Dell Products Inc.

Emc IP Holding Company LLC

Emc Corporation

Keyur B Desai

Xiaobing Zhang

George Mathew

Dominick J Santangelo

Meeilee Chong

Donna Barry Lewis

Murthy Mamidi

Srisailendra Yallapragada

Shobhan Chinnam

Dhawal Bhagwat

Andrew R Huber

Shefali Gautam

Jerrold Heyman

Fani Jenkins

William Formyduval

Abhishek Pareek

Divya Chukkapalli

Systems and methods for allocating resources are disclosed. Resources such as streams are allocated using a stream credit system. Credits are issued to the clients in a manner that ensure the system is operating in a safe allocation state. The credits can be used not only to allocate resources but also to throttle clients where necessary. Credits can be granted fully, partially, and in a number greater than a request. Zero or negative credits can also be issued to throttle clients.

Stream allocation using stream credits

A cache system that includes a reverse cache and a main cache is disclosed. The reverse cache is configured to identify candidates for insertion into a main cache. The reverse cache stores entries such as fingerprints and index values, which are representations of or that identify data. When the entry has been accessed multiple times or is a candidate for promotion based on operation of the reverse cache, data corresponding to the entry is promoted to the main cache. The main cache is configured to evict entries using recency, frequency, and time-adjustments. The main cache and the reverse cache may be similarly configured with a recent list and a frequent list but operate differently.

Methods for cache insertion and cache eviction in a cache system that includes a reverse cache and a main cache

Determining a failure recovery line in a distributed scaleout computing system. Each node or process of a distributed system has or is associated with a vector clock that includes a logical clock for each node in the distributed system. When failure is detected, a recovery operation may be performed using the vector clock. After the recovery operation, the vector clock is updated such that the failure recovery line is available in the computing system.

Failure recovery in a scaleout system using vector clock

A reverse cache for inserting data into a main cache is disclosed. The reverse cache is configured to identify candidates for insertion into a main cache. The reverse cache stores entries such as fingerprints and index values, which are representations of or that identify data. When the entry has been accessed multiple times or is a candidate for promotion based on operation of the reverse cache, data corresponding to the entry is promoted to the main cache.

Methods for cache insertion using ghost lists

A system provides affinity-based load-balancing distribution of similarity groups. A system generates service hash values by applying a consistent hash function to identifiers of deduplication object services. The system maps the deduplication object services to locations where the service hash values map on a hash ring. The system generates, for each of multiple similarity groups, a corresponding group hash value by applying the consistent hash function to a corresponding group identifier. The system assigns, each of the multiple similarity groups, to one of the deduplication object services which is mapped to one of the locations on the hash ring, which is a first successor to a location where the corresponding group hash value maps on the hash ring, thereby enabling each deduplication object service to deduplicate a balanced load of data file segments identified by assigned similarity groups.

Affinity-based load-balancing distribution of similarity groups

One example method includes a cache eviction operation. Entries in a cache are maintained in an entry list that includes a recent list, a recent ghost list, a frequent list, and a frequent ghost list. When an eviction operation is initiated or triggered, timestamps of last access for the entries in the entry list are adjusted by corresponding adjustment values. Candidate entries for eviction are identified based on the adjusted timestamps of last access. At least some of the candidates are evicted from the cache.

Ghost list cache eviction

One example method includes determining a hazard value such as a Weibull hazard function value for an operation such as a recovery operation. Enablement and disablement of a read ahead cache are performed based on the hazard value. The hazard value is representative of whether accesses to a file, such as a backup, being recovered or read are sequential or random. The read ahead cache is enabled when the accesses are sequential or sufficiently sequential.

Managing restore workloads using a hazard function

A cache system that includes a reverse cache and a main cache is disclosed. The reverse cache is configured to identify candidates for insertion into a main cache. The reverse cache stores entries such as fingerprints and index values, which are representations of or that identify data. When the entry has been accessed multiple times or is a candidate for promotion based on operation of the reverse cache, data corresponding to the entry is promoted to the main cache. The main cache is configured to evict entries using recency, frequency, and time-adjustments. The main cache and the reverse cache may be similarly configured with a recent list, a frequent list, and/or ghost lists, but operate differently.

Methods for cache insertion and cache eviction using ghost list in a cache system that includes a reverse cache and a main cache

A reverse cache for inserting data into a main cache is disclosed. The reverse cache is configured to identify candidates for insertion into a main cache. The reverse cache stores entries such as fingerprints, which are representations of data. When the entry has been accessed multiple times or is a candidate for promotion based on operation of the reverse cache, data corresponding to the entry is promoted to the main cache.

Method for optimized cache insertion

One example method includes performing replication operations and/or failure recovery operations for replication operations in a computing system using matrix clocks and replication controllers. The replication controller includes or has access to vectors of a matrix clock. The replication controller also includes a multiplexer. The multiplexer is configured as an m×n multiplexer using the matrix clock to determine m and n. The operation is performed using the configured multiplexer. This allows an imbalance of resources to be managed by the replication controller. The matrix clock can be updated, which may result in the multiplexer being reconfigured.

Failure recovery in a replication environment using a matrix clock and a replication controller

One example method includes performing replication operations and/or failure recovery operations for replication operations in a computing system using matrix clocks. Each node or process in a computing system is associated with a matrix clock that reflects streams required and/or available at the source and destination nodes. As the replication operation is performed, the matrix clocks are updated. When a failure is detected or when insufficient streams are available, the nodes may be throttled until sufficient streams are available. The matrix clocks facilitate a recovery line such that replication can resume when sufficient streams are available.

Failure recovery in a replication environment using matrix clocks

Systems and methods for reconstructing a namespace in response to a loss of data are described herein. To reconstruct the namespace, metadata super segments within a set of file containers may be located in response to a determination that original namespace metadata associated with files stored within the containers is unreadable. Content handles linking the super segment to metadata for each of the files may then be reconstructed from the super segments. Both inode numbers and M-tree identifiers are also extracted from the super segments for each of the files referenced by the super segments. Then, a plurality of U-trees is assembled by aggregating and linking the inode numbers of the files referenced within the super segments based on the M-tree identifiers. Finally, the recovered namespace metadata tree data structure for all of the files within the plurality of LP containers is assembled by aggregating and linking the U-trees.

Growing community of inventors

Cary, NC, United States of America

Keyur B Desai