Subsystems that affect disk and other I/O

There are two methods of transferring data between memory and disk:

• Normal I/O uses read(S) and write(S) system calls. These can go through the buffer cache using the block device interface or direct to disk using the raw device interface. The read and write calls block the process until they complete -- that is, they are synchronous.
  

• 
  
  
  Asynchronous I/O (AIO) allows non-blocking access to raw disk devices. This allows processes to carry out other tasks while the kernel performs the I/O requests. If possible, you should enable AIO for database systems if they support it; this maximizes transaction throughput and minimizes delays.

  The two forms of AIO supported are the aio(HW) driver and the POSIX.1b aio functions.

  See ``Viewing AIO activity'' for information about how to monitor AIO activity.

  Synchronous I/O operations to the raw disk device force the
  
  process requesting the operation to wait for it to complete. Database applications typically use synchronous I/O to ensure the integrity of the data being written to disk. For example, the journal logs that a database uses to recover in the event of system failure are written to disk using synchronous I/O.

In a similar way, the system maintains a namei cache (for translating names to inodes) of most recently used filenames in order to speed up locating files within filesystems. See ``How the namei cache works'' for a description of its operation.

Finally, the multiphysical buffers use a small pool of memory (generally 160KB to 256KB in size). They are used for various purposes as described in ``How multiphysical buffers are used''.

For a description of how to monitor the activity of block devices including disks, see ``Viewing disk and other block I/O activity''.

Disk I/O and networked filesystem (such as SCO® NFS®) performance are affected by filesystem fragmentation and other filesystem-related factors as described in ``Filesystem factors affecting disk performance''.