Vendor specific vs raw values
As described on the S.M.A.R.T. page of the hard disks, you may see that the evaluation of attributes are possible by using the vendor specific method (checking the value and the worst fields) or by using a different method which is independent from the manufacturer (checking the data field). How these two different methods work and what is the difference?
Vendor specific values
The current status of the hard disk are constantly examined by many sensors. The raw values measured are stored (in the “data” field of the corresponding S.M.A.R.T. attribute) and evaluated by an algorythm (coded internally in the firmware of the hard disk).
After this evaluation, the vendor specific result for the attribute is stored in the value field. It indicates how “good” the attribute is (compared to the number in the threshold field. When the hard disk is new (never used before) this value has a theoretical maximum value and then it decreases – until it reaches the threshold level.
This way one attribute is perfect if the following inequality is TRUE:
A – f(r) >= C
Where:
-
A – theoretical maximum value, the “best” possible value for the attribute.
-
f – a function to calculate the decrement based on the raw (r) values. Usually it is a linear function, so r is multiplied by a B constant.
-
C – the vendor specific threshold level, under this level, the attribute is considered problematic.
Using this method has some disadvantages. The A, B, C, values (or the f function) is not defined exactly (these may vary from model to model even on two hard disks from the same manufacturer). Other disadvantage is that the attributes are evaluated independently, the relationship between them is ignored.
In the real life it is possible to find non realistic values which may indicate improper status of the hard disk. Some hard disk manufacturers may define 60-70 years or even more for a hard disk total lifetime when the corresponding attribute is examined. Also it is possible that many thousands of bad sectors are needed until the hard disk will show itself problematic if the above method is used. Of course, many problems may occur long time before reaching this threshold. For example, examining and reallocating data from sectors on a bad (damaged) data area may take lots of time (even hours) – and it may not be completed without problems. During this process, the system is usually not responding, so a problematic hard disk which generates problems again and again may significantly reduce the overall system stability.
Most hard disk analyser applications use the above method to determine the hard disk health status. In many cases, displayed hard disk status is much better than the actual because of the moderate threshold values. Some manufacturers may change the values to show the health status of their driver better than others. This may mislead some softwares or the users.
It is important that after one S.M.A.R.T. attribute reaches the threshold level, the hard disk should run for 24 hours (to run a possible backup). In most cases it is not enough and only partial data can be saved during this time. Saving data from a problematic, almost dead hard disk can be very painful, long and expensive process. It is better to identify and prepare for the problem long time before a catastrophic situation.
Analysing raw values
Hard Disk Sentinel was designed to overcome this situation. The default evaluation methods check the raw values of the S.M.A.R.T. attributes. This way it is possible to determine the hard disk health independently from the manufacturers and models. Note: it is possible to select vendor specific evaluation method in the condition calculation. But the constant use of this method is not recommended (because of the problems described above), as it may give improper results.
When raw values are evaluated, there is no pre-defined (vendor specific) threshold for the attributes. But increasing problems reduce the disk health (constantly and drastically) and this predicts the possible failure and data loss.
The detailed explanation of the calculation methods are described here.