Abstract: Systems and methods for consistent hashing using multiple hash rings. An example method may comprise: assigning two or more tokens to each node of a plurality of nodes, the two or more tokens belonging to two or more distinct cyclic sequences of tokens, wherein each node is assigned a token within each cyclic sequence; receiving a request comprising an attribute of an object; determining, based on the attribute, a sequence identifier and an object position, the sequence identifier identifying a sequence of the two or more cyclic sequences of tokens, the object position identifying a position of the object within the sequence; and identifying, based on the sequence identifier and the object position, a node for servicing the request.

Abstract: A processing device causes a plurality of storage servers to create tenant identifiers for a tenant of a multi-tenant distributed file system that is distributed over the plurality of storage servers, the plurality of storage servers having underlying local file systems that interface with the multi-tenant distributed file system. The processing device maps the tenant to the tenant identifiers. The processing device assigns a global storage quota to the tenant. The processing device divides the global storage quota into a plurality of local storage quotas and assigns a local storage quota of the plurality of local storage quotas to each storage server of the plurality of storage servers.

Abstract: Systems and methods for consistent hashing using multiple hash rings. An example method may comprise: assigning two or more tokens to each node of a plurality of nodes, the two or more tokens belonging to two or more distinct cyclic sequences of tokens, wherein each node is assigned a token within each cyclic sequence; receiving a request comprising an attribute of an object; determining, based on the attribute, a sequence identifier and an object position, the sequence identifier identifying a sequence of the two or more cyclic sequences of tokens, the object position identifying a position of the object within the sequence; and identifying, based on the sequence identifier and the object position, a node for servicing the request.

Abstract: A processing device causes a plurality of storage servers to create tenant identifiers for a tenant of a multi-tenant distributed file system that is distributed over the plurality of storage servers, the plurality of storage servers having underlying local file systems that interface with the multi-tenant distributed file system. The processing device maps the tenant to the tenant identifiers. The processing device assigns a global storage quota to the tenant. The processing device divides the global storage quota into a plurality of local storage quotas and assigns a local storage quota of the plurality of local storage quotas to each storage server of the plurality of storage servers.

Abstract: In a data storage system, a storage management device includes at least a target mode driver, a data classifier in communication with the target mode driver, a plurality of storage buffers in communication with the target mode driver, a plurality of I/O managers in communication with the data classifier, and a physical store interface in communication with the plurality of I/O managers and with the plurality of storage buffers.

Abstract: A storage management device can receive a write operation that includes a data payload, store a first instance of the data payload at a first storage buffer in the storage management device, and evaluate a first cost equation to identify a second storage buffer in the storage management device, different from the first storage buffer, at which to optimally store a second instance of the data payload.

Abstract: In a multiprocessor system that includes a plurality of processor modules, each one of which includes its own internal clock, one of the plurality of processor modules is designated as a master processor module having a master internal clock. Each other processor module is designated as a slave processor module having a slave processor module internal clock. Each slave processor module synchronizes its internal clock with the master internal clock.

Abstract: In a multiprocessor system that includes a plurality of processor modules, each one of which includes its own internal clock, one of the plurality of processor modules is designated as a master processor module having a master internal clock. Each other processor module is designated as a slave processor module having a slave processor module internal clock. Each slave processor module synchronizes its internal clock with the master internal clock.

Abstract: A storage management device can receive a write operation that includes a data payload, store a first instance of the data payload at a first storage buffer in the storage management device, and evaluate a first cost equation to identify a second storage buffer in the storage management device, different from the first storage buffer, at which to optimally store a second instance of the data payload.

Abstract: In a data storage system, a storage management device includes at least a target mode driver, a data classifier in communication with the target mode driver, a plurality of storage buffers in communication with the target mode driver, a plurality of I/O managers in communication with the data classifier, and a physical store interface in communication with the plurality of I/O managers and with the plurality of storage buffers.