|
DADDY BOB'S COMPUTER Q & A
Hard
Drive Fragmentation
First,
I’ll have to explain a little about how a
hard is laid out and stores your data. The
easiest way to do this is to make
comparisons with something everyone is
probably familiar with. So, I’ll start out
by comparing the hard drive to a large
building like a warehouse.
When the
hard drive is first bought, it is like the
newly built warehouse, pretty much empty in
side. It has to be prepared to hold whatever
it is you are going to store in it. In the
warehouse, we will lay out many rows. In
each row there will be shelves from top to
bottom, and in these shelves we will create
bins that will hold the product. For this
comparison, all these bins will be the same
size.
Compare
this to the hard drive where the rows become
tracks, the shelves become cylinders, and
the bins become sectors. In the case of the
hard drive, each of these sectors is the
same size capable of holding 256 bytes of
data. This general layout is created on the
hard drive by a process known as formatting.
When a
hard drive is formatted, the tracks,
cylinders, and sectors are laid out
magnetically on the hard drive. As in a new
warehouse, where all the new bins are
completely empty, so it is with the hard
drive, all the sectors are empty. This is
why re-formatting a hard drive destroys or
removes all previously stored data.
Now, to
keep track of where the stuff is stored, we
need an office where the files are kept. In
the hard drive this is called the directory
or catalog. If we place this office in or
near the center of the warehouse, we will
have access to more of the storage space all
around the office than if it were placed at
one end. The same thing goes with the hard
drive. The directory is placed near the
center so the distance needed to travel to
the data is shorter and therefore takes less
time to get there and back.
In the
warehouse we use men, trucks and other
handling tools to store or retrieve the
items. In the hard drive, we use the
“read/write” heads to go back and forth
between the directory and the sectors to
either get (read) data, or store (write) it.
When a
loaded truck comes into the warehouse, there
is a bill of lading that tells what is in
the truck, how much it weighs, and the
warehouse man calculates how much space will
be require to store it. The warehouse man
then allocates a number units required to
hold the items. To make it easier to keep
track of, there are always the same number
of bins in each allocation unit.
The size
of the allocation unit will vary with the
size of the warehouse. If the allocation
unit contains 50 bins, but only 40 are
actually used, then the other 10 are wasted,
and never available for use. Therefore, the
smaller the size of the allocation unit is
the better.
Compare
this to the hard drive, the allocation unit
becomes the cluster, so all space allocated
on the hard drive is done by the cluster.
Now, if the cluster contains 50 sectors but
only 40 are actually used, then the other 10
are wasted and never used. So again, the
smaller the size of the cluster is the
better.
Finally,
I’ll get to fragmentation. In the warehouse,
new items are being received and sent out
all the time. In the hard drive, new data is
being written and read all the time. To keep
things as manageable as possible, space is
always allocated for new arrivals starting
as close to the office/directory as
possible, then outward from there.
When the
first shipment comes in it is decided that
10 units will be needed and the items are
stored in this area. The next shipment has
units allocated the same way and is stored
next to this, and so on until the warehouse
begins to fill up. In the hard drive the
incoming data is stored the same way,
starting at the directory and moving
outwards. So far, so good and no problems.
Now, from
time to time, items are removed leaving
empty bins. These are marked in the office
as being available. In the hard drive, data
is erased or deleted, and the sectors are
marked in the directory as being available.
For the sake of this example, we’ll say that
there are now 10 empty bins/sectors
available close to the office/directory when in comes a new load that will
take 30 bins/sectors.
The 10
bins are allocated here first being closest
to the office, and then the search goes on
for the next available 20 bins. These may be
scattered all over the warehouse. The same
goes for the hard drive. The available
sectors closest to the directory are
allocated first, and then the remaining ones
are allocated from there. This causes the
data to be stored in non adjacent areas, and
it is said to have become fragmented.
The more
that the drive is used; the more data is
written and erased, the more it will become
fragmented. Compare the hard drive’s
read/write head to being a warehouse man
with a small hand truck, and a short memory.
He checks the office, finds the first bin
and goes to it and gets the items. In this
bin, the location of the next bin is
kept, so it is obtained and then the man
goes there. These items are collected, the
location of the next bin is found, and this
continues until all the required items have
been collected.
This
constant traveling, which can be back and
forth all over the warehouse takes time.
The more scattered the items are the longer
it takes to get them all. In the hard drive,
the small hand truck is compared with the
hard drive’s small read/write head. The more
traveling it has to do, moving from sector
to
sector to get the data, the longer
it takes to retrieve it.
After
days, weeks, and months of use, this
constant saving and deleting data of
different size files will cause the hard drive to
become very fragmented. Even if you don’t
think that you are saving or deleting that much
manually, the computer and the applications
it runs are doing this constantly. Just take
a look at the hard drive’s light on your
computer, and you will see it flashing
frequently when you are not doing anything.
The more
fragmented the data is the longer it takes
for the hard drive to retrieve it. So, the
solution is to defrag the hard drive.
Defragging moves all the separated segments
of a file to one place so that it becomes
contiguous. (Defined as Sharing an edge or
boundary; touching; neighboring; adjacent;
connecting without a break.) This reduces
the amounting of traveling necessary, and makes it
faster for the read/write head
to find and retrieve the data.
In modern
computers with their fast CPUs and
memory, the one still slow bottleneck is the
hard drive. Therefore, anything that speeds
up the hard drive's function will enhance the
overall computer’s performance. It becomes evident that keeping your hard drive defragmented is advantageous and something
that should be done periodically.
Fortunately, some programs, like Microsoft’s
Live OneCare, do this automatically when
their “tune up” feature is run.
|
