International Business Machines Corporation

Wisconsin Alumni Research Foundation

Microsoft Technology Licensing, LLC

Ashok Anand

Manish Sethi

Srinivasa Akella

Amitkumar Manoharrao Paradkar

Pradeep Varma

Donald Philip Pazel

Beth Rush Tibbitts

Aaron Gember

Manish Satyapal Gupta

Ramachandran Ramjee

Akshat Verma

George Varghese

Dipayan James Gangopadhyay

Vyas Sekar

Philippe G Charles

Srinivasan Seshan

Bhavish Aggarwal

Robert Grandl

Athula Balachandran

Theophilus Aderemi Benson

Pushkar Chitnis

Chitra Muthukrishnan

Aspects of the present invention provide high-performance indexing for data-intensive systems in which 'slicing' is used to organize indexing data on an SSD such that related entries are located together. Slicing enables combining multiple reads into a single 'slice read' of related items, offering high read performance. Small in-memory indexes, such as hash tables, bloom filters or LSH tables, may be used as buffers for insert operations to resolve slow random writes on the SSD. When full, these buffers are written to the SSD. The internal architecture of the SSD may also be leveraged to achieve higher performance via parallelism. Such parallelism may occur at the channel-level, the package-level, the die-level and/or the plane-level. Consequently, memory and compute resources are freed for use by higher layer applications, and better performance may be achieved.

High-performance indexing for data-intensive systems

A network employing multiple redundancy-aware routers that can eliminate the transmission of redundant data is greatly improved by steering redundant data preferentially into common data paths possibly contrary to other routing paradigms. By collecting redundant data in certain pathways, the effectiveness of the redundancy-aware routers is substantially increased.

Network routing system providing increased network bandwidth

Logical data stores are placed on storages to minimize store request time. The stores are sorted. A store counter and a storage counter are each set to one. (A), (B), and (C) are repeated until the storage counter exceeds the number of storages within the array. (A) is setting a load for the storage specified by the storage counter to zero. (B) is performing (i), (ii), and (iii) while the load for the storage specified by the storage counter is less an average determined load over all the storages. (i) is allocating the store specified by the store counter to the storage specified by the storage counter; and, (ii) is incrementing the load for this storage by this storage's request arrival rate multiplied by an expected service time for the requests of this store. (iii) is incrementing the store counter by one. (C) is incrementing the storage counter by one.

Method and system for placement of logical data stores to minimize request response time

A virtual network virtual machine may be implemented on a cloud computing facility to control communication among virtual machines executing applications and virtual machines executing middlebox functions. This virtual network virtual machine may provide for automatic scaling of middleboxes according to a heuristic algorithm that monitors the effectiveness of each middlebox on the network performance as application virtual machines are scaled. The virtual machine virtual network may also locate virtual machines in actual hardware to further optimize performance.

Cloud-based middlebox management system

An end host redundancy elimination system and method to provide redundancy elimination as an end system service. Embodiments of the system and method use optimization techniques that reduce server central processing unit (CPU) load and memory footprint as compared to existing approaches. For server storage, embodiments of the system and method use a suite of highly-optimized data structures for managing metadata and cached payloads. An optimized asymmetric max-match technique exploits the inherent structure in data maintained at the server and client and ensures that client processing load is negligible. A load-adaptive fingerprinting technique is used that is much faster than current fingerprinting techniques while still delivering similar compression. Load-adaptive means that embodiments of the fingerprinting technique can adapt CPU usage depending on server load. Embodiments of the system and method operate above the transmission control protocol (TCP) layer, thereby reducing the number of roundtrips needed for data transfer.

Asymmetric end host redundancy elimination for networks

Techniques for co-relating at least one of a functional design and at least one implementation artifact of a solution with at least one infrastructure component of a target deployment environment for the solution are provided. The techniques include obtaining at least one of a functional design and at least one implementation artifact of a solution, obtaining at least one infrastructure component of a target deployment environment for the solution, and co-relating at least one of a functional design and at least one implementation artifact of a solution with at least one infrastructure component of a target deployment environment for the solution, wherein co-relating comprises discovering at least one system level dependency among the at least one of a functional design and at least one implementation artifact and the at least one infrastructure component.

Determining system level dependencies

A network employing redundancy-aware hardware may actively allocate decompression tasks among different devices along a single path to improve data throughput. The allocation can be performed by a hash or similar process operating on a header of the packets to distribute caching according to predefined ranges of hash values without significant additional communication overhead. Decompression of packets may be similarly distributed by marking shim values to match the earlier caching of antecedent packets. Nodes may use coordinated cache sizes and organizations to eliminate the need for separate cache protocol communications.

Architecture and system for coordinated network-wide redundancy elimination

A method for preventing deadlock in a distributed computing system includes the steps of: receiving as input a sorted set of containers defining a unique global sequence of containers for servicing process requests; populating at least one table based at least in part on off-line analysis of call graphs defining corresponding transactions for a given order of the containers in the sorted set; storing within each container at least a portion of the table; and allocating one or more threads in a given container according to at least a portion of the table stored within the given container.

Prevention of deadlock in a distributed computing environment

A system for preventing deadlock in a distributed computing system includes a memory and at least one processor coupled to the memory. The processor is operative: to receive as input a sorted set of containers defining a unique global sequence of containers for servicing process requests; to populate at least one table based at least in part on off-line analysis of call graphs defining corresponding transactions for a given order of the containers in the sorted set; to store within each container at least a portion of the at least one table; and to allocate one or more threads in a given container according to at least a portion of the at least one table stored within the given container.

A system (and method, and method for deploying computing infrastructure) for facilitating fixing program development errors found during different phases of program development using different programming tools, such as compilers, linkers, library managers, and debuggers, and can be integrated into a program development environment encompassing all tools. The exemplary aspects of the present invention can detect a qualified set of problems from which to recover with sufficient information, as typically provided by a user or other intelligent or artificial intelligent means, and upon error detection, can initiate a communication with the user (or other support software), to acquire information that could remedy the problem.

System and method for interactive and integrated software development process and phases

The provision of methods and apparatus to determine package identity for an application build. The idea is based on providing a mapping of package with affiliated attributes such as paths used for include or linking, along with reverse mappings. This mapping could be provided through any number of conventional environments such as program development environments or operating system registries: Tools would access this information through a programmatic interface, and use that information to identify packages affiliated with some information. For example, a symbol in a program would be defined in some file. That file would be in a path, which maps to the package identity. Thus, the symbol definition can be conceptually affiliated with a package.

Systems and methods for determining software package identity during a system build

Methods and arrangements for managing error messages from a set of programming tools within a program development environment, such as compilers, linkers, library builders, etc. The error messages are arranged as a directed acyclic graph, with the nodes as error messages, and edges as dependency relationships. In accordance with at least one presently preferred embodiment of the present invention, there are broadly contemplated the following: allowing one or more programming tool to report errors to an error manager component; deducing dependencies amongst error messages, based on error dependency definition provided to the error manager; and delivering messages and dependencies to client components. Although framed within the context of program development tools, this solution could be applied to many other tools outside this domain, under suitable and specified requirements.

Growing community of inventors

Madison, WI, United States of America

Ashok Anand