August 5, 2020 • 49 mins
When it comes to long-term data storage on computers, there are two main camps. In this episode, we learn how hard disk drives are different from solid state drives.
Learn more about your ad-choices at https://www.iheartpodcastnetwork.com
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:04):
Welcome to text Stuff, a production from my Heart Radio.
Hey there, and welcome to tech Stuff. I'm your host,
Jonathan Strickland. I'm an executive producer with I Heart Radio
and I love all things tech and I recently had
a couple of listeners right to me and ask if
(00:25):
I could do an episode about solid state drives, which
is a method of data storage. So today we're going
to learn about different ways to store information with computer
systems and what makes each one special. And there are
a lot of different ways that computer scientists have created
to store information, either temporarily or you know, permanently or
(00:48):
semi permanently using computer systems. To go through all of
them and to explain how all of them work would
actually take a few episodes. A lot of them work
in similar ways but with different manifestations. And also a
lot of those methods are actually totally obsolete today, so
we're not gonna go over everything. Instead, we're going to
(01:10):
have a quick refresher on ROM, RAM, cash memory, and
then storage systems. ROM and RAM are both types of
computer memory. The purpose of computer memory is to have
a way to reference instructions quickly to run processes, and
by that I mean for the CPU to be able
to get to the information it needs. Typically, we refer
(01:31):
to memory as being a type of data storage that
a CPU can access directly, as opposed to permanent storage,
which must be retrieved before the CPU can access it.
A processor needs two major things to carry out tasks.
It needs a list of instructions also known as what
to do, and then data that's the stuff you're performing
(01:54):
operations upon. So with an absurdly simple analogy, it would
be like a teacher telling a student, Hey, I'm going
to give you some numbers, and I want you to
add those numbers together. So the student already knows the instructions, right,
They know that they are to add any numbers that
the teacher gives them. Then the teacher gives a list
(02:16):
of numbers, which would be the data in our example,
and the student would carry out the instructions adding them.
Computer processors do something similar, though at a speed and
level of complexity that's a little harder for us to grasp.
But without memory, the processor has nothing to draw upon
to actually do anything. Wrong. Stands for read only memory,
(02:39):
and as the name suggests, this is memory that the
computer can reference, but it doesn't change it or add
to it, at least not under normal circumstances. There's some
extreme exceptions, but we're not really going to get into
them here. So you can think of this like messages
that have been etched in stone, but you lack the
(03:00):
ability or the tools to carve in anything into stone.
So you can read these messages that already exist, but
you can't change them in any way. Typically, read only
memory contains basic instructions that a computer system needs in
order for all of its components to work together and
to boot up. So for a computer to actually be
(03:21):
able to detect and interact with the various components the
physical hardware that make up the computer. That's a necessary
part of ROM, and it really is just a basic
set of instructions that allow everything else to happen, like
going through the initial process of recognizing inputting output devices.
(03:42):
Without those instructions, the computer wouldn't do anything meaningful. It
would just be a bunch of pieces that don't actually
work together. RAM or random access memory, is kind of
like short term memory for humans. This is where a
computer can store information that's elevant to whatever the computer
is doing at that very moment. So if you're running
(04:05):
a program, a computer will load relevant data in RAM
for quick reference. It's kind of like how I write episodes.
I take a lot of notes and then I've got
my notes to refer to when i need to, you know,
reference something. Accessing RAM is fast generally speaking, though there
(04:25):
are some potential bottlenecks. I'll mention those late in this episode.
But what does the random access part of RAM mean?
It means that the processor can access the data on
a RAM chip wherever that data might be physically stored
on that chip, and that accessing any part of the
memory should generally take the same amount of time regardless
(04:48):
of where the data is stored. In contrast, there are
some types of storage that would require a computer to
scan through all the data recorded from the beginning of
the storage until hill it hits the relevant patch of information.
It's kind of like the difference between using a chapter
select on a DVD or Blu Ray or just fast
(05:09):
forward scrubbing through a movie to get to a specific scene.
If you have a DVD or Blue ray that has chapters,
you can just jump right to the relevant section and
you access that specific part of the story instantly. Without chapters,
then you have to go through the whole movie sequentially
to get to the part you actually want. RAM is
(05:29):
more like the chapter select approach. RAM has a limited capacity.
Now this depends on the type of RAM you've got
installing your PC or your computational device. Some machines, like
a lot of PCs, are designed in such a way
that you can upgrade RAM over time. You can add
to it and create greater RAM capacity. But even upgraded,
(05:52):
there will be a limit as to how much data
can exist in RAM at any given time. You can't
just keep updating RAM forever. Motherboards won't accept that. Processors
can't work with it, so there are actual limitations that
are dependent upon outside factors. Even with upgraded RAM, there
is a limit to how much data can exist in
(06:14):
RAM all at a given time. You can't load every
single thing from storage into RAM. It wouldn't make sense
for me to copy all of my sources word for
word in my notes, right because then my notes aren't
notes anymore. They are copies of the original sources, and
I wouldn't really be able to refer to them very quickly.
RAM is also temporary, by which I mean that the
(06:37):
data that is inside RAM only sticks around for as
long as the device is powered. Computer systems dump the
information and RAM whenever the computer shuts down or restarts,
so effectively the memory gets white. RAM is thus a
type of volatile memory that means it works as long
as the power is going to the system. You need
(06:59):
a non volu a toll form of memory, something that's
a more persistent, permanent method to store data in larger
volumes if you want to be able to access it
in subsequent sessions. ROM is non volatile, but then again,
it's also unchangeable, so that doesn't do you any good either.
You need something that you can actually update that is
(07:19):
also non volatile if you want to be able to
hold onto data between sessions. Before I move on to that, though,
I should also mention cash memory c a c ch
E memory. This allows processors to access specific, frequently referred
to data at very fast speeds, faster than RAM. It
(07:41):
has less capacity for storage than RAM does, but it
can hold stuff that the processor is going to need
to refer to a lot in order to complete whatever
the task at hand happens to be. RAM capacity tends
to be in the gigabyte range these days, but CASH
tends to be much lower, like in the megabyte range,
and just for the purposes of clarity, A byte is
(08:03):
a unit of information that's equal to eight bits, and
a bit is a piece of binary information a zero
or a one. A megabyte is essentially one million bytes,
and a gigabyte is essentially one billion bytes. A terabyte
is essentially one trillion bytes. If you were to look
at a computer motherboard, you would see the CPU or
(08:25):
central processing unit, which is what executes the programs, and
physically closest to the CPU would be the cash memory,
which holds data that's going to be referenced frequently by
the CPU. Next would be the RAM, So the CPU
would check for information in cash memory first to see
if it's there. If not, it would send a fetch
request for information stored in RAM to see if it's there.
(08:48):
And if the data isn't there, then the CPU has
to send out a request to fetch data from non
volatile storage. Non Volatile memory is necessary if you want
to save data for longer than the immediate present. The
tradeoff is it takes a processor a little bit longer
to access that data. So in my notes example, let's
say that I'm doing this episode and there's something I
(09:10):
wanted to talk about, but I didn't write it down
in my notes. I do happen to know, however, that
it's in one of the large, dusty books that surround
me at all times. I am cursed with them. So
I would take a book aside, and then I would
start searching through the book to find the relevant information.
And this takes a bit longer than just glancing at
my notes would. And that's kind of what it's like
(09:32):
for a computer to reference information that's stored on a
hard drive or solid state drive. When I was growing up,
my family's first real computer was an Apple to Eat,
and that computer did not have a hard drive. Instead,
you would save information onto five and a quarter inch
floppy disc ets the computer had a disk drive. You
(09:53):
would slide the floppy disk into the disk drive and
then you could access whatever information was stored on it,
or you could save information to it. If the computer
needed to reference something from the disk, everything would be
pretty much put on hold while the computer searched the
disks contents for the specific information, then pull it up
loaded into RAM, and then the computer program could continue.
(10:15):
This process is particularly noticeable if you're running something really
process or intensive like computer game. More complicated games such
as those that have like really nice graphics, take up
a lot of space. From a data perspective, the developers
will typically design a game so that the computer running
the game will load chunks of the game into its memory,
(10:38):
but if you navigate to a new chunk, the computer
has to reference the information in storage and then update everything,
and that leads to the dreaded loading screen, and developers
have found a lot of different ways to kind of
deal with this. A common one was to put in
a loading screen whenever you would go through a door
that represented a major change of environment, such as if
(10:59):
you were to go from the outside world of the
game and enter the inside world, like going into a castle.
Between being outside and inside, you know, when you would
hit that door and you'd say open, you'd get treated
to a loading screen. So part of what we're going
to learn about today is why loading screens are even
a thing and what type of storage results in different
(11:20):
weight times. And let's start with hard disk drives a ka,
the platter based drives. Alright, So back in the day,
we used to store data on either floppy disks or
hard disks. Although floppy disks are really a thing of
the past at this point, unless you're using a truly
old computer system, like a legacy computer system. Hard disks
(11:43):
actually predate floppy disks, and we didn't call them hard
disks originally because there was no floppy disk to refer to.
You wouldn't call one without the other, right, there can
be no good without evil. Well, originally we called these
fixed disks, or sometimes we even were to them as Winchesters.
And no, it wasn't a pair of brothers who went
(12:04):
around attacking supernatural bad guys. In this case, the term
Winchester actually came from IBM. It was IBM computer scientists
who pioneered the design of the platter based hard drive
back in the nineteen fifties, and the code name was Winchester.
But then once floppy disks came along, folks would refer
(12:26):
to fixed discs as hard disks. And then you had
the differentiation. You had the floppy disks, which were external,
then you would insert them into a drive and then
remove them when you were done, and you had hard
disks which stayed inside the computer. So hard disks are
disc shaped there around with a hub or or hole
(12:47):
in the middle, and they are contained within a sealed container,
typically made of something like aluminum. And the reason why
is because aluminum is a material that is non magnetic
under normal conditions. If you went to truly extreme conditions,
you could magnetize aluminum, but it would be well outside
(13:09):
the conditions you would find someone's personal computer in. So
old hard discs could only hold a few megabytes worth
of data and they measured like twenty inches in diameter.
They were huge, you know. The much later there would
be closer to three and a half inches in diameter.
So typically hard disk drives actually have stacks of discs.
(13:30):
It's not just a single disc like a single platter,
it's actually a stack of them, and each platter in
that stack is separated by a small amount of space,
so there's actually free space between each stack. If you
think of one, two, and three, you've got a little
bit of space between each of those. And that's really
(13:51):
important and I'll get into that in a second. But
floppy disks are a disc of thin plastic that has
a coding of magnetic material on top of it, and
the plastic disc is inside an envelope or disket made
of thicker plastic and there have been several sizes of
floppy discs over the years. There were eight inch discs,
(14:14):
five and a quarter inch discs like my Apple to
e had, and then three and a half inch diskts,
which my IBM compatible computer used. The eight and the
five and aquarre inch discs were pretty thin. They were
made out of a thinner plastic material and that gave
us the name floppy disc because they were flexible, though
(14:36):
you were not supposed to bend them in any way
that would possibly ruin everything. In fact, if you really
bent it, you had just destroyed that disk. Uh. This
was a piece of information that would have been useful
to a lot of people back in the early eighties
when they weren't aware that floppy did not mean you
can fold it. But the terminology would become more confusing
(14:58):
later on when three and a half inch discs, which
are in a much thicker plastic case one that is
not flexible, When those came around, it confused everything because
those discs weren't floppy like the five and a quarter
inch ones, So some folks would mistakenly refer to three
and a half inch discs as hard disks, but they
were still a type of external storage. You would insert
(15:20):
a floppy disk into a disk drive, you would read
or write to that disk, and then you could inject
the disc and replace it with another one. And that's
how our old Apple to E worked. If you were
to write a page of text, you would save that
page to a floppy disk for later retrieval because the
computer had no way to store information on it permanently.
Later on, c d s or compact discs would largely
(15:44):
replace the need for floppy disks, particularly when computers began
to include drives that could read or write to c
d s. But by then we were also looking at
computers that had internal storage in the form of a
hard disk drive. And really the hard disk and floppy
disk systems are fairly similar to each other, so rather
than explain how each one works, I'll focus on hard
(16:06):
disks so that we can contrast that with solid state
drives in a little bit. Before we get into any
of that, however, let's take a quick break. The first
thing to get into our heads is that hard disk
(16:26):
drives and floppy drives for that matter, are electro mechanical systems,
so they have moving parts and if you were able
to see through a computer, you know, Superman style, and
you were able to see it's hard drive in motion.
You might think it bears some resemblance to how a
turntable plays the tracks on a vinyl record, And there
(16:48):
is some similarity there, but only to a very superficial point. See,
a vinyl record has physical grooves in it. The groove
is a three dimensional groove with a little ridges and
dips and edges, and the stylus or needle of the
record player vibrates as it travels along this groove, and
those vibrations passed to a piece of electric crystal or
(17:11):
a tiny electromagnet, and that transforms the kinetic energy the
energy get movement into electrical energy, and that electrical signal
then passes on to amplifiers that boost that signal. That
then goes on to speakers and it plays out as
the sound that's on the record. Hard disc doesn't have
a physical groove in it. Instead, it's a platter made
(17:34):
out of something like ceramic glass or an aluminum alloy,
and it has a mirror like finish. In fact, it's
highly reflective. The disc has ferromagnetic particles bonded to it.
These particles, if they are exposed to a magnetic field
become magnetized themselves, and they will hold on to that
(17:55):
magnetic property. So if you create a system where you
give meaning to the specific magnetic orientation of domains of
particles domains or sectors or regions of these particles on
the platter, you can designate that as stuff like zeros
and ones. For example, you could say that a domain
(18:15):
that is magnetized so it aligns in the north direction
is a one, and a domain that's aligned in the
south direction is a zero. And then by applying a
precise magnetic fluctuation to the domains, you could arrange them
into meaningful representations of information. So with a hard disk drive,
you do in fact have a disk, or more likely
(18:36):
several disks or platters in a stack, and there is
a hole or hub in the center of these disks,
and those fit around a spindle. The spindle has a
motor that can spend the disc super fast. I'll get
into how fast in a second. Then you've got a
mechanical arm and this has the little read right heads
(18:56):
and those are transducers. These act kind of like the
knee doll on a turntable, but there have their own
special properties, and this arm can move from the inner
edge of the disc to the outer edge in a
fraction of a second. In fact, it would not be
unusual to have one of these be able to move
between those two edges fifty times per second. The arm
(19:18):
itself is split so that the right head, as in
the w R I T E head, the head that
writes data to the disc, fits on one side of
the platter, and the read head that reads information off
the disc can fit on the other side of the platter.
So the platter will spend between these two heads and
(19:40):
they will be separated from that platter by just a tiny, tiny,
tiny amount of space. This is one of those points
where our vinyl record analogy really breaks down, because it
would be like you would have a stylus or needle
on either side of a record as it plays on
the turntable, and that just doesn't happen. Now. Typically the
arms motion controlled with electro magnets. Then you are likely
(20:03):
dealing with a stack of discs, so you would also
be working with a stack of read write heads mounted
on this arm. You would have one pair of read
write head transducers for every disc on the stack, and
the arms would be separated like timees on a fork,
so they can fit between those spinning disks, and it
(20:23):
gets pretty snug in those hard drives. So if you
had three platters in your hard drive, you would have
six read right heads right, one read and one right
head for each of the three platters. A transducer, by
the way, is a type of electronic device that converts
one form of energy into another form of energy, and
(20:44):
there's a lot of stuff that falls into that category.
It's a broad category. So a microphone has a transducer.
It converts the kinetic energy from air pressure fluctuations a
k a. Sound into electrical signals. A speaker does the
same thing, but in the opposite direction. It takes electrical
signals and converts those into kinetic energy by driving the
(21:06):
diaphragm of a speaker to create fluctuations and air pressure,
and we experience that a sound. A digital thermometer converts
thermal energy into electrical energy, and then that can be
measured and displayed on a little screen. The transducers in
a hard disk drives read right head convert electrical energy
into magnetic energy. The arm positions the read right head
(21:29):
at a very specific point along the spinning platter, and
those platters are spinning at like a hundred seventy miles
per hour two d seventy two kilometers per hour. They
could be spinning at a rate of thirty six hundred RPMs.
Those are the old slow hard disk drives. If you
can believe it, pms ten thousand revolutions per minute. It's
(21:53):
incredible how fast they spend. And the transducer on these
read right heads applies a magnetic fluctuation to that spinning disk,
aligning magnetic particles either in a north or south orientation
to indicate those ones and zeros recording information in binary data. Now,
to read data, a transducer works more or less in reverse.
(22:15):
The arm moves out to a certain distance from the
edge of the disk as the disc spins up, and
the moving magnetic particles traveling below the read right head
induce an electrical signal to flow through the head, which
then can be sent on to a processor. So instead
of making a magnetic flux affect the platter, the actual
(22:37):
magnetic field that is generated by the particles on the
platter affect the transducer. It's a very elegant kind of solution. Now,
The data on a hard disk falls into areas known
as sectors and tracks. You can think of a track
as a concentric circle, sort of like an archery target.
These circles get larger as you get to the outer edge.
(22:59):
It also means that they travel at a different speed.
The outer edge of a record travels at a faster
speed than the inner edge of a record, which doesn't
seem to make sense at first because you think it's
all rotating at the same rate. But you have to
remember that outer edge represents a further distance, so the
outer edge is going further in the same amount of
(23:20):
time as the inner edge, which means it has to
be traveling faster. So so sectors are like wedges within
those concentric circles. If you think of the platter as
like a pie, the sectors would be the slices of
pie along these concentric circles. Mm hmm pie. Tracks are
(23:43):
numbered with zero being the closest to the outermost edge
of the disc, and then counting upward from there until
you get all the way to the inner part of
the disc. Sectors can hold a set number of bites,
like five twelve bites. That's not very many bites. At all,
they have a limited capacity. So in addition, the computer
typically groups certain sectors together into what are called clusters.
(24:07):
The computer has to keep track of which sectors in
which tracks have free space in them before saving a
file to the hard drive. So when you're looking at
a file management system and you see you have limited
space on a computer device that has hard disk drive,
you know that that actually corresponds with actual available physical
space on the platters themselves. Also, one way some machines
(24:31):
organized hard drive platters is in cylinders. So we've got
our stack of hard drive platters right there, all one
round top of the other, separated by a thin amount
of space. If you were to look at one track,
one concentric circle on the top platter, you could imagine
that the corresponding concentric circle on platters two and three
(24:54):
are grouped with that same circle on the top platter,
And then you've got yourself a cylinder for um platter
one down to platter three. They'll remember, these platterers are
not in contact with one another, so it's a virtual cylinder.
Not all computer systems use this method for organizing information
on a hard disc. However, ideally, files get stored on
(25:16):
adjacent sectors within a cluster, or adjacent clusters or clusters
that are vertically aligned within a cylinder. In other words,
the group together kind of geographically. But as hard disk
drive space fills up, that just might not be possible.
You might not have enough adjacent sectors to be able
to do that, and then it becomes necessary for the
(25:37):
drive to store sections of a file's data into different
sectors on the hard disk itself. The system keeps track
of where all these bits of the files are, but
it does mean that the transducer has to move around
a lot more to read all the relevant data stored
on the hard drive in order to send that files
information to the CPU, and that's something that can cause
(25:57):
a little tiny delay. I mentioned and that this is
an incredibly precise technology, which is extra impressive considering the
speeds we're talking about with regard to both the arms
movement and the revolutions per minute of the platters. But
the word precise, ironically doesn't give you an idea of
what I'm talking about with modern hard disk drives. So again,
(26:21):
let's imagine the grooves on a vinyl record album. Those
grooves are typically between point zero four millimeters and point
zero eight millimeters wide, or between forty to eighty microns wide.
The bands of information on a hard disc can measure
less than one hundred nanometers in width. A nanometer is
(26:43):
one billionth of a meter, a micron is just one
millionth of a meter, and a human hair typically has
a width of between eighty thousand to one hundred thousand nanometers.
So imagine that these bands of information are measuring less
than a hundred meters wide. That is incredible. In fact,
(27:03):
if you were to measure one inch in from the
edge of a disk, you could fit around three hundred
thousand tracts of information side by side in that space.
Based on that with we refer to the amount of
data that a hard disc can store on its physical
structure as aerial density, not aerial like doing tricks on
the trapeze, aerial as an a R E A L
(27:27):
part of area, and these days that that can be
greater than a ter a bit per square inch, and
a terra bit, like I said, as a trillion bits,
that's a hundred twenty five billion bytes. By comparison, IBM
S three fifty Raymack disc way back in nineteen fifty
six could only hold two thousand bits per square inch.
(27:50):
The increase in aerial density over time has followed a
path similar to what we see with semiconductors and with
Boar's law. To make all this possible, numerous discoveries and
advancements were acquired. Increasing aerial density meant not just shrinking
down components, but also expanding our understanding of stuff like
(28:11):
quantum effects and magnetism. It would take me an entire
series of podcasts to go through the various parts and
ideas and discoveries that all contributed to our ability to
store this much information on a hard disk drive. But
one bit I do want to mention specifically, just because
it's super cool. So you might know that one of
(28:36):
the challenges of keeping up with Moore's law has to
do with a quantum effect called tunneling. Well, in a
similar way, magnetic storage had its own physical limitation. Once
you try to squeeze the magnetic domains or regions into
smaller physical spaces, once you try to pack those zeros
and ones in even more tightly, you encountered something called
(29:00):
the super paramagnetic effect. Yeah, it's something that Mary Poppins
would pricing about. Super para magnetic anyway. The issue is
that when it's packed into such a small space, the
magnetization of individual domains could end up switching very easily,
particularly if there was any heat applied to the area.
(29:22):
So if your magnetization switches and your storage of information
is dependent upon magnetization, that would mean some of your
zeros would become ones, and some of your ones would
become zeros, so your files would become corrupted and unusable.
But the solution to this problem was actually pretty straightforward. See.
(29:44):
Up to that point, the magnetic direction of those little
domains had been longitudinal with regard to the platter. That
means the magnetic polls pointed along the same plane as
the platter, and scientists decided to change this so that
the magnetic fields were now perpendicular with respect to the platter.
(30:04):
So you can think of the magnetic fields is pointing
up and down from the platter surface, rather than say
forward or backward. This solved the superpara magnetic effect, and
it meant that engineers could increase the aerial density of
disks even further. Now, the fact that we're talking about
a mechanical process means that whenever you want to write
(30:25):
information to a disk, or you want to retrieve information
from a disk, that arm must move into place. The
disc has to spin up, the arm has to go
to each sector to pull up the relevant bits of
data that make up that file, and this takes a
little bit of time, and that explains part of the
delay to get information from storage into the computer's random
(30:47):
access memory where the CPU can make some use of it. Now,
it's not a huge amount of time. Typically, the seek time,
that is the delay between a CPU requesting a file
and when the first bite of data is sent from
storage to memory typically falls in the ten to twenty
millisecond range, so it's not like it's, you know, order
(31:08):
out for a pizza because you just decided to open
up a word document. However, the actual rate at which
a hard drive can deliver data to the CPU is
a different story. This is called the data rate, and
it tends to have a fairly wide range depending on
the hard drive, topping out at around two megabytes per second.
(31:28):
So if you're using a device with a hard drive
like this, you've probably experienced some delays as a CPU
requests data and then waits for it to be delivered.
The bigger and more complex the file, the longer the
weight tends to be, and because hard drives have moving components,
stuff can wear down over time. There are a few
(31:49):
potential points of failure, from the mechanical arm with the
transducers mounted on it, to the spinning motor that turns
the disks, to the alignment of the bladders themselves. If
you have a top that has a physical hard drive
in it and you were to accidentally drop that laptop
while it was working, there's a good chance that you
could dislodge those platters and then you've got a ruined
(32:11):
hard drive. Also, if any dust gets into that hard
disk drive, it can create read write errors or even
be enough to cause the arm to collide with the
hard disk and ruin everything. See these days, the read
write heads might just only be a few nanometers away
from the surface of the disk, so to us, if
we were to look at it, it would seem like
(32:32):
the two pieces are actually in contact with one another,
because that that space between them is so small that
even visible light is too big to show it. The
distance between them is about the distance of the width
of a couple of bands of d N and A.
It's tiny, so a single mote of dust would be
like a gargantuan boulder by comparison, and it's really hard
(32:55):
for me to get my mind wrapped around it, because
once we start talking about this level of scale, I
can kind of conceptualize it, but I can't visualize it now.
That's one of the reasons that these hard disks are
sealed in aluminum cases. It's to protect the platters from
dust and other contaminants. It's also why you should never
(33:16):
open up a hard disk drive unless you're okay with
the fact that it's never going to serve a useful
purpose outside of perhaps being an instruction to others on
how disk drives work or or showing people what it
looks like, because the chances are it's never going to
run again. This is also why stuff like clean rooms
need to exist. Clean rooms are facilities that use powerful
(33:38):
filtration and h VAC systems, along with incredibly strict protocols
to prevent the introduction of dust particles. Stuff like hard
drives and semiconductor chips need to be produced in clean
room facilities to avoid the possibility of even just one
moat of dust getting in there and ruining everything. Hard
disk drives tend to be fairly heavy and they require
(33:59):
a decent amount of power to operate, but they also
are cheap and they tend to be pretty high capacity.
Meaning we've advanced the science of designing hard drives to
a point where you can store an enormous amount of
information on a physical hard drive. But now it's time
for us to turn our attention to the alternative long
term storage solution, that of the solid state drive or
(34:20):
s s D. And when we come back, I'll tell
you all about it. But first let's take another quick
break with solid state drives. Were no longer talking about
mechanical systems. There are no moving parts. We're also no
(34:40):
longer talking about magnetic media, so we're not saving data
by magnetizing small areas on a chip or anything like that.
It's a form of nonvolatile memory, so the data does
stick around even if the computer or device is powered off.
The secret sauce in this case is that an s
s D follows into the storage medium of semiconductor chips.
(35:05):
Ss D chips share some similarities with other chips that
are on your computer. For example, remember when I was
talking about rom and Ram at the beginning of the
episode well. A ROM chip is a microchip that is
physically programmed to carry out specific sets of instructions, including
those necessary to boot up a computer. RAM chips are
microchips that can temporarily hold information for quick reference by
(35:28):
the CPU. The RAM and RAM chips are mounted on
the motherboard that's the main circuit board for a computer,
and SSD is not mounted on the motherboard like a
physical hard drive. It lives separate from the motherboard. It
connects to the motherboard via cables. In fact, if you
had a PC with a hard disk drive, you could
(35:50):
open up your computer, you could disconnect your hard disk
drive and you could install a solid state drive in
its place without really changing anything else in the computer
or The type of storage ss d s provide has
a somewhat confusing name. It's called flash memory. I say
it's confusing because we talked about random access memory or RAM.
(36:13):
But that stuff is volatile, right, It goes away when
you power the device down, that information is gone. But
flash memory and ss d s does not go away.
The data stays put. So we've got two different kinds
of memory here, one of which is actually storage. But
don't blame me because I don't come up with the names.
(36:33):
I'm just reporting them. Flash memory can come in a
couple of different varieties, and the type we find an
SSD drives is nan flash in A and D, but
there's also nor flash. So what the heck is up
with those names and how are these two things different? Well,
let's start with the names because they're based on a
very foundational component of computer science. It harkens back to
(36:57):
logic gates, which depend upon Boolean functions or Boolean algebra.
This is all about binary variables, so it's a variable
that can represent one of two values, like a zero
or a one. Logic gates are the practical implementation of
Boolean algebra. The gates determine what output is sent out
(37:20):
based on the incoming input. And let's use a simple
example with the and gate. The and gate accepts two
variables two variables as input, and the values for those
inputs can be either a zero or a one. The
and gate will output a one only if both inputs
(37:41):
are also ones. So if you feed two one bits
into the input of the hand gate, you get a
one bit as the output. But if you were to
feed in two zeros or a zero and a one,
or alternatively a one and a zero. We do differentiate
between these, then the end gate would produce a zero
(38:04):
as its output. So it's a set of rules. It says,
if I get two ones, I give you a one.
If I get anything else, I give you a zero.
By contrast, and or gate produces a one, also known
as a high output if at least one input is
also a one. So the or gate will send out
a one if the inputs are one in zero, zero
(38:26):
and one or one in one, and it only produces
a zero if the inputs are both zero. So this case,
you give me two zeros, I give you a zero.
You give me any other combination, I give you a one.
Logic gates are a way to build out complex responses
to inputs. They are the instructions that tell the computer
(38:47):
what outcome to produce given a specific input. So let's
talk specifically about nand and nor. These are sort of
the bizarro versions of the and A or gates. A
nand gate produces a one output with every pair of
inputs except for one one. So in other words, if
(39:08):
you give me a zero, zero, a zero one, or
a one zero, I will give you a one output.
If you were to give me one one, I would
give you a zero output. A nore gate will only
produce a one output if both inputs are zero. So
you give me a zero one, a one zero, or
a one one. I give you a big fat zero.
(39:29):
You give me a zero zero. Hey, it's your lucky day.
I give you a one. Now. I've done an episode
on logic gates many years ago to explain how why
these are important, how they work in the realm of
computer science, and what this actually all means, and it
may be time for me to revisit that and to
kind of give it a deeper treatment, because it really
gives you an appreciation of the logical design that you
(39:52):
have to create so that computers will do the stuff
you want them to do. But for now, let's just
put that aside and go back to nand versus NOR
flash memory. With both nand and NOR flash memory cards,
you have transistors arranged in cells. They're laid out in
a grid format, so you've got rows and you've got
(40:14):
columns of transistors. In NOR flash cells, the grids are
wired in parallel to one another, so you can think
of them as being wired side by side. In nand
flash cells. They're wired in series, which means you go
from one to the next one and you wire them
all in order in a sequence. NAND cells have a
(40:36):
greater density of transistors and they also use fewer wires
than NOR cells. They can read and write data faster
than NOR flash memory, so nand flash is great for
the solid state drive, whereas NOR flash tends to be
used for read only purposes, kind of like the wrong
chips on a motherboard. If you were to put an
SSD and a hard disk drive mint for the exact
(40:59):
same drive bay in a computer case next to each other,
like if you were to take out a hard disk
drive at an SSD drive and you put them side
by side, they would look fairly similar there both be
and metal you know, aluminum cases, and they would be
the same size. But the SSD would have no mechanical
parts and it would likely have a good amount of
(41:19):
unused space inside the case. The reason for that is
for a convenience, the solid state drive needs to match
the physical shape and size of the hard disk drive
so that it can fit into the computer case properly,
so it's really just there so it'll it'll be able
to fit the model of the computer case, it's not
(41:41):
necessary for the ss D to actually function. The nand
semiconductor chips in an ss D have transistors arranged in
a grid, which means that the grid has columns and rows,
and a chain of transistors conducting a current would represent
the value of one. A chain that is not conducting
current represents a zero. And at first, before you've stored
(42:04):
any data on a solid state drive, you haven't. You
haven't saved anything to it. All the transistors would be
carrying currents, so they would all be set to one.
Saving data to the drive means that the solid state
drive will actually start to block current to specific transistors
to switch them from a one to a zero. Now,
(42:25):
at each intersection of this column and row, you get
a pair of transistors that form a cell. One of
the two transistors is a control gate and the other
is a floating gate. To quote the house Stuff Works
article on it quote, when current reaches the control gate,
electrons flow onto the floating gate, creating a net positive
(42:47):
charge that interrupts current flow. By applying precise voltages to
the transistors, a unique pattern of ones and zeros emerges.
End quote that article. By the way, was written by
William Harris, not written by me. It's a great article.
I highly recommend reading How solid state Drives Work if
you want to learn more. One big advantage of solid
(43:10):
state drives over hard drives is that with no moving parts,
the computer can access data from any part of the
solid state drive with the same speed as any other part.
There's no arm that needs to move into position, there's
no platter that needs to spend, and this means that
data can move from storage to RAM or into cash
memory much faster than it would with a hard disk drive,
(43:33):
and it's fast enough to make a noticeable difference. And
they also use less power than hard disk drives do,
so that's another bonus. Now, a few years ago, there
were some pretty big differences in storage capacity between hard
drives and solid state drives. For a while, the hard
drive had a really good head start, and so for
(43:53):
a few years if you wanted a lot of storage,
really the hard drive was the way to go. You
can get much higher copa a city hard drives. But
today solid state drives are really caught up and it's
possible to buy a solid state drive with the same
storage capacity as a high capacity hard disk. Drive. However,
solid state drives are much more expensive now. The cost
(44:15):
fluctuates based on numerous market factors, but you're likely to
spend double or more than what it would cost to
get a hard disk drive that has the exact same
storage capacity, so they are much more expensive. Interestingly, solid
state drives actually do wear out over time, despite the
fact that they don't have moving parts. So the application
(44:37):
of voltages on transistors, you know, changing the charge of
those transistors that slowly wears out the transistors, and after
a number of cycles a cycle being going from say
a one to a zero back to a one. After
a certain number of those, the cells will start to
wear out. Now, the typical number of cycles ranges in
(44:59):
the thoul posens of cycles, and computers are really good
at using up available storage space that hasn't been through
a lot of cycles already, so typically you don't have
to worry about the solid state drive giving out before
some other component on your machine gives out. So in
other words, you're far more likely to need to upgrade
your computer due to your processor or something else other
(45:21):
than the solid state drive. It would be unusual for
you to use a solid state drive so long that
that cycle thing becomes a real issue. One interesting thing
to remember is that there's always going to be bottlenecks
for data transfer. You might speed up in one area,
but you will start to find restrictions in other areas.
(45:41):
The limitation might be in the amount of RAM you
have in your machine. The RAMS capacity is going to
limit how much can be loaded into memory, which is
why a lot of folks advocate for adding more RAM
to a machine if you want to make it go faster.
Of course, this only works if the machine actually has
the capability to accept more RAM. You might have a
device where you can't upgrade the RAM, or you might
(46:04):
have a device where you've got as much RAM in
it as the motherboard can support. But then there's also
the bus, and a bus in a computer is a
connection between different components within the computer itself. Can actually
also be external components that are attached to the computer.
Bus is a very generic term, but you can think
of the bus as a highway between two different components,
(46:28):
and data travels down this highway to get from one
to the other. So like the memory to the CPU,
and a lot of devices placed the memory physically close
to the CPU to improve data transfer efficiency. It's kind
of blows my mind to think about that, that the
difference between you know, a centimeter can make a big
difference in and transfer efficiency, and it's it's kind of crazy,
(46:53):
but like a highway, a bus has a capacity limit
to how much data can actually cross it in a
of an instant, a bus will have a limit on
how many bits per second can move across it. So
if you're trying to build a fast PC, you've got
to take a lot of different things into consideration, including
the processors, which might include not just a central processing unit,
(47:16):
but maybe one or more graphics processing units, the RAM
that supports those processors, the storage system you're going to
be using, and more so making your computer go faster.
There are a lot of different approaches you can take.
Adding more rams usually a pretty good one, but switching
to a different kind of storage if you're using a
(47:38):
hard disk drive, if you switch to a solid state drive,
that can really help out a lot. It's also generally
more reliable than a hard disk drive. Fewer failures happen
with them. In general, they're always exceptions, and it's always, always,
always a good idea to back up your data. Back
it up, on an external drive or back it up
(48:00):
on a cloud service. Back it up somewhere just in
case one of those catastrophic failures does happen, you'll still
be able to get to the important information. So I
hope that all of this was useful. It's really interesting stuff.
Like I said, we can do a full episode on
things like logic Gates further down the road. Logic is
one of those things I really enjoy because it's all
(48:21):
about just learning basic rules, and those rules are solid,
Like the only thing that changes is what you're feeding
into those rules. But the rules themselves are dependable. And
in the world I live in now, when I find
something that's dependable, I hug it, I hug logic Gates. Guys. Okay,
(48:43):
well that was weird. If you have any suggestions for
future episodes of tech Stuff, whether it's a technology, a
trend in tech, a person in tech, a company, anything
like that, reach out to me. You can get in
touch over on Twitter. The handle for the show is
tech Stuff HSB you and I'll talk to you again
really soon. Y. Text Stuff is an I Heart Radio production.
(49:11):
For more podcasts from I Heart Radio, visit the I
Heart Radio app Apple podcasts, or wherever you listen to
your favorite shows.
Welcome to text Stuff, a production from my Heart Radio.
Hey there, and welcome to tech Stuff. I'm your host,
Jonathan Strickland. I'm an executive producer with I Heart Radio
and I love all things tech and I recently had
a couple of listeners right to me and ask if
(00:25):
I could do an episode about solid state drives, which
is a method of data storage. So today we're going
to learn about different ways to store information with computer
systems and what makes each one special. And there are
a lot of different ways that computer scientists have created
to store information, either temporarily or you know, permanently or
(00:48):
semi permanently using computer systems. To go through all of
them and to explain how all of them work would
actually take a few episodes. A lot of them work
in similar ways but with different manifestations. And also a
lot of those methods are actually totally obsolete today, so
we're not gonna go over everything. Instead, we're going to
(01:10):
have a quick refresher on ROM, RAM, cash memory, and
then storage systems. ROM and RAM are both types of
computer memory. The purpose of computer memory is to have
a way to reference instructions quickly to run processes, and
by that I mean for the CPU to be able
to get to the information it needs. Typically, we refer
(01:31):
to memory as being a type of data storage that
a CPU can access directly, as opposed to permanent storage,
which must be retrieved before the CPU can access it.
A processor needs two major things to carry out tasks.
It needs a list of instructions also known as what
to do, and then data that's the stuff you're performing
(01:54):
operations upon. So with an absurdly simple analogy, it would
be like a teacher telling a student, Hey, I'm going
to give you some numbers, and I want you to
add those numbers together. So the student already knows the instructions, right,
They know that they are to add any numbers that
the teacher gives them. Then the teacher gives a list
(02:16):
of numbers, which would be the data in our example,
and the student would carry out the instructions adding them.
Computer processors do something similar, though at a speed and
level of complexity that's a little harder for us to grasp.
But without memory, the processor has nothing to draw upon
to actually do anything. Wrong. Stands for read only memory,
(02:39):
and as the name suggests, this is memory that the
computer can reference, but it doesn't change it or add
to it, at least not under normal circumstances. There's some
extreme exceptions, but we're not really going to get into
them here. So you can think of this like messages
that have been etched in stone, but you lack the
(03:00):
ability or the tools to carve in anything into stone.
So you can read these messages that already exist, but
you can't change them in any way. Typically, read only
memory contains basic instructions that a computer system needs in
order for all of its components to work together and
to boot up. So for a computer to actually be
(03:21):
able to detect and interact with the various components the
physical hardware that make up the computer. That's a necessary
part of ROM, and it really is just a basic
set of instructions that allow everything else to happen, like
going through the initial process of recognizing inputting output devices.
(03:42):
Without those instructions, the computer wouldn't do anything meaningful. It
would just be a bunch of pieces that don't actually
work together. RAM or random access memory, is kind of
like short term memory for humans. This is where a
computer can store information that's elevant to whatever the computer
is doing at that very moment. So if you're running
(04:05):
a program, a computer will load relevant data in RAM
for quick reference. It's kind of like how I write episodes.
I take a lot of notes and then I've got
my notes to refer to when i need to, you know,
reference something. Accessing RAM is fast generally speaking, though there
(04:25):
are some potential bottlenecks. I'll mention those late in this episode.
But what does the random access part of RAM mean?
It means that the processor can access the data on
a RAM chip wherever that data might be physically stored
on that chip, and that accessing any part of the
memory should generally take the same amount of time regardless
(04:48):
of where the data is stored. In contrast, there are
some types of storage that would require a computer to
scan through all the data recorded from the beginning of
the storage until hill it hits the relevant patch of information.
It's kind of like the difference between using a chapter
select on a DVD or Blu Ray or just fast
(05:09):
forward scrubbing through a movie to get to a specific scene.
If you have a DVD or Blue ray that has chapters,
you can just jump right to the relevant section and
you access that specific part of the story instantly. Without chapters,
then you have to go through the whole movie sequentially
to get to the part you actually want. RAM is
(05:29):
more like the chapter select approach. RAM has a limited capacity.
Now this depends on the type of RAM you've got
installing your PC or your computational device. Some machines, like
a lot of PCs, are designed in such a way
that you can upgrade RAM over time. You can add
to it and create greater RAM capacity. But even upgraded,
(05:52):
there will be a limit as to how much data
can exist in RAM at any given time. You can't
just keep updating RAM forever. Motherboards won't accept that. Processors
can't work with it, so there are actual limitations that
are dependent upon outside factors. Even with upgraded RAM, there
is a limit to how much data can exist in
(06:14):
RAM all at a given time. You can't load every
single thing from storage into RAM. It wouldn't make sense
for me to copy all of my sources word for
word in my notes, right because then my notes aren't
notes anymore. They are copies of the original sources, and
I wouldn't really be able to refer to them very quickly.
RAM is also temporary, by which I mean that the
(06:37):
data that is inside RAM only sticks around for as
long as the device is powered. Computer systems dump the
information and RAM whenever the computer shuts down or restarts,
so effectively the memory gets white. RAM is thus a
type of volatile memory that means it works as long
as the power is going to the system. You need
(06:59):
a non volu a toll form of memory, something that's
a more persistent, permanent method to store data in larger
volumes if you want to be able to access it
in subsequent sessions. ROM is non volatile, but then again,
it's also unchangeable, so that doesn't do you any good either.
You need something that you can actually update that is
(07:19):
also non volatile if you want to be able to
hold onto data between sessions. Before I move on to that, though,
I should also mention cash memory c a c ch
E memory. This allows processors to access specific, frequently referred
to data at very fast speeds, faster than RAM. It
(07:41):
has less capacity for storage than RAM does, but it
can hold stuff that the processor is going to need
to refer to a lot in order to complete whatever
the task at hand happens to be. RAM capacity tends
to be in the gigabyte range these days, but CASH
tends to be much lower, like in the megabyte range,
and just for the purposes of clarity, A byte is
(08:03):
a unit of information that's equal to eight bits, and
a bit is a piece of binary information a zero
or a one. A megabyte is essentially one million bytes,
and a gigabyte is essentially one billion bytes. A terabyte
is essentially one trillion bytes. If you were to look
at a computer motherboard, you would see the CPU or
(08:25):
central processing unit, which is what executes the programs, and
physically closest to the CPU would be the cash memory,
which holds data that's going to be referenced frequently by
the CPU. Next would be the RAM, So the CPU
would check for information in cash memory first to see
if it's there. If not, it would send a fetch
request for information stored in RAM to see if it's there.
(08:48):
And if the data isn't there, then the CPU has
to send out a request to fetch data from non
volatile storage. Non Volatile memory is necessary if you want
to save data for longer than the immediate present. The
tradeoff is it takes a processor a little bit longer
to access that data. So in my notes example, let's
say that I'm doing this episode and there's something I
(09:10):
wanted to talk about, but I didn't write it down
in my notes. I do happen to know, however, that
it's in one of the large, dusty books that surround
me at all times. I am cursed with them. So
I would take a book aside, and then I would
start searching through the book to find the relevant information.
And this takes a bit longer than just glancing at
my notes would. And that's kind of what it's like
(09:32):
for a computer to reference information that's stored on a
hard drive or solid state drive. When I was growing up,
my family's first real computer was an Apple to Eat,
and that computer did not have a hard drive. Instead,
you would save information onto five and a quarter inch
floppy disc ets the computer had a disk drive. You
(09:53):
would slide the floppy disk into the disk drive and
then you could access whatever information was stored on it,
or you could save information to it. If the computer
needed to reference something from the disk, everything would be
pretty much put on hold while the computer searched the
disks contents for the specific information, then pull it up
loaded into RAM, and then the computer program could continue.
(10:15):
This process is particularly noticeable if you're running something really
process or intensive like computer game. More complicated games such
as those that have like really nice graphics, take up
a lot of space. From a data perspective, the developers
will typically design a game so that the computer running
the game will load chunks of the game into its memory,
(10:38):
but if you navigate to a new chunk, the computer
has to reference the information in storage and then update everything,
and that leads to the dreaded loading screen, and developers
have found a lot of different ways to kind of
deal with this. A common one was to put in
a loading screen whenever you would go through a door
that represented a major change of environment, such as if
(10:59):
you were to go from the outside world of the
game and enter the inside world, like going into a castle.
Between being outside and inside, you know, when you would
hit that door and you'd say open, you'd get treated
to a loading screen. So part of what we're going
to learn about today is why loading screens are even
a thing and what type of storage results in different
(11:20):
weight times. And let's start with hard disk drives a ka,
the platter based drives. Alright, So back in the day,
we used to store data on either floppy disks or
hard disks. Although floppy disks are really a thing of
the past at this point, unless you're using a truly
old computer system, like a legacy computer system. Hard disks
(11:43):
actually predate floppy disks, and we didn't call them hard
disks originally because there was no floppy disk to refer to.
You wouldn't call one without the other, right, there can
be no good without evil. Well, originally we called these
fixed disks, or sometimes we even were to them as Winchesters.
And no, it wasn't a pair of brothers who went
(12:04):
around attacking supernatural bad guys. In this case, the term
Winchester actually came from IBM. It was IBM computer scientists
who pioneered the design of the platter based hard drive
back in the nineteen fifties, and the code name was Winchester.
But then once floppy disks came along, folks would refer
(12:26):
to fixed discs as hard disks. And then you had
the differentiation. You had the floppy disks, which were external,
then you would insert them into a drive and then
remove them when you were done, and you had hard
disks which stayed inside the computer. So hard disks are
disc shaped there around with a hub or or hole
(12:47):
in the middle, and they are contained within a sealed container,
typically made of something like aluminum. And the reason why
is because aluminum is a material that is non magnetic
under normal conditions. If you went to truly extreme conditions,
you could magnetize aluminum, but it would be well outside
(13:09):
the conditions you would find someone's personal computer in. So
old hard discs could only hold a few megabytes worth
of data and they measured like twenty inches in diameter.
They were huge, you know. The much later there would
be closer to three and a half inches in diameter.
So typically hard disk drives actually have stacks of discs.
(13:30):
It's not just a single disc like a single platter,
it's actually a stack of them, and each platter in
that stack is separated by a small amount of space,
so there's actually free space between each stack. If you
think of one, two, and three, you've got a little
bit of space between each of those. And that's really
(13:51):
important and I'll get into that in a second. But
floppy disks are a disc of thin plastic that has
a coding of magnetic material on top of it, and
the plastic disc is inside an envelope or disket made
of thicker plastic and there have been several sizes of
floppy discs over the years. There were eight inch discs,
(14:14):
five and a quarter inch discs like my Apple to
e had, and then three and a half inch diskts,
which my IBM compatible computer used. The eight and the
five and aquarre inch discs were pretty thin. They were
made out of a thinner plastic material and that gave
us the name floppy disc because they were flexible, though
(14:36):
you were not supposed to bend them in any way
that would possibly ruin everything. In fact, if you really
bent it, you had just destroyed that disk. Uh. This
was a piece of information that would have been useful
to a lot of people back in the early eighties
when they weren't aware that floppy did not mean you
can fold it. But the terminology would become more confusing
(14:58):
later on when three and a half inch discs, which
are in a much thicker plastic case one that is
not flexible, When those came around, it confused everything because
those discs weren't floppy like the five and a quarter
inch ones, So some folks would mistakenly refer to three
and a half inch discs as hard disks, but they
were still a type of external storage. You would insert
(15:20):
a floppy disk into a disk drive, you would read
or write to that disk, and then you could inject
the disc and replace it with another one. And that's
how our old Apple to E worked. If you were
to write a page of text, you would save that
page to a floppy disk for later retrieval because the
computer had no way to store information on it permanently.
Later on, c d s or compact discs would largely
(15:44):
replace the need for floppy disks, particularly when computers began
to include drives that could read or write to c
d s. But by then we were also looking at
computers that had internal storage in the form of a
hard disk drive. And really the hard disk and floppy
disk systems are fairly similar to each other, so rather
than explain how each one works, I'll focus on hard
(16:06):
disks so that we can contrast that with solid state
drives in a little bit. Before we get into any
of that, however, let's take a quick break. The first
thing to get into our heads is that hard disk
(16:26):
drives and floppy drives for that matter, are electro mechanical systems,
so they have moving parts and if you were able
to see through a computer, you know, Superman style, and
you were able to see it's hard drive in motion.
You might think it bears some resemblance to how a
turntable plays the tracks on a vinyl record, And there
(16:48):
is some similarity there, but only to a very superficial point. See,
a vinyl record has physical grooves in it. The groove
is a three dimensional groove with a little ridges and
dips and edges, and the stylus or needle of the
record player vibrates as it travels along this groove, and
those vibrations passed to a piece of electric crystal or
(17:11):
a tiny electromagnet, and that transforms the kinetic energy the
energy get movement into electrical energy, and that electrical signal
then passes on to amplifiers that boost that signal. That
then goes on to speakers and it plays out as
the sound that's on the record. Hard disc doesn't have
a physical groove in it. Instead, it's a platter made
(17:34):
out of something like ceramic glass or an aluminum alloy,
and it has a mirror like finish. In fact, it's
highly reflective. The disc has ferromagnetic particles bonded to it.
These particles, if they are exposed to a magnetic field
become magnetized themselves, and they will hold on to that
(17:55):
magnetic property. So if you create a system where you
give meaning to the specific magnetic orientation of domains of
particles domains or sectors or regions of these particles on
the platter, you can designate that as stuff like zeros
and ones. For example, you could say that a domain
(18:15):
that is magnetized so it aligns in the north direction
is a one, and a domain that's aligned in the
south direction is a zero. And then by applying a
precise magnetic fluctuation to the domains, you could arrange them
into meaningful representations of information. So with a hard disk drive,
you do in fact have a disk, or more likely
(18:36):
several disks or platters in a stack, and there is
a hole or hub in the center of these disks,
and those fit around a spindle. The spindle has a
motor that can spend the disc super fast. I'll get
into how fast in a second. Then you've got a
mechanical arm and this has the little read right heads
(18:56):
and those are transducers. These act kind of like the
knee doll on a turntable, but there have their own
special properties, and this arm can move from the inner
edge of the disc to the outer edge in a
fraction of a second. In fact, it would not be
unusual to have one of these be able to move
between those two edges fifty times per second. The arm
(19:18):
itself is split so that the right head, as in
the w R I T E head, the head that
writes data to the disc, fits on one side of
the platter, and the read head that reads information off
the disc can fit on the other side of the platter.
So the platter will spend between these two heads and
(19:40):
they will be separated from that platter by just a tiny, tiny,
tiny amount of space. This is one of those points
where our vinyl record analogy really breaks down, because it
would be like you would have a stylus or needle
on either side of a record as it plays on
the turntable, and that just doesn't happen. Now. Typically the
arms motion controlled with electro magnets. Then you are likely
(20:03):
dealing with a stack of discs, so you would also
be working with a stack of read write heads mounted
on this arm. You would have one pair of read
write head transducers for every disc on the stack, and
the arms would be separated like timees on a fork,
so they can fit between those spinning disks, and it
(20:23):
gets pretty snug in those hard drives. So if you
had three platters in your hard drive, you would have
six read right heads right, one read and one right
head for each of the three platters. A transducer, by
the way, is a type of electronic device that converts
one form of energy into another form of energy, and
(20:44):
there's a lot of stuff that falls into that category.
It's a broad category. So a microphone has a transducer.
It converts the kinetic energy from air pressure fluctuations a
k a. Sound into electrical signals. A speaker does the
same thing, but in the opposite direction. It takes electrical
signals and converts those into kinetic energy by driving the
(21:06):
diaphragm of a speaker to create fluctuations and air pressure,
and we experience that a sound. A digital thermometer converts
thermal energy into electrical energy, and then that can be
measured and displayed on a little screen. The transducers in
a hard disk drives read right head convert electrical energy
into magnetic energy. The arm positions the read right head
(21:29):
at a very specific point along the spinning platter, and
those platters are spinning at like a hundred seventy miles
per hour two d seventy two kilometers per hour. They
could be spinning at a rate of thirty six hundred RPMs.
Those are the old slow hard disk drives. If you
can believe it, pms ten thousand revolutions per minute. It's
(21:53):
incredible how fast they spend. And the transducer on these
read right heads applies a magnetic fluctuation to that spinning disk,
aligning magnetic particles either in a north or south orientation
to indicate those ones and zeros recording information in binary data. Now,
to read data, a transducer works more or less in reverse.
(22:15):
The arm moves out to a certain distance from the
edge of the disk as the disc spins up, and
the moving magnetic particles traveling below the read right head
induce an electrical signal to flow through the head, which
then can be sent on to a processor. So instead
of making a magnetic flux affect the platter, the actual
(22:37):
magnetic field that is generated by the particles on the
platter affect the transducer. It's a very elegant kind of solution. Now,
The data on a hard disk falls into areas known
as sectors and tracks. You can think of a track
as a concentric circle, sort of like an archery target.
These circles get larger as you get to the outer edge.
(22:59):
It also means that they travel at a different speed.
The outer edge of a record travels at a faster
speed than the inner edge of a record, which doesn't
seem to make sense at first because you think it's
all rotating at the same rate. But you have to
remember that outer edge represents a further distance, so the
outer edge is going further in the same amount of
(23:20):
time as the inner edge, which means it has to
be traveling faster. So so sectors are like wedges within
those concentric circles. If you think of the platter as
like a pie, the sectors would be the slices of
pie along these concentric circles. Mm hmm pie. Tracks are
(23:43):
numbered with zero being the closest to the outermost edge
of the disc, and then counting upward from there until
you get all the way to the inner part of
the disc. Sectors can hold a set number of bites,
like five twelve bites. That's not very many bites. At all,
they have a limited capacity. So in addition, the computer
typically groups certain sectors together into what are called clusters.
(24:07):
The computer has to keep track of which sectors in
which tracks have free space in them before saving a
file to the hard drive. So when you're looking at
a file management system and you see you have limited
space on a computer device that has hard disk drive,
you know that that actually corresponds with actual available physical
space on the platters themselves. Also, one way some machines
(24:31):
organized hard drive platters is in cylinders. So we've got
our stack of hard drive platters right there, all one
round top of the other, separated by a thin amount
of space. If you were to look at one track,
one concentric circle on the top platter, you could imagine
that the corresponding concentric circle on platters two and three
(24:54):
are grouped with that same circle on the top platter,
And then you've got yourself a cylinder for um platter
one down to platter three. They'll remember, these platterers are
not in contact with one another, so it's a virtual cylinder.
Not all computer systems use this method for organizing information
on a hard disc. However, ideally, files get stored on
(25:16):
adjacent sectors within a cluster, or adjacent clusters or clusters
that are vertically aligned within a cylinder. In other words,
the group together kind of geographically. But as hard disk
drive space fills up, that just might not be possible.
You might not have enough adjacent sectors to be able
to do that, and then it becomes necessary for the
(25:37):
drive to store sections of a file's data into different
sectors on the hard disk itself. The system keeps track
of where all these bits of the files are, but
it does mean that the transducer has to move around
a lot more to read all the relevant data stored
on the hard drive in order to send that files
information to the CPU, and that's something that can cause
(25:57):
a little tiny delay. I mentioned and that this is
an incredibly precise technology, which is extra impressive considering the
speeds we're talking about with regard to both the arms
movement and the revolutions per minute of the platters. But
the word precise, ironically doesn't give you an idea of
what I'm talking about with modern hard disk drives. So again,
(26:21):
let's imagine the grooves on a vinyl record album. Those
grooves are typically between point zero four millimeters and point
zero eight millimeters wide, or between forty to eighty microns wide.
The bands of information on a hard disc can measure
less than one hundred nanometers in width. A nanometer is
(26:43):
one billionth of a meter, a micron is just one
millionth of a meter, and a human hair typically has
a width of between eighty thousand to one hundred thousand nanometers.
So imagine that these bands of information are measuring less
than a hundred meters wide. That is incredible. In fact,
(27:03):
if you were to measure one inch in from the
edge of a disk, you could fit around three hundred
thousand tracts of information side by side in that space.
Based on that with we refer to the amount of
data that a hard disc can store on its physical
structure as aerial density, not aerial like doing tricks on
the trapeze, aerial as an a R E A L
(27:27):
part of area, and these days that that can be
greater than a ter a bit per square inch, and
a terra bit, like I said, as a trillion bits,
that's a hundred twenty five billion bytes. By comparison, IBM
S three fifty Raymack disc way back in nineteen fifty
six could only hold two thousand bits per square inch.
(27:50):
The increase in aerial density over time has followed a
path similar to what we see with semiconductors and with
Boar's law. To make all this possible, numerous discoveries and
advancements were acquired. Increasing aerial density meant not just shrinking
down components, but also expanding our understanding of stuff like
(28:11):
quantum effects and magnetism. It would take me an entire
series of podcasts to go through the various parts and
ideas and discoveries that all contributed to our ability to
store this much information on a hard disk drive. But
one bit I do want to mention specifically, just because
it's super cool. So you might know that one of
(28:36):
the challenges of keeping up with Moore's law has to
do with a quantum effect called tunneling. Well, in a
similar way, magnetic storage had its own physical limitation. Once
you try to squeeze the magnetic domains or regions into
smaller physical spaces, once you try to pack those zeros
and ones in even more tightly, you encountered something called
(29:00):
the super paramagnetic effect. Yeah, it's something that Mary Poppins
would pricing about. Super para magnetic anyway. The issue is
that when it's packed into such a small space, the
magnetization of individual domains could end up switching very easily,
particularly if there was any heat applied to the area.
(29:22):
So if your magnetization switches and your storage of information
is dependent upon magnetization, that would mean some of your
zeros would become ones, and some of your ones would
become zeros, so your files would become corrupted and unusable.
But the solution to this problem was actually pretty straightforward. See.
(29:44):
Up to that point, the magnetic direction of those little
domains had been longitudinal with regard to the platter. That
means the magnetic polls pointed along the same plane as
the platter, and scientists decided to change this so that
the magnetic fields were now perpendicular with respect to the platter.
(30:04):
So you can think of the magnetic fields is pointing
up and down from the platter surface, rather than say
forward or backward. This solved the superpara magnetic effect, and
it meant that engineers could increase the aerial density of
disks even further. Now, the fact that we're talking about
a mechanical process means that whenever you want to write
(30:25):
information to a disk, or you want to retrieve information
from a disk, that arm must move into place. The
disc has to spin up, the arm has to go
to each sector to pull up the relevant bits of
data that make up that file, and this takes a
little bit of time, and that explains part of the
delay to get information from storage into the computer's random
(30:47):
access memory where the CPU can make some use of it. Now,
it's not a huge amount of time. Typically, the seek time,
that is the delay between a CPU requesting a file
and when the first bite of data is sent from
storage to memory typically falls in the ten to twenty
millisecond range, so it's not like it's, you know, order
(31:08):
out for a pizza because you just decided to open
up a word document. However, the actual rate at which
a hard drive can deliver data to the CPU is
a different story. This is called the data rate, and
it tends to have a fairly wide range depending on
the hard drive, topping out at around two megabytes per second.
(31:28):
So if you're using a device with a hard drive
like this, you've probably experienced some delays as a CPU
requests data and then waits for it to be delivered.
The bigger and more complex the file, the longer the
weight tends to be, and because hard drives have moving components,
stuff can wear down over time. There are a few
(31:49):
potential points of failure, from the mechanical arm with the
transducers mounted on it, to the spinning motor that turns
the disks, to the alignment of the bladders themselves. If
you have a top that has a physical hard drive
in it and you were to accidentally drop that laptop
while it was working, there's a good chance that you
could dislodge those platters and then you've got a ruined
(32:11):
hard drive. Also, if any dust gets into that hard
disk drive, it can create read write errors or even
be enough to cause the arm to collide with the
hard disk and ruin everything. See these days, the read
write heads might just only be a few nanometers away
from the surface of the disk, so to us, if
we were to look at it, it would seem like
(32:32):
the two pieces are actually in contact with one another,
because that that space between them is so small that
even visible light is too big to show it. The
distance between them is about the distance of the width
of a couple of bands of d N and A.
It's tiny, so a single mote of dust would be
like a gargantuan boulder by comparison, and it's really hard
(32:55):
for me to get my mind wrapped around it, because
once we start talking about this level of scale, I
can kind of conceptualize it, but I can't visualize it now.
That's one of the reasons that these hard disks are
sealed in aluminum cases. It's to protect the platters from
dust and other contaminants. It's also why you should never
(33:16):
open up a hard disk drive unless you're okay with
the fact that it's never going to serve a useful
purpose outside of perhaps being an instruction to others on
how disk drives work or or showing people what it
looks like, because the chances are it's never going to
run again. This is also why stuff like clean rooms
need to exist. Clean rooms are facilities that use powerful
(33:38):
filtration and h VAC systems, along with incredibly strict protocols
to prevent the introduction of dust particles. Stuff like hard
drives and semiconductor chips need to be produced in clean
room facilities to avoid the possibility of even just one
moat of dust getting in there and ruining everything. Hard
disk drives tend to be fairly heavy and they require
(33:59):
a decent amount of power to operate, but they also
are cheap and they tend to be pretty high capacity.
Meaning we've advanced the science of designing hard drives to
a point where you can store an enormous amount of
information on a physical hard drive. But now it's time
for us to turn our attention to the alternative long
term storage solution, that of the solid state drive or
(34:20):
s s D. And when we come back, I'll tell
you all about it. But first let's take another quick
break with solid state drives. Were no longer talking about
mechanical systems. There are no moving parts. We're also no
(34:40):
longer talking about magnetic media, so we're not saving data
by magnetizing small areas on a chip or anything like that.
It's a form of nonvolatile memory, so the data does
stick around even if the computer or device is powered off.
The secret sauce in this case is that an s
s D follows into the storage medium of semiconductor chips.
(35:05):
Ss D chips share some similarities with other chips that
are on your computer. For example, remember when I was
talking about rom and Ram at the beginning of the
episode well. A ROM chip is a microchip that is
physically programmed to carry out specific sets of instructions, including
those necessary to boot up a computer. RAM chips are
microchips that can temporarily hold information for quick reference by
(35:28):
the CPU. The RAM and RAM chips are mounted on
the motherboard that's the main circuit board for a computer,
and SSD is not mounted on the motherboard like a
physical hard drive. It lives separate from the motherboard. It
connects to the motherboard via cables. In fact, if you
had a PC with a hard disk drive, you could
(35:50):
open up your computer, you could disconnect your hard disk
drive and you could install a solid state drive in
its place without really changing anything else in the computer
or The type of storage ss d s provide has
a somewhat confusing name. It's called flash memory. I say
it's confusing because we talked about random access memory or RAM.
(36:13):
But that stuff is volatile, right, It goes away when
you power the device down, that information is gone. But
flash memory and ss d s does not go away.
The data stays put. So we've got two different kinds
of memory here, one of which is actually storage. But
don't blame me because I don't come up with the names.
(36:33):
I'm just reporting them. Flash memory can come in a
couple of different varieties, and the type we find an
SSD drives is nan flash in A and D, but
there's also nor flash. So what the heck is up
with those names and how are these two things different? Well,
let's start with the names because they're based on a
very foundational component of computer science. It harkens back to
(36:57):
logic gates, which depend upon Boolean functions or Boolean algebra.
This is all about binary variables, so it's a variable
that can represent one of two values, like a zero
or a one. Logic gates are the practical implementation of
Boolean algebra. The gates determine what output is sent out
(37:20):
based on the incoming input. And let's use a simple
example with the and gate. The and gate accepts two
variables two variables as input, and the values for those
inputs can be either a zero or a one. The
and gate will output a one only if both inputs
(37:41):
are also ones. So if you feed two one bits
into the input of the hand gate, you get a
one bit as the output. But if you were to
feed in two zeros or a zero and a one,
or alternatively a one and a zero. We do differentiate
between these, then the end gate would produce a zero
(38:04):
as its output. So it's a set of rules. It says,
if I get two ones, I give you a one.
If I get anything else, I give you a zero.
By contrast, and or gate produces a one, also known
as a high output if at least one input is
also a one. So the or gate will send out
a one if the inputs are one in zero, zero
(38:26):
and one or one in one, and it only produces
a zero if the inputs are both zero. So this case,
you give me two zeros, I give you a zero.
You give me any other combination, I give you a one.
Logic gates are a way to build out complex responses
to inputs. They are the instructions that tell the computer
(38:47):
what outcome to produce given a specific input. So let's
talk specifically about nand and nor. These are sort of
the bizarro versions of the and A or gates. A
nand gate produces a one output with every pair of
inputs except for one one. So in other words, if
(39:08):
you give me a zero, zero, a zero one, or
a one zero, I will give you a one output.
If you were to give me one one, I would
give you a zero output. A nore gate will only
produce a one output if both inputs are zero. So
you give me a zero one, a one zero, or
a one one. I give you a big fat zero.
(39:29):
You give me a zero zero. Hey, it's your lucky day.
I give you a one. Now. I've done an episode
on logic gates many years ago to explain how why
these are important, how they work in the realm of
computer science, and what this actually all means, and it
may be time for me to revisit that and to
kind of give it a deeper treatment, because it really
gives you an appreciation of the logical design that you
(39:52):
have to create so that computers will do the stuff
you want them to do. But for now, let's just
put that aside and go back to nand versus NOR
flash memory. With both nand and NOR flash memory cards,
you have transistors arranged in cells. They're laid out in
a grid format, so you've got rows and you've got
(40:14):
columns of transistors. In NOR flash cells, the grids are
wired in parallel to one another, so you can think
of them as being wired side by side. In nand
flash cells. They're wired in series, which means you go
from one to the next one and you wire them
all in order in a sequence. NAND cells have a
(40:36):
greater density of transistors and they also use fewer wires
than NOR cells. They can read and write data faster
than NOR flash memory, so nand flash is great for
the solid state drive, whereas NOR flash tends to be
used for read only purposes, kind of like the wrong
chips on a motherboard. If you were to put an
SSD and a hard disk drive mint for the exact
(40:59):
same drive bay in a computer case next to each other,
like if you were to take out a hard disk
drive at an SSD drive and you put them side
by side, they would look fairly similar there both be
and metal you know, aluminum cases, and they would be
the same size. But the SSD would have no mechanical
parts and it would likely have a good amount of
(41:19):
unused space inside the case. The reason for that is
for a convenience, the solid state drive needs to match
the physical shape and size of the hard disk drive
so that it can fit into the computer case properly,
so it's really just there so it'll it'll be able
to fit the model of the computer case, it's not
(41:41):
necessary for the ss D to actually function. The nand
semiconductor chips in an ss D have transistors arranged in
a grid, which means that the grid has columns and rows,
and a chain of transistors conducting a current would represent
the value of one. A chain that is not conducting
current represents a zero. And at first, before you've stored
(42:04):
any data on a solid state drive, you haven't. You
haven't saved anything to it. All the transistors would be
carrying currents, so they would all be set to one.
Saving data to the drive means that the solid state
drive will actually start to block current to specific transistors
to switch them from a one to a zero. Now,
(42:25):
at each intersection of this column and row, you get
a pair of transistors that form a cell. One of
the two transistors is a control gate and the other
is a floating gate. To quote the house Stuff Works
article on it quote, when current reaches the control gate,
electrons flow onto the floating gate, creating a net positive
(42:47):
charge that interrupts current flow. By applying precise voltages to
the transistors, a unique pattern of ones and zeros emerges.
End quote that article. By the way, was written by
William Harris, not written by me. It's a great article.
I highly recommend reading How solid state Drives Work if
you want to learn more. One big advantage of solid
(43:10):
state drives over hard drives is that with no moving parts,
the computer can access data from any part of the
solid state drive with the same speed as any other part.
There's no arm that needs to move into position, there's
no platter that needs to spend, and this means that
data can move from storage to RAM or into cash
memory much faster than it would with a hard disk drive,
(43:33):
and it's fast enough to make a noticeable difference. And
they also use less power than hard disk drives do,
so that's another bonus. Now, a few years ago, there
were some pretty big differences in storage capacity between hard
drives and solid state drives. For a while, the hard
drive had a really good head start, and so for
(43:53):
a few years if you wanted a lot of storage,
really the hard drive was the way to go. You
can get much higher copa a city hard drives. But
today solid state drives are really caught up and it's
possible to buy a solid state drive with the same
storage capacity as a high capacity hard disk. Drive. However,
solid state drives are much more expensive now. The cost
(44:15):
fluctuates based on numerous market factors, but you're likely to
spend double or more than what it would cost to
get a hard disk drive that has the exact same
storage capacity, so they are much more expensive. Interestingly, solid
state drives actually do wear out over time, despite the
fact that they don't have moving parts. So the application
(44:37):
of voltages on transistors, you know, changing the charge of
those transistors that slowly wears out the transistors, and after
a number of cycles a cycle being going from say
a one to a zero back to a one. After
a certain number of those, the cells will start to
wear out. Now, the typical number of cycles ranges in
(44:59):
the thoul posens of cycles, and computers are really good
at using up available storage space that hasn't been through
a lot of cycles already, so typically you don't have
to worry about the solid state drive giving out before
some other component on your machine gives out. So in
other words, you're far more likely to need to upgrade
your computer due to your processor or something else other
(45:21):
than the solid state drive. It would be unusual for
you to use a solid state drive so long that
that cycle thing becomes a real issue. One interesting thing
to remember is that there's always going to be bottlenecks
for data transfer. You might speed up in one area,
but you will start to find restrictions in other areas.
(45:41):
The limitation might be in the amount of RAM you
have in your machine. The RAMS capacity is going to
limit how much can be loaded into memory, which is
why a lot of folks advocate for adding more RAM
to a machine if you want to make it go faster.
Of course, this only works if the machine actually has
the capability to accept more RAM. You might have a
device where you can't upgrade the RAM, or you might
(46:04):
have a device where you've got as much RAM in
it as the motherboard can support. But then there's also
the bus, and a bus in a computer is a
connection between different components within the computer itself. Can actually
also be external components that are attached to the computer.
Bus is a very generic term, but you can think
of the bus as a highway between two different components,
(46:28):
and data travels down this highway to get from one
to the other. So like the memory to the CPU,
and a lot of devices placed the memory physically close
to the CPU to improve data transfer efficiency. It's kind
of blows my mind to think about that, that the
difference between you know, a centimeter can make a big
difference in and transfer efficiency, and it's it's kind of crazy,
(46:53):
but like a highway, a bus has a capacity limit
to how much data can actually cross it in a
of an instant, a bus will have a limit on
how many bits per second can move across it. So
if you're trying to build a fast PC, you've got
to take a lot of different things into consideration, including
the processors, which might include not just a central processing unit,
(47:16):
but maybe one or more graphics processing units, the RAM
that supports those processors, the storage system you're going to
be using, and more so making your computer go faster.
There are a lot of different approaches you can take.
Adding more rams usually a pretty good one, but switching
to a different kind of storage if you're using a
(47:38):
hard disk drive, if you switch to a solid state drive,
that can really help out a lot. It's also generally
more reliable than a hard disk drive. Fewer failures happen
with them. In general, they're always exceptions, and it's always, always,
always a good idea to back up your data. Back
it up, on an external drive or back it up
(48:00):
on a cloud service. Back it up somewhere just in
case one of those catastrophic failures does happen, you'll still
be able to get to the important information. So I
hope that all of this was useful. It's really interesting stuff.
Like I said, we can do a full episode on
things like logic Gates further down the road. Logic is
one of those things I really enjoy because it's all
(48:21):
about just learning basic rules, and those rules are solid,
Like the only thing that changes is what you're feeding
into those rules. But the rules themselves are dependable. And
in the world I live in now, when I find
something that's dependable, I hug it, I hug logic Gates. Guys. Okay,
(48:43):
well that was weird. If you have any suggestions for
future episodes of tech Stuff, whether it's a technology, a
trend in tech, a person in tech, a company, anything
like that, reach out to me. You can get in
touch over on Twitter. The handle for the show is
tech Stuff HSB you and I'll talk to you again
really soon. Y. Text Stuff is an I Heart Radio production.
(49:11):
For more podcasts from I Heart Radio, visit the I
Heart Radio app Apple podcasts, or wherever you listen to
your favorite shows.
TechStuff News
Hosts And Creators
Oz Woloshyn
Karah Preiss
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.