Why time since last access is inadequate as the only criterion for activity-based relocation:

• Access age is a binary measure. The time since last access of a file is computed by subtracting the time at which the fsppadm enforce command is issued from the POSIX atime in the file's metadata. If a file is opened the day before the fsppadm enforce command, its time since last access is one day, even though it may have been inactive for the month preceding. If the intent of a policy rule is to relocate inactive files to lower tier volumes, it will perform badly against files that happen to be accessed, however casually, within the interval defined by the value of the <ACCAGE> pa-rameter.

• Access age is a poor indicator of resumption of significant activity. Using ACCAGE, the time since last access, as a criterion for relocating inactive files to lower tier volumes may fail to schedule some relocations that should be performed, but at least this method results in less relocation activity than necessary. Using ACCAGE as a criterion for relocating previously inactive files that have become active is worse, because this method is likely to schedule relocation activity that is not warranted. If a policy rule's intent is to cause files that have experienced I/O activity in the recent past to be relocated to higher performing, perhaps more failure tolerant storage, ACCAGE is too coarse a filter. For example, in a rule specifying that files on tier2 volumes that have been accessed within the last three days should be relocated to tier1 volumes, no distinction is made between a file that was browsed by a single user and a file that actually was used intensively by applications.