Acer SmartRAID V Uživatelský manuál stáhnout pdf (Strana 173)

7-5

Theory of Operation

Pipelined Parallel Processing (P

)

Traditional RAID controller designs execute host commands sequentially.

Each command initiates a sequence of hardware activities that are

processed one after the other until the command is complete. If any of

these activities are delayed, hardware components must wait. These

delays compromise performance. DPT’s P

technology addresses this

problem in five important areas:

• Real-Time Control System – P

technology can process a virtually

unlimited number of host commands in parallel. The SmartRAID V

Real-Time Control System (RTCS) breaks each command down into

sub-tasks that are then routed to one of the following hardware

processors on the controller:

– DMA processor

– XOR processor

– SCSI or Fibre Channel interface.

Host Commands

DMA Engine

XOR Engine

SCSI / Fibre Channel

Interface

Controller CPU

• Hardware Request Schedulers – Each hardware processor has an

associated hardware request scheduler that coordinates its queued

requests and manages the queue so that all subtasks are processed

efficiently. The disk hardware request scheduler manages one request

queue for each disk drive attached to the controller. To achieve

maximum throughput from each drive, sorted disk queues are

maintained by grouping requests when possible. This results in an

increased number of read and write operations per drive.

1 2 ... 168 169 170 171 172 173 174 175 176 177 178 ... 255 256

Žádné komentáře

Acer SmartRAID V Uživatelský manuál Strana 173