* rotational or flash-based devices, and to get the job done quickly
* for applications consisting in many I/O-bound processes.
*
+ * NOTE: if the main or only goal, with a given device, is to achieve
+ * the maximum-possible throughput at all times, then do switch off
+ * all low-latency heuristics for that device, by setting low_latency
+ * to 0.
+ *
* BFQ is described in [1], where also a reference to the initial, more
* theoretical paper on BFQ can be found. The interested reader can find
* in the latter paper full details on the main algorithm, as well as