International Business Machines Corporation

Ajay K Singh

Ranjit Noronha

Sandeep R Patil

Kailas S Zadbuke

Abhishek Dave

Ravindra P Walde

A mechanism is provided in a data processing system for performing a remote direct memory access operation. Responsive to receiving in a network interface controller a hash value of data to be copied from a source address in a source node to a destination address in a destination node in the remote direct memory access operation, the network interface controller performs a lookup operation in a translation protection table in the network interface controller to match the hash value to a hash value for data existing in memory of the destination node. Responsive to the network interface controller finding a match in the translation protection table, the network interface controller completes the remote direct memory access operation without transferring the data from the source node to the destination node.

Boosting remote direct memory access performance using cryptographic hash based approach

A mechanism is provided in a data processing system for on-demand caching in a wide area network (WAN) separated distributed file system or clustered file system. The mechanism monitors file access by a plurality of cache sites in the WAN separated distributed file system or clustered file system. The mechanism identifies access patterns by cache sites. The mechanism shares the access patterns with the plurality of cache sites. A given cache site within the plurality of cache sites combines the access patterns with local access information and identifies files to pre-fetch based on the combined information.

On-demand caching in a WAN separated distributed file system or clustered file system cache

Resiliency in file replication scenarios via maintenance of multiple replication queues with differing susceptibility to various sources of data loss. A plurality of requests are received at a first file system location to write to one or more file system objects at the first file system location. Details about the plurality of requests are recorded in at least two replication queues which differ, at least in part, as to types of failures which will cause data loss. Responsive to detection of data loss in one of the replication queues, details about the plurality of requests are retrieved from another of the replication queues.

File-based cluster-to-cluster replication recovery

A mechanism is provided in a data processing system for reliable asynchronous solid-state device based de-duplication. Responsive to receiving a write request to write data to the file system, the mechanism sends the write request to the file system, and in parallel, computes a hash key for the write data. The mechanism looks up the hash key in a de-duplication table. The de-duplication table is stored in a memory or a solid-state storage device. Responsive to the hash key not existing in the de-duplication table, the mechanism writes the write data to a storage device, writes a journal transaction comprising the hash key, and updates the de-duplication table to reference the write data in the storage device.

Growing community of inventors

Wakad, India

Ajay K Singh