Preface Solid-state disk drive If you want to use FTL to try to to the logical address and therefore the physical address betwee...
Preface
Solid-state
disk drive If you want to use FTL to try to to the logical address and
therefore the physical address between the conversion, if the SSD read, write,
delete and other normal add the case of abnormal power-down, may cause mapping
table due to too late to update and lost there's a fault that the SSD can't be
recognized by the system.
At
an equivalent time, so as to enhance read and write performance, usually use
SDRAM to try to to the cache, if the method of reading and writing encountered
abnormal power-down, SDRAM data could also be too late to write down Nand Flash
data loss, or update the mapping table too late to write down Nand Flash The mapping
table is missing.
Abnormal
power outage caused by the phenomenon
SSD
abnormal power outage usually there are three sorts of failure phenomenon:
1,
SSD can't reproduce the system identification, the necessity to rebuild the
mapping table or by simple and crude thanks to re-production are often used;
2,
repeatedly after power outage , SSD there are tons of "new bad
block";
The
mechanism behind the new block is that when the SSD reads, writes, or erases
unsuccessfully, it'll be identified as a nasty block. Of course, these blocks
aren't really bad blocks, simply because they're abnormal power outage caused
by the incorrect judgment.
3,
SDRAM data loss;
Common
power-down protection mechanism
Each
power-down protection mechanism to know different, different for the user, the
protection mechanism is totally different, generally there'll be the subsequent
two practices:
1,
save all the info in SDRAM
Abnormal
power-down, SDRAM all data must be fully written to Nand Flash, generally ,
SDRAM capacity is about to 1000% of the quantity of SSD bare capacity, for
little capacity SSD, SDRAM got to write Nand Flash data Relatively small,
through the super capacitor or tantalum capacitor can still write data.
However, if the SSD capacity is large enough, for example: 8TB, then, SDRAM got
to write Nand Flash data are going to be very large, if still believe super
capacitor or tantalum capacitor to try to to power supply, will inevitably face
the subsequent three tricky problem:
a,
the necessity for more tantalum capacitor particles to try to to the
protection, within the actual engineering practice, this is often a really
serious test, the engineers face the thickness, the quality size limit, PCB
area isn't enough to use;
b,
albeit there's enough capacitance to try to to the protection, when the
implementation of the "restart", the SSD won't start properly, you
want to first pack up for a few time before restarting, because: SSD got to put
all the tantalum capacitor after power Identified;
c,
when the utilization of a couple of years after the tantalum capacitor or super
capacitor after aging, when the tantalum capacitor power supply can't achieve
the initial design target value, the user still has data loss after power loss
or SSD can't identify the potential risks, if the initial design that's , to
try to to redundant capacitors, then, will return to the matter "b"
death cycle.
It
is gratifying that the issues of b and c are perfect solutions to unravel these
thorny problems only need the engineers enough mind and knowledge only.
2,
only save the SDRAM user data, without saving the mapping table
This
will reduce the utilization of SDRAM and therefore the use of tantalum
capacitors, "do not save the mapping table" doesn't mean that the
mapping table is lost, just don't save the last data write update map, when the
SSD re-power, trying to find the last mapping table to save lots of the new
data written to re-build the mapping table, the drawbacks of this approach
isn't enough mechanism to line the reasonable, then rebuild the mapping table
are going to be longer, SSD takes a while to normal access to normal
For
controllers without SDRAM design, all data is written on to Nand Flash. When
data is lost, the info that's not written to Nand Flash are going to be
returned to the host. If no additional data must be saved, High reliability
requirements of the appliance , no SDRAM design is king, its representative may
be a German industrial brand master, its only drawback is that the performance
isn't ok , in fact, many applications and therefore the need for the very best
performance, and Is "enough" performance.
Test
methods and principles
Specific
test, SSD need as a system disk and as a disk from the 2 cases of testing,
therefore the main disk and do from the disk test method is that the only
difference is that the most disk got to test the pc to power off the machine,
and from the disk only SSD are often on the facility off.
a,
respectively, of the SSD as a blank disc, data is written within the 25% and
50% when writing data, the write data for 85% and 3000, respectively, the
abnormal power down 100% test write data, each down and therefore the power-on
interval of three seconds;
The
principle of writing different capacity data to the disk is: When the SSD write
a particular amount of knowledge , the background began to garbage pickup ,
garbage pickup means the relocation of knowledge , data migration means the
mapping table updates, at this point Abnormal power outage is typically a drag
.
b,
when the traditional write data, the SSD abnormal power-down
c,
when the info is deleted when the abnormal power-down
In
the windows, delete the info also got to perform eight actions, and therefore
the establishment of an equivalent document, the mapping table also got to
update.
d,
when the SSD read file is abnormal power down, test 3000 times, power off interval
of three seconds;
e,
when the traditional shutdown process abnormal power down, test 3000 times;
f,
when the traditional start of the OS abnormal power down, test 3000 times;
No comments