Abstract: Systems and methods for RAID data storage in which data is written across a subset of the RAID drives, where the subset is selected based on drive performance. For instance, if a write will use N?2 of a total of N drives, the system may be configured to determine the two most heavily loaded drives (e.g., based on the respective weighted queue depths of the drives), and may exclude these drives from the write. The data may then be written to the remaining N?2 drives. The system may be configured to determine the RAID encoding for each write request independently of other writes, so the number of drives which are excluded may vary between write requests.

Abstract: Systems and methods for RAID data storage in which each write request identifies a user address and a data length, the system determining a RAID encoding that meets the user's service level requirements, selecting disks to which the data will be written, and writing the data to the disks using the identified RAID encoding. The system may store the metadata for the write in a metadata tree in which the key includes the user address and data length, and the corresponding value includes the physical address(es) of the data on the disks and the RAID encoding used to write the data. The system may use less than all of the disks to store the data, and different writes may use different RAID encodings and different disks (or different numbers of disks), and may be mapped to different addresses on different drives.

Abstract: A scheduler is disclosed. The scheduler can include a time-wheel structure configured to hold scheduling elements, an enqueuer configured to place a scheduling element on the time-wheel structure, and a delay manager configured to direct the scheduling element through the time-wheel structure and remove the scheduling element from the time-wheel structure. The time-wheel structure can include a plurality of decades that can rotate, and each of the plurality of decades can rotate respectively at one or more different rates of rotation. Multiple scheduling elements can be on the time-wheel structure at least partially during the same time. The scheduling elements can be on different decades or on the same decade. One of the plurality of decades can comprise an entry configured to hold a plurality of scheduling elements.