International Business Machines Corporation

National Security Agency

Linda Marie Duyanovich

Sandeep Madhav Uttamchandani

Kristal T Pollack

Juan C Gomez

William Frank Micka

Robert Wesley Shomler

Boris Klots

James T Brady

David Allan Pease

John Davis Palmer

Toby Lyn Marek

Guillermo Alvarez

Li Yin

David R Nowlen

Elizabeth Suzanne Richards

Michael Leo Walker

James Vernon Carlson

Paul Harold Hilton

A policy-based data management system, method, and apparatus are disclosed. The system, method, and apparatus are configured to operate over a distributed storage system such as a storage area network (SAN). Files to be stored on the network are each assigned a service class and a storage pool based on the application of policies to file attributes such as file name, type, user, etc. The service class and storage pool designations are stored as metadata. Files may be retrieved using the metadata to identify the storage pool where the file is stored, and the service class listed within the metadata may be used to control the manner in which the file is handled. A metadata server may be utilized to provide the appropriate service class of files in response to requests from remote clients that may be of different computing platforms.

System, method, and apparatus for policy-based data management

Disclosed are system and method embodiments for determining the root-causes of a performance objective violation, such as an end-to-end service level objection (SLO) violation, in a large-scale system with multi-tiered applications. This determination is made using a hybrid of component-level snapshots of the state of the system during a period in which an abnormal event occurred (i.e., black box mapping) and of known events and their causes (i.e., white-box mapping). Specifically, in response to a query about a violation (e.g., why did the response time for application aincrease from rto r), a processor will access and correlate the black-box and white-box mappings to determine a short-list of probable causes for the violation.

Diagnostic system and method

Disclosed is an autonomic abnormality detection device having a plurality of agents, a server with a one or more processors, a data storage device and a corrective actions engine. The device is adapted to detect and diagnose abnormalities in system components. Particularly, the device uses agents to track performance/workload measurements of system components and dynamically compiles a history of those performance/workload measurements for each component. In order to detect abnormalities a processor compares current performance/workload measurements for a component to the compiled histories for that component and for other components. The processor can further be adapted to determine possible causes of a detected abnormality and to report the abnormality, including the possible causes, to a corrective actions engine.

System and method for recording behavior history for abnormality detection

A computer and method for problem detection and determination for automated system management in a system, wherein the method comprises monitoring system state, workload, and performance parameters of the system; comparing the monitored parameters against normal system performance behavior of the system, wherein the normal system performance behavior is maintained as a mapping of a system state and workload-to-performance parameters; summarizing performance abnormalities at a specified layer in the system as computation and data-processing attributes, wherein the performance abnormalities comprise deviations from the normal system performance behavior; correlating the performance abnormalities across multiple layers in the system using an attribute-based framework; and communicating a root-cause of the performance abnormalities.

Technique for mapping goal violations to anamolies within a system

Embodiments herein present a method, system, computer program product, etc. for automated management using a hybrid of prediction models and feedback-based systems. The method begins by calculating confidence values of models. Next, the method selects a first model based on the confidence values and processes the first model through a constraint solver to produce first workload throttling values. Following this, workloads are repeatedly processed through a feedback-based execution engine, wherein the feedback-based execution engine is controlled by the first workload throttling values. The first workload throttling values are applied incrementally to the feedback-based execution engine, during repetitions of the processing of the workloads, with a step-size that is proportional to the confidence values. The processing of the workloads is repeated until an objective function is maximized, wherein the objective function specifies performance goals of the workloads.

Self-modulation in a model-based automated management framework

Non-disruptive data migration moves are performed form a source DASD to target DASD in a data subsystem. Both source and target DASDs comprise a cache and a nonvolatile store (NVS). Relative to the host, the target DASD generally operates as a source, managing requests for access to source DASD data by issuing I/O requests to the source DASD. First, the target DASD initiates a background data transfer, where data is copied from the source DASD to the target DASD preparing the target DASD to replace the source DASD. The data transfer includes copying data items directly from the source DASD's cache to that of the target DASD's cache, as well as copying data items from the source DASD's NVS to the target DASD's NVS. During background data transfer, the host can still perform data access operations, such as reads and writes, using the target DASD. Namely, for data already copied from the source DASD to the target DASD, the host reads this data from the target DASD. Host requests to read data not yet copied to the source DASD causes staging of the data from the source DASD to the target DASD. For host write requests, the data to be written to the target DASD, if no source data is needed to complete the write. To satisfy host write requests requiring data not yet copied from the source DASD, the data is copied from the source DASD to the target DASD.

Target DASD controlled data migration move

A multi-node video server system in accordance with the invention comprises disk storage associated with a first node which stores at least a portion of a video presentation in the form of plural data blocks, each data block comprising a viewing time segment of the presentation. RAM buffer in a second node receives and stores data blocks of the video presentation from the disk storage in the first node, upon an initial request by a first viewing terminal to view a portion of the video presentation. A communication module in the second node outputs the data blocks, as an isochronous data stream, to requesting terminals. A host controller is coupled to the video server nodes and receives requests from viewing terminals for the video presentation. The host controller causes the communication module in the second node to connect to a viewing terminal which renders the initial request. Additional viewing terminals requesting the same video presentation are also connected to the same communication module if their requests are received within a time period from the initial request that does not exceed a sum of viewing time segments represented by data blocks stored in the RAM buffer. In this manner, both the first requester and subsequent requesting viewing terminals are enabled to receive the audiovisual presentation directly from the same buffer without requiring additional disk accesses for the subsequent viewers. Avoidance of additional internal video server communications is thus achieved.

Multi-node media server that provides video to a plurality of terminals

Primary and secondary data processing systems are coupled via a communication system. Data storage in both systems is provided by a log structured array (LSA) system that stores data in compressed form. Each time data are updated within LSA, the updated data are stored in a data storage location different from the original data. Selected data recorded in a primary data storage of the primary system is remote dual copied to the secondary system for congruent storage in a secondary data storage, such as for disaster recovery purposes. The primary systems creates a remote copying session. Within such remote copying session, a series of 'pending write update' sessions are ESTABLISHED. All data updated within each pending write update session is a consistency group of data. Within each pending write update session update data are retentively stored in both the primary and secondary systems (such as in a non-volatile cache). Addressability of such update data within a pending write update session is through separate pending write directories. The main (primary and secondary) directory for data is not updated such that all original (not updated) data are addressable. A COMMIT command terminates each pending write update session. Congruence of the sequence of updating in the primary system is maintained in the secondary system updating procedure.

Growing community of inventors

Saratoga, CA, United States of America

Linda Marie Duyanovich