulting in the ability of the provider to take on and/or retain customers). Determination of an “optimal” compression level that strikes an appropriate balance among resource usage may include consideration of a current messaging load (e.g., bytes of data per unit time), a current compression level, and current usage levels of one or more of the various resources.

Abstract: The disclosed embodiments provide a system for analyzing data from a monitored system. During operation, the system identifies a difference between a performance of an application and a service-level agreement (SLA) of the application. Next, the system determines a correlation between the performance of the application and a disk input/output (I/O) performance of a data storage device used by the application. When the correlation exceeds a threshold, the system outputs a recommendation to migrate the application between the data storage device and a different type of data storage device.

Abstract: Techniques are provided for assigning application instances to different machines. Resource usage of a particular resource by multiple application instances over a period of time is determined. Two or more application instances may be assigned to the same machine if the combined resource usage of the application instances does not exceed the resource capacity of the machine. One or more different types of resources may be considered for machine assignment, such as CPU, memory, network I/O, and disk I/O.

Abstract: A system, method, and apparatus are provided for performing a transparent hybrid data storage scheme in which data are stored as blocks distributed among one or more flash-based storage devices (e.g., solid state drives) and one or more magnetic storage devices (e.g., magnetic disk drives). Files larger than a given size (e.g., 1 MB) are segmented into blocks of that size and stored on one or more devices; blocks of one file may be stored on devices of different types. Periodically, a utility function calculates utility values for each of some or all stored blocks based on frequency of access to the block, frequency of access of a particular type (e.g., random, sequential), a preference regarding where to store the block or the corresponding file, and/or other factors. Blocks having the highest utility values are subject to migration between devices of different types and/or the same type (e.g., for load-balancing).

Abstract: A system, apparatus, and methods are provided for managing on-demand profiling of one or more instances of a software application executing on a plurality of machines within one or more data centers. During operation, the system executes the one or more instances of the software application on the plurality of machines. Next, the system publishes, to a command channel, a command message that comprises a profiling request, wherein the profiling request specifies a subset of the machines. The system then receives, via a data channel, one or more data messages from the subset of the machines, wherein the data messages comprise data gathered by the subset of the machines in response to receiving the command message. Next, the system then evaluates the performance of the software application by aggregating and processing the data messages. Responsive to detecting an anomaly in the performance, the system then executes one or more remedies.

Abstract: The disclosed embodiments provide a method, apparatus, and system for obtaining user ratings and/or feedback for a software application. During operation, for each of a plurality of jobs executed by a computing system component, wherein each job includes an execution of a corresponding job definition: the system retrieves metadata about the job from the computing system component and calculates an inefficiency metric for the job based on the metadata, wherein a higher inefficiency metric corresponds to a more inefficient job. Next, the system ranks the plurality of jobs based on each job's inefficiency metric and selects one or more top-ranked jobs from the ranking. The system then selects one or more job definitions corresponding to the one or more top-ranked jobs. Next, the system sends optimization requests to users associated with the selected job definitions.

Abstract: The disclosed embodiments provide a system for analyzing data from a monitored system. During operation, the system identifies a difference between a performance of an application and a service-level agreement (SLA) of the application. Next, the system determines a correlation between the performance of the application and a disk input/output (I/O) performance of a data storage device used by the application. When the correlation exceeds a threshold, the system outputs a recommendation to migrate the application between the data storage device and a different type of data storage device.

Abstract: A system, method, and apparatus are provided for performing a transparent hybrid data storage scheme in which data are stored as blocks distributed among one or more flash-based storage devices (e.g., solid state drives) and one or more magnetic storage devices (e.g., magnetic disk drives). Files larger than a given size (e.g., 1 MB) are segmented into blocks of that size and stored on one or more devices; blocks of one file may be stored on devices of different types. Periodically, a utility function calculates utility values for each of some or all stored blocks based on frequency of access to the block, frequency of access of a particular type (e.g., random, sequential), a preference regarding where to store the block or the corresponding file, and/or other factors. Blocks having the highest utility values are subject to migration between devices of different types and/or the same type (e.g., for load-balancing).

Abstract: A system, method, and apparatus are provided for determining an appropriate time to disrupt operation of a computer system, subsystem, or component, such as by shutting it down or taking it offline. Historical measurements of work accumulated on the component at different times are used to generate one or more forecasts regarding future amounts of work that will accumulate at different times. Accumulated work may include all job/tasks (or other executable objects) that have been initiated but not yet completed at the time the measurement is taken, and may be expressed in terms of execution time and/or component resources (e.g., cpu, memory). When a request is received to disrupt component operations, based on an urgency of the disruption a corresponding accumulated work threshold is chosen to represent the maximum amount of accumulated work that can be in process and still allow the disruption, and the disruption is scheduled accordingly.

Abstract: In order to prevent data thrashing and the resulting performance degradation, a computer system may maintain an application-layer cache space to more effectively use physical memory and, thus, significantly improve an application-memory hit ratio and reduce disk input-output operations. In particular, the computer system may maintain a managed memory cache that is separate from a page cache. The managed memory cache may be managed according to predefined caching rules that are separate from the caching rules in the operating system that are used to manage the page cache, and these caching rules may be application-aware. Subsequently, when data for an application is accessed, the computer system may prefetch the data and associated information from disk and store the information in the managed memory cache based on data correlations associated with the application.

Abstract: A system, method, and apparatus are provided for assigning or allocating multiple content objects, within a content page (e.g., web page) or other content collection (e.g., a set of pages), to different content delivery networks for delivery in response to a content request. The objects are ranked by importance (e.g., importance in rendering or presenting the page), and the networks are ranked by performance (e.g., throughput). In order of importance, the objects are assigned to the best-performing network that is “available.” Some or all networks are initially available, and a given network becomes “unavailable” after it has been assigned its portion of the objects (e.g., based on content, number of objects, amount of data, percentage). If a total accumulated cost of delivering the objects exceeds a target before all objects have been allocated, the allocation process may terminate early and the remaining objects may be assigned to the least-expensive network.

Abstract: In order to prevent data thrashing and the resulting performance degradation, a computer system may maintain an application-layer cache space to more effectively use physical memory and, thus, significantly improve an application-memory hit ratio and reduce disk input-output operations. In particular, the computer system may maintain a managed memory cache that is separate from an operating systems' default page cache. The managed memory cache may be managed according to predefined caching rules that are separate from rules used to manage the page cache. Moreover, at least one of the data entries in the managed memory cache may have a page size that is smaller than a minimum page size of the page cache. Furthermore, at least some of the data entries in the managed memory cache may have different page sizes and, more generally, different associated predefined caching rules.

Abstract: In order to prevent data thrashing and the resulting performance degradation, a computer system may maintain an application-layer cache space to more effectively use physical memory and, thus, significantly improve an application-memory hit ratio and reduce disk input-output operations. In particular, the computer system may maintain a managed memory cache that is separate from a page cache. The managed memory cache may be managed according to predefined caching rules that are separate from the caching rules in the operating system that are used to manage the page cache, and these caching rules may be application-aware. Subsequently, when data for an application is accessed, the computer system may prefetch the data and associated information from disk and store the information in the managed memory cache based on data correlations associated with the application.

Abstract: A system, methods, and apparatus are provided for performing capacity planning within a system that experiences high volumes of data having high velocity and high variability. Based on historical traffic, a forecast is generated for one or more relatively coarse time periods (e.g., weeks, days), and is decomposed to yield finer-grained forecasts (e.g., for hours, minutes) by applying a distribution index also generated from historical traffic. Estimated replication latency for the forecast period can be calculated from the traffic forecast and an expected level of replication capacity. Further, a required amount of replication capacity can be determined based on a traffic forecast and a maximum replication latency permitted by a service level agreement (SLA) of an event consumer. In addition, replication headroom can be computed, to identify a maximum level of traffic that can be sustained without violating an SLA and/or a date/time at which a violation may occur.

Abstract: Intelligent adjustment of compression levels of messages being brokered or otherwise handled by message service providers in a cloud-computing environment to strike a balance between the reduction in network usage and storage requirement resources owing to message compression on one hand and the increase in CPU resource usage owing to message compression on the other hand for reducing usage-based charges from the cloud customer's point of view and reducing the equipment and operational cost from the cloud provider's point of view (resulting in the ability of the provider to take on and/or retain customers). Determination of an “optimal” compression level that strikes an appropriate balance among resource usage may include consideration of a current messaging load (e.g., bytes of data per unit time), a current compression level, and current usage levels of one or more of the various resources.

Abstract: A system, apparatus, and methods are provided for dynamically adjusting or modifying the operating frequency of a CPU or other processor component, based on a current performance of an application (or applications) executed by the CPU and a target performance, such as a performance level identified in a service level agreement (SLA) associated with the application(s). Multiple thresholds or zones of performance are defined, which may be altered during operation (e.g., based on a profile or predicted CPU workload), and each threshold/zone is associated with a frequency adjustment to implement when the application's performance triggers the threshold or zone. Illustrative adjustments include maximizing the operating frequency, minimizing it, and scaling it up or down one or more discrete levels within a range of frequencies at which the CPU can operate. Thus, CPU operation is adjusted with a primary goal of satisfying application SLAs and a secondary goal of conserving energy.

Abstract: Intelligent adjustment of compression levels of messages being brokered or otherwise handled by message service providers in a cloud-computing environment to strike a balance between the reduction in network usage and storage requirement resources owing to message compression on one hand and the increase in CPU resource usage owing to message compression on the other hand for reducing usage-based charges from the cloud customer's point of view and reducing the equipment and operational cost from the cloud provider's point of view (resulting in the ability of the provider to take on and/or retain customers). Determination of an “optimal” compression level that strikes an appropriate balance among resource usage may include consideration of a current messaging load (e.g., bytes of data per unit time), a current compression level, and current usage levels of one or more of the various resources.

Abstract: A system, methods and apparatus are provided for determining the locus of a bottleneck in a data transfer between a data receiver (e.g., a client device) and a data sender (e.g., a computer server). The locus may be one of a receiver realm encompassing the data receiver (especially a receiver application that consumes the data), a sender realm encompassing the data sender (especially a sender application that produces the data), and a communication link realm that encompasses the communication link(s) over which the data are conveyed (and possibly network-layer protocols and lower that use the communication link(s)). A monitor entity may employ a state-machine model to represent and track progress of a given data transfer between states, using information collected from the data receiver and data sender to identify state transitions. Given a time at which a transfer was delayed or halted, the monitor outputs the locus of the problem.

Abstract: Intelligent adjustment of compression levels of messages being brokered or otherwise handled by message service providers in a cloud-computing environment to strike a balance between the reduction in network usage and storage requirement resources owing to message compression on one hand and the increase in CPU resource usage owing to message compression on the other hand for reducing usage-based charges from the cloud customer's point of view and reducing the equipment and operational cost from the cloud provider's point of view (resulting in the ability of the provider to take on and/or retain customers). Determination of an “optimal” compression level that strikes an appropriate balance among resource usage may include consideration of a current messaging load (e.g., bytes of data per unit time), a current compression level, and current usage levels of one or more of the various resources.