April 26, 2018 • 30 mins
What are some of the modern uses for cloud computing? How safe is it? And who owns all that data anyway?
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):
Get in touch with technology with tech Stuff from how
stuff Works dot com. Hey there, this is Jonathan Strickland
and you are listening to tech Stuff. I am an
executive producer with How Stuff Works and I love all
things tech, and today we're going to rejoin the overview
of cloud computing. What exactly is cloud computing and how
(00:28):
does it work. We talked a little bit about the
history of cloud computing in the last episode and kind
of give some definitions of it. So if you haven't
listened to that, go check that out. It is a
bonus episode. This is also a bonus episode. I am
in Las Vegas, Nevada. Since that's the correct way it's pronounced,
I've learned, and I am covering Sweet World eighteen. It's
(00:49):
a big cloud computing conference held by net Sweet and
so as part of that, I thought I would do
a couple of episodes about cloud computing, and in this one,
we're going to look a little bit more into what
cloud computing can do these days. Turns out it is
a big, big business, a multibillion dollar business these days,
(01:09):
and I'll talk a little bit more about that and
uh enjoy So the ability to connect computers together also
gave rise to computing models similar to cloud computing. For example,
there's cluster computing. That's where you couple lots of different
machines together to work as a unified system. Typically, I
mean the coupling can be loose or tight. It doesn't
(01:32):
really it doesn't necessitate tightly coupling machines together, but frequently
that is the way it's done. And often these machines
are all in the same location, so they may all
be in the same data center. Clusters just tend to
be more tightly connected than other computer models. So there's
also grid computing. Grid computing is less tightly coupled than
(01:54):
cluster computing. You may have computers that are part of
this grid in remote locations there and nowhere close to
each other, and they're all working together to solve various problems.
But typically they're working kind of the way a multi
core processor works. And by that I mean that the
various computers in the grid are working on different parts
(02:16):
of a problem. You could have the problem divide it
up into various segments, and the various grid computers are
working on each of them are working on a segment
of that problem. This you can see examples of this,
and things like the various at home projects like folding
at Home or set at Home that use user computers
(02:38):
as uh as units in a grid computing system to
solve very difficult problems that would normally take a supercomputer
ages to complete if it were working on it just
by itself. So both cluster computers and grid computers, both
of those models are sort of like virtual supercomputers because
(03:00):
they are grouping the assets of various regular powered machines
together to make it more than what it was, so
you can create kind of a virtual supercomputer now. In
Professor Ramna Chilapa at Emory University here in Atlanta described
cloud computing as a method that would be defined not
(03:21):
by technological limits, but rather by economic factors. So, in
other words, the professor was saying that the costs associated
with scaling processes would make it imperative for businesses to
offload that burden by making use of cloud computing providers,
and that this would in turn create the opportunity for
such providers to kind of coalesce and become the partner
(03:44):
these companies needed to grow and do work. In other words,
he was saying the economic environment out there is such
that there is a need for services that can provide
these sort of cloud computing processes and storage. Uh So,
because there's a need, sooner or later, they're going to
be businesses filling that need. It's it's like nature of
(04:07):
whorring a vacuum kind of concept. So cloud computing isn't
just about technology but also the bottom line, and that
should not come as a surprise since many factors that
we frequently associate with technological advances are actually tied closely
to economic factors. For example, Moore's law, which we frequently
simplify by saying processing power doubles every two years or so,
(04:31):
was originally an observation about how economic factors would drive
the necessity to develop a means to double the number
of transistors on a square inch of silicon substrate. In
other words, money makes the world go round. In Evan
Goldberg founded a company called net Ledger. The company was
offering up software as a service, which means it's a
(04:52):
good time to tackle a few common buzzwords associated with
cloud computing. I'll get back to Mr Goldberg in just
a minute. Now. For one thing, you'll find a lot
of as a service buzz terms associated with cloud computing.
Software as a service is a big one that is
also known as S A A S, with both s
as capitalized in the lower A the as in lower
(05:13):
case UM. And this model, you don't sell software packages
to a customer. Instead, you set up essentially a subscription
service so that the customer can access the software. You
maintain ownership of the software itself. You as the provider.
You own the software. You're not selling instances of it
to people. You provide access to the software in return
(05:35):
for money. Then there's platform as a service that's P
A A S. Again the P and the S or
upper case the a's or lower case. That's the case
for all of these as a service buzzwords. With platform
as a service, you create a virtual platform for the
customer for the purposes of developing programs or apps. So
(05:58):
developers use the virtual platform them when creating whatever app
they're working on at the time, and it can serve
as a place where developers share tools and processes to
help speed up development. There's a variant of this called
mobile back end as a service that, as the name suggests,
performs the task of being the back end operations of
a mobile application. But this is all about creating the
(06:21):
tools the developer needs in order to create and test
and then deploy the software they are making. Next, you've
got the infrastructure as a service or i a a S.
This is even more robust than the platform as a service,
and it's really meant for businesses for enterprises. So let's
(06:42):
imagine that you start your own business and you're making toys,
and you start out small, maybe you're even working out
of your own home, and you're doing all of this
by hand. But your business grows, your demand increases, and
you are physically incapable of meaning that demand all by yourself.
So you need to scale up operations. And this kind
of starts slowly. You know, at first it's not too hectic.
(07:06):
You can hire on some people and take on that responsibility.
But if demand continues to increase, you have to take
the next step. You have to start forming relationships with
factories to make your toys. And this is both to
increase the supply as well as to reduce the cost.
By producing them in bulk, you can reduce the cost
on a per unit basis to produce and also still
(07:28):
meet the demand that your customers have. You have to
manage supply chains. Though you have to make sure all
the components you need to make your toys are getting
to the factories. You have to figure out how to
get the manufactured toys to retailers or to customers. You
have to manage all the money that's involved in those transactions,
whether it's payments to services like the factories or accepting
(07:50):
payments from customers. And for some businesses, this becomes a
barrier to growth because it's so overwhelming. The transformation of
small company to mid size company or mid size company
to large company can be really daunting. So the lure
of infrastructure as a service is that companies that have
already designed systems, usually suites of programs that streamlined various processes,
(08:15):
will share that knowledge and capability with you for a price.
So you pay a subscription and then you get all
access to all these tools that may help you do
things like track your supply chain, to track production and
even see how things have changed over time. Like there
are some very sophisticated programs out there that allow you
to get very granular with data. Data analysis is a
(08:39):
huge part of cloud computing services because it provides another
value to the customer if you can say, not only
will we show you what the UH you know where
you are at any given time, will show you trends
and will even start to predict risks that might be
in place, like maybe we see that because of this
(09:00):
one element in the supply chain, you're going to have
a bottleneck in the manufacturing process, which might mean you
need to be able to communicate out to customers that
there's currently a shortage of whatever the product is and
that you will be turning this around as soon as
you can. Or it might even be something about how
producing UH your product in one region might be more
(09:22):
economically viable than another region, even within the same country,
so that you can get your products to market with
a lower cost, not just economic, but environmental. It could
be something as as UH intrinsic as how many miles
have to be traveled in order to get your product
to market. So it gets really really complicated, and you
(09:44):
can see how if you're running your own business, once
you start layering on these different considerations, it rapidly gets
outside of most people's comfort zone, so that that's kind
of the selling point for infrastructure as a service. There's
a more recent off ring that's called function as a
service or f a a S. This creates a level
(10:05):
of abstraction between the developer and virtual machines, so that
developers only have to worry about creating very narrowly functional
blocks of code, and essentially they're just creating the instructions
that sit on top of this service, and the service
does everything else. Like it removes the necessity to program
(10:25):
for a specific stack of technology, so the developer just
has to worry about the code that he or she
needs to do whatever it is they want to do.
And typically this kind of service executes upon a real
world event happening. So something happens in the real world
that triggers the function, and the function then gets executed
(10:49):
by this service. So an example it would be, let's
say you go to a website and there's a form
you want to fill out in order to get access
to something. So you fill out the form, you can
complete it, and you submit it. That submission could be
the real world event that then triggers this function, this
hypothetical function I'm making up now. Unlike infrastructure as a
(11:11):
service that's a perpetual model that means you have to
keep paying a subscription for as long as you are
relying upon that infrastructure. It is a day today thing.
Function as a service typically only triggers a fee whenever
the function itself gets triggered, so you're not paying every
single day for this thing to exist. You're paying only
(11:32):
when people are using it, which means that you can
save money if it's not something that people are going
crazy and using all the time. If it is, then
maybe you have to look at a different model for
economic purposes. But in general, it means that you can
save a lot of money this way because it's very lightweight.
It just exists on the top of the provider's services.
(11:53):
Now there are other service models out there. I'm getting
really tired of saying the word service, but it's that's
just the term used. So there's storage as a service,
that's what sounds like. That's your online document databases or
photo albums or file storage or whatever. There's also video
as a service. There's compute as a service. There's probably
half a dozen more, but it all comes down to
(12:15):
the concept of someone else running a process on lots
of computers. You pay that someone a fee to make
use of those assets, and it frees you up from
having to build out your own massive computer system. Hey guys,
before we go any further, I need to take a
quick break to thank our sponsor. Let's get back to
(12:40):
net Ledger and Evan Goldberg. The company ran accounting software
that had a web enabled user interface, so customers could
pay to use the service over the Internet rather than
purchase a full software package and installed on their own machines.
And one of the selling points of this approaches there's
never a need to upgrade your software. And by that
I mean if you buy a software package, like a
(13:02):
conventional traditional software package, chances are sooner or later there's
going to be an updated version coming along, which creates
a lot of pressure on customers. So do you upgrade
to the latest version? If so, you probably have to
spend more money to do it. If you don't upgrade
to the latest version, you might find that some of
your data you work with becomes incompatible with your older
(13:24):
legacy systems because it's meant for the newer versions of
the software. So there could be features that the newer
version has that your old version does not have and
it doesn't support, and so you start running into compatibility issues.
But with software as a service, the provider makes all
the upgrades to the service on their end on the
(13:45):
on the back end, So you're accessing the software through
the Internet, whether it's through the web or through an app,
whatever it may be. You don't have to worry about
installing upgrades or patches necessarily. I mean sometimes you do,
but that's what the app world. If you're talking about
web browsers, you typically don't. So all you're doing is
just accessing the service. All of the upgrades are being
(14:09):
done behind the scenes on the provider side. So one
of the benefits of this model is that you always
have access to the latest version of the service as
long as you remain a customer. Net Ledgers features grew
the company evolved into net Suite, which better indicated that
the company was offering up multiple internet based services. A
short time after Goldberg found a net Suite, there was
(14:31):
another cloud based services company that popped up called Salesforce
dot Com. The founders of salesforce dot Com included one
former Oracle executive named Mark bennie Off and three software
developers named Frank Dominguez, Parker Harris, and Dave Mollenhoff, who
had come from Left Coast Software. Their first product was
(14:53):
a type of sales automation software, and Salesforce rapidly gained
attention as it was offering up enterprise level services over
a relatively simple web based interface. The company became incredibly
successful and was able to weather the dot com crash
and grow into a multibillion dollar global company. Net Suite
also made it through the crash and became a much
(15:15):
larger organization over time, and in two thousand and sixteen,
the company was acquired by Oracle. I'll talk more about
both net Suite and Oracle in upcoming episodes. But in
two thousand two, Amazon made a move to maximize company efficiency.
The standard practice at the time was to have enough capacity,
like computer capacity, to do your work while only using
(15:36):
ten of your capacity. But that's not very efficient. It
means that nine cent of your computer capacity goes unused,
and so Amazon began to explore other options that would
allow them to leverage that. The answer turned out to
be cloud storage and cloud computing, and in two thousand
six they expanded this by introducing Amazon Web Services, which
(15:59):
had a whole bunch of different web based services that
companies can use, including mechanical turk so quick side note
about Mechanical Turk because I just love this story. So
Amazon markets it as a tool that allows customers to
leverage human intelligence for specific tasks because humans tend to
be better at certain things than computers. And by better,
I mean we can do that particular type of work
(16:22):
much more quickly and reliably. But it can be expensive
to hire humans to do those tasks, especially if the
tasks are very simple, and if you only need people
for a short amount of time, then hiring a huge workforce,
a temporary workforce, that's that's an an enormous expense, and
not just money, but also in time. And so Amazon
(16:44):
Mechanical Turk is an on demand workforce service. It's kind
of like a group of people employed by Amazon to
do whatever it is you need them to do with
your human task issues. And then then once that's, once
that projects done, they go on to do something else.
And it sounds a little creepy if I'm being totally honest,
(17:05):
But the reason why I wanted to talk about them
was because of what mechanical Turk is a reference to.
It's actually referencing an old piece of clockwork chicanery. The
original device was called simply the turk, and it appeared
to be a clockwork automaton that had the figure of
a turk sitting at a chessboard. It was made by
(17:27):
a guy named wolf Gang Fawn Kimberland. He built it
in the late seventeen hundreds, like in the seventeen seventies,
and upon casual glance you would think it was a
robot of ingenious design, capable of defeating even skilled players
in chess. But in reality, there was a human being
hidden within the cabinet of the machine who was guiding
(17:49):
the turk's movements, and a playing chess against a human opponents.
So really it was just two humans playing chess, only
one of the humans was hidden out of sight. Just
I love that story, so I wanted to tell it quickly.
Around the same time the Amazon was introducing web services,
Google debuted Google Docs, which was actually based off two
earlier products. One of them was Google Spreadsheets, which Google acquired.
(18:12):
They bought it from a company called two Web Technologies,
and they also acquired a company called Rightly that ended
up being the basis for the Google Docs part, the
actual document offering of the product. How those moves began
to bring cloud computing into the world of the home
user for the first time. Earlier implementations were almost completely
(18:34):
focused on business to business operations, which means most of
us never see it right when we talk about business
to business stuff, it's interesting because it tells us how
the stuff we interact with, how that happens, how it
gets done in the background, but we don't see it directly.
That's why when we talk about companies like IBM, most
(18:55):
of us have very little experience working directly with IBM
products unless it's within the realm of the office, because
IBM doesn't really make products for the home consumer. The
same sort of thing here was that cloud computing for
the longest time was not for the home consumer. It
was for businesses. But with the emergence of these kind
(19:19):
of applications, we were starting to see people get access
to that sort of stuff for the first time in
a big way. And that's when it became necessary to
figure out how to explain that technology to people. And
that's when cloud computing really became a buzzword, and also
how I came to write sixteen articles about the stuff.
(19:39):
These days, many companies are involved with cloud services, along
with net Suite and Salesforce there's Amazon, Google, Microsoft has
its Azure platform, IBM is a big one, and there's
tons of others. And the services being offered by these
companies have grown significantly over time as well. For example,
(19:59):
i EM offers you the chance to work on a
quantum computer over the cloud, which still blows my mind.
We got just a bit more to talk about with
cloud computing, but before I get into it, let's take
another quick break to thank our sponsor. One thing I
(20:20):
feel we should talk about is the different types of
cloud computing models. And I almost said different types of clouds,
but I need to be more serious. That's a cloud joke.
One type is the public cloud model. Now, these are
services in which all the assets are run by a
third party. So I talked about Google Docs. That's a
great example of a public cloud computer model. So if
(20:41):
your company relies on Google Suite for all of its processes, really,
then you'll be relying on a public cloud. Then you've
got private clouds and those models. The company itself, whatever
company it is that's running the processes, it actually owns
all the equipment and the services that run on that equipment.
(21:03):
So let's say you work for a company ABC, The
private cloud is also owned by ABC, and it's got
all these different data centers and they're running all these processes.
So you're still accessing apps and programs and storage that
exists on other computers. It's just that you work for
(21:24):
the same company that owns those computers. It's not owned
by a third party. Now, you would want to use
the private cloud approach if you had mission critical applications
that you wanted under your full control and you didn't
want to entrust that to a third party. So if
you're handling really sensitive information or the processes you rely upon,
or really big trade secrets, that might be the way
(21:44):
you want to go. Or if your needs are just
so particular that there's not really a provider out there
that can meet your needs because they're so different from
what everyone else needs, this would be the way you
would go. There's a third model called the hybrid cloud
that merges those two. You have both public and private
cloud entities. So a company with a hybrid cloud approach
(22:07):
would have a public cloud stuff to do to handle
certain tasks, private cloud to handle other tasks, and as
you might imagine, the more critical elements to the company's
operations would probably run on the private cloud, while more
mundane processes might be pushed to the public cloud service,
and there's some sort of layer of communication and automation
(22:28):
that allows for the exchange of data between those two
clouds in a controlled manner so that the output from
one can work with the information on the other. In fact,
without that communication channel between the two clouds, you do
not actually have a hybrid cloud. You would instead have
what is called a multi cloud approach. You would have
two distinct cloud services that do not communicate with each other.
(22:52):
And there are valid reasons to go with a multi
cloud approach where there is no communication between clouds. So again,
let's talk about a really big company. Really big companies
could have departments. They're so large and have such specific
needs they do not need to communicate directly with other
departments to get their their work done. Their their work
(23:14):
may not relate directly with other departments, so in those
cases it might make more sense to have a separate
public or private cloud available to those departments either, so
you can have multi cloud approach that uses either private
or public or both. The important element here is that
there isn't that communication channel between them. So as companies
(23:35):
grow they often find that the departments inside them end
up being larger than the original company was when it
first got started, and at that size, finding processes that
work and scale is critical. Now, at the top of
the show, I mentioned that one of the big concerns
about cloud computing was security, and with so many companies
in trusting data and processes to third parties, security is
(23:57):
an absolute necessity without a company would go out of business.
Third parties would collapse if they were found to be insecure,
and like I said, that could create a domino effect,
a disastrous result among the third party's customer base. So
typically cloud computing applications rely upon profiles that require authentication
through some means, most commonly through a user name and password.
(24:20):
There are cloud based services that actually managed this as well.
It's called identity as a Service or i d a
a s NOW. Those services and not only managed user
log in information, but can designate different levels of access.
So for example, the head of a department might need
administrator level access to that department's data, while an employee
(24:41):
further down the organizational chain might only require limited access.
So you need to have a way of denoting that
so that you don't have some low level employee suddenly
have access to say, everyone's pace stub. That would be bad.
Cloud providers take security seriously and it shows because they
(25:02):
might prove to be attempting target to hackers. But the
big providers of technically been more resilient to attacks than
private enterprise centers have been in the past. The weak
spot tends not to be the data centers or the
providers or the security around them, but rather that tenuous
link between provider and customer. It's that old problem where
(25:25):
you find out that the weakest link in the security
system is the people. It's always the people, because people
might choose really bad passwords, which creates a security vulnerability
that's difficult to protect against. You might use something like
two factor authentication to help fight against that problem. That
can help. That's where a person needs not only a
(25:45):
password but some other form of authentication like a physical
token in order to access the services. And at that
stage and an authorized person would need to get hold
not just to the password, but also of some sort
of physical object a token or a cell phone or
something that the target owns. So it doesn't eliminate risk,
(26:05):
but it cuts down on it significantly. There are a
lot of other buzzword like topics that tie in with
cloud computing. There's big data that's all about how to
leverage the huge amounts of information that's being mined on
a daily basis. How do you do that in a
way that's meaningful and efficient. That's frequently associated with cloud computing.
Artificial intelligence and machine learning also get thrown into the mix.
(26:28):
There are services that use machine learning to sort through
big data on the cloud, for example. That way you
get a big, heaping handful of buzzwords all at once,
that's all in an effort to produce a particular result.
Sometimes you use machine learning and big data for legitimate
research purposes that can further our scientific understanding. Sometimes it's
(26:49):
for less lofty reasons. It might be to find the
best way to advertise to different groups of people in
order to more effectively sell products to them. Sometimes you
can get downright creepy, which as the recent Cambridge Analytical story.
Have to do a full episode about what that was
all about. Why raised such a fuss at some point,
But that's that's for another time. It's a complicated issue
(27:11):
that I can't just if I ran into it here,
we would easily go another hour, and UH actually have
to be at the conference keynote in a little bit,
so I can't do that now. At the top of
the show, I also mentioned that there can be an
issue about ownership, like who owns the data on these services,
and usually there's some pretty complicated terms of service that
lays all this out, where you're essentially granting a license
(27:35):
to the third party to have a sort of physical
possession of all that information. And the reason that's necessary
is because the way cloud computing works, the way you
can access it through different devices and in different places,
you have to give the company permission to be able
to show that information. So if I store a file
(27:59):
on my personal computer and I'm the only one who
has physical access to my personal computer, I'm reasonably assured
that I'm the only person who can see it. There's
no real permissions that need to be made or anything
like that. If I'm storing my data on someone else's
machine and I want it to be clear that still
my data even though it's sitting on someone else's machine,
(28:22):
then we have to draw up these agreements. And one
of those agreements might be well, if you request to
see and work with this data, I need to have
the permission to actually serve it to you. If you
choose to share it with someone, I have to have
the permission that allows me to share it so that
I'm not legally responsible If you later on say, oh
(28:44):
I didn't I don't want that person to have it. Well,
if you chose to share it, by the nature of
the cloud computing model, then you know that that's what happened,
that it got shared. So it's largely to protect the
third party providers, but it's also just this very complicated
language that gives them the permission they need to do
the business as they do it. So it can get
(29:07):
pretty complicated. On casual glance, it looks like you're signing
over all your data to another party, um, and you're
not really doing that, at least not in most agreements.
Any ethical agreement would not include that sort of information
in it. But that's pretty much the overview of cloud computing.
Will conclude our story on that. It's a fascinating model
(29:29):
that has its roots all the way back in the
early days of of of mainframe computers. I hope you
found this episode helpful and understanding what the heck that
buzzword meant in the first place. If you have suggestions
for future episodes of tech Stuff, whether it is to
explain something within tech, to talk about a specific technology
or a company or a person in technology. Maybe you
(29:51):
have someone in mind that I should interview or have
on as a guest. Reach out and let me know.
You can send me and email the addresses tech Stuff
at how stuff works dot com or draw me a
line on Facebook or Twitter. The handle for both of
those is text stuff hs W. Follow us on Instagram.
We've got an Instagram account that we post to regularly,
(30:12):
so check that out. And remember, on normal weeks, I
record this show on Wednesdays and Fridays. You can actually
watch me record it live in the studio. You can
just go to the link Twitch dot tv slash tech Stuff.
You'll see the schedule there, and there's a chat room
and everything. You can join in be part of the crowd.
And I look forward to seeing you and I'll talk
(30:32):
to you again really soon. For more on this and
bouthands of other topics, is it how staff works dot
com
Get in touch with technology with tech Stuff from how
stuff Works dot com. Hey there, this is Jonathan Strickland
and you are listening to tech Stuff. I am an
executive producer with How Stuff Works and I love all
things tech, and today we're going to rejoin the overview
of cloud computing. What exactly is cloud computing and how
(00:28):
does it work. We talked a little bit about the
history of cloud computing in the last episode and kind
of give some definitions of it. So if you haven't
listened to that, go check that out. It is a
bonus episode. This is also a bonus episode. I am
in Las Vegas, Nevada. Since that's the correct way it's pronounced,
I've learned, and I am covering Sweet World eighteen. It's
(00:49):
a big cloud computing conference held by net Sweet and
so as part of that, I thought I would do
a couple of episodes about cloud computing, and in this one,
we're going to look a little bit more into what
cloud computing can do these days. Turns out it is
a big, big business, a multibillion dollar business these days,
(01:09):
and I'll talk a little bit more about that and
uh enjoy So the ability to connect computers together also
gave rise to computing models similar to cloud computing. For example,
there's cluster computing. That's where you couple lots of different
machines together to work as a unified system. Typically, I
mean the coupling can be loose or tight. It doesn't
(01:32):
really it doesn't necessitate tightly coupling machines together, but frequently
that is the way it's done. And often these machines
are all in the same location, so they may all
be in the same data center. Clusters just tend to
be more tightly connected than other computer models. So there's
also grid computing. Grid computing is less tightly coupled than
(01:54):
cluster computing. You may have computers that are part of
this grid in remote locations there and nowhere close to
each other, and they're all working together to solve various problems.
But typically they're working kind of the way a multi
core processor works. And by that I mean that the
various computers in the grid are working on different parts
(02:16):
of a problem. You could have the problem divide it
up into various segments, and the various grid computers are
working on each of them are working on a segment
of that problem. This you can see examples of this,
and things like the various at home projects like folding
at Home or set at Home that use user computers
(02:38):
as uh as units in a grid computing system to
solve very difficult problems that would normally take a supercomputer
ages to complete if it were working on it just
by itself. So both cluster computers and grid computers, both
of those models are sort of like virtual supercomputers because
(03:00):
they are grouping the assets of various regular powered machines
together to make it more than what it was, so
you can create kind of a virtual supercomputer now. In
Professor Ramna Chilapa at Emory University here in Atlanta described
cloud computing as a method that would be defined not
(03:21):
by technological limits, but rather by economic factors. So, in
other words, the professor was saying that the costs associated
with scaling processes would make it imperative for businesses to
offload that burden by making use of cloud computing providers,
and that this would in turn create the opportunity for
such providers to kind of coalesce and become the partner
(03:44):
these companies needed to grow and do work. In other words,
he was saying the economic environment out there is such
that there is a need for services that can provide
these sort of cloud computing processes and storage. Uh So,
because there's a need, sooner or later, they're going to
be businesses filling that need. It's it's like nature of
(04:07):
whorring a vacuum kind of concept. So cloud computing isn't
just about technology but also the bottom line, and that
should not come as a surprise since many factors that
we frequently associate with technological advances are actually tied closely
to economic factors. For example, Moore's law, which we frequently
simplify by saying processing power doubles every two years or so,
(04:31):
was originally an observation about how economic factors would drive
the necessity to develop a means to double the number
of transistors on a square inch of silicon substrate. In
other words, money makes the world go round. In Evan
Goldberg founded a company called net Ledger. The company was
offering up software as a service, which means it's a
(04:52):
good time to tackle a few common buzzwords associated with
cloud computing. I'll get back to Mr Goldberg in just
a minute. Now. For one thing, you'll find a lot
of as a service buzz terms associated with cloud computing.
Software as a service is a big one that is
also known as S A A S, with both s
as capitalized in the lower A the as in lower
(05:13):
case UM. And this model, you don't sell software packages
to a customer. Instead, you set up essentially a subscription
service so that the customer can access the software. You
maintain ownership of the software itself. You as the provider.
You own the software. You're not selling instances of it
to people. You provide access to the software in return
(05:35):
for money. Then there's platform as a service that's P
A A S. Again the P and the S or
upper case the a's or lower case. That's the case
for all of these as a service buzzwords. With platform
as a service, you create a virtual platform for the
customer for the purposes of developing programs or apps. So
(05:58):
developers use the virtual platform them when creating whatever app
they're working on at the time, and it can serve
as a place where developers share tools and processes to
help speed up development. There's a variant of this called
mobile back end as a service that, as the name suggests,
performs the task of being the back end operations of
a mobile application. But this is all about creating the
(06:21):
tools the developer needs in order to create and test
and then deploy the software they are making. Next, you've
got the infrastructure as a service or i a a S.
This is even more robust than the platform as a service,
and it's really meant for businesses for enterprises. So let's
(06:42):
imagine that you start your own business and you're making toys,
and you start out small, maybe you're even working out
of your own home, and you're doing all of this
by hand. But your business grows, your demand increases, and
you are physically incapable of meaning that demand all by yourself.
So you need to scale up operations. And this kind
of starts slowly. You know, at first it's not too hectic.
(07:06):
You can hire on some people and take on that responsibility.
But if demand continues to increase, you have to take
the next step. You have to start forming relationships with
factories to make your toys. And this is both to
increase the supply as well as to reduce the cost.
By producing them in bulk, you can reduce the cost
on a per unit basis to produce and also still
(07:28):
meet the demand that your customers have. You have to
manage supply chains. Though you have to make sure all
the components you need to make your toys are getting
to the factories. You have to figure out how to
get the manufactured toys to retailers or to customers. You
have to manage all the money that's involved in those transactions,
whether it's payments to services like the factories or accepting
(07:50):
payments from customers. And for some businesses, this becomes a
barrier to growth because it's so overwhelming. The transformation of
small company to mid size company or mid size company
to large company can be really daunting. So the lure
of infrastructure as a service is that companies that have
already designed systems, usually suites of programs that streamlined various processes,
(08:15):
will share that knowledge and capability with you for a price.
So you pay a subscription and then you get all
access to all these tools that may help you do
things like track your supply chain, to track production and
even see how things have changed over time. Like there
are some very sophisticated programs out there that allow you
to get very granular with data. Data analysis is a
(08:39):
huge part of cloud computing services because it provides another
value to the customer if you can say, not only
will we show you what the UH you know where
you are at any given time, will show you trends
and will even start to predict risks that might be
in place, like maybe we see that because of this
(09:00):
one element in the supply chain, you're going to have
a bottleneck in the manufacturing process, which might mean you
need to be able to communicate out to customers that
there's currently a shortage of whatever the product is and
that you will be turning this around as soon as
you can. Or it might even be something about how
producing UH your product in one region might be more
(09:22):
economically viable than another region, even within the same country,
so that you can get your products to market with
a lower cost, not just economic, but environmental. It could
be something as as UH intrinsic as how many miles
have to be traveled in order to get your product
to market. So it gets really really complicated, and you
(09:44):
can see how if you're running your own business, once
you start layering on these different considerations, it rapidly gets
outside of most people's comfort zone, so that that's kind
of the selling point for infrastructure as a service. There's
a more recent off ring that's called function as a
service or f a a S. This creates a level
(10:05):
of abstraction between the developer and virtual machines, so that
developers only have to worry about creating very narrowly functional
blocks of code, and essentially they're just creating the instructions
that sit on top of this service, and the service
does everything else. Like it removes the necessity to program
(10:25):
for a specific stack of technology, so the developer just
has to worry about the code that he or she
needs to do whatever it is they want to do.
And typically this kind of service executes upon a real
world event happening. So something happens in the real world
that triggers the function, and the function then gets executed
(10:49):
by this service. So an example it would be, let's
say you go to a website and there's a form
you want to fill out in order to get access
to something. So you fill out the form, you can
complete it, and you submit it. That submission could be
the real world event that then triggers this function, this
hypothetical function I'm making up now. Unlike infrastructure as a
(11:11):
service that's a perpetual model that means you have to
keep paying a subscription for as long as you are
relying upon that infrastructure. It is a day today thing.
Function as a service typically only triggers a fee whenever
the function itself gets triggered, so you're not paying every
single day for this thing to exist. You're paying only
(11:32):
when people are using it, which means that you can
save money if it's not something that people are going
crazy and using all the time. If it is, then
maybe you have to look at a different model for
economic purposes. But in general, it means that you can
save a lot of money this way because it's very lightweight.
It just exists on the top of the provider's services.
(11:53):
Now there are other service models out there. I'm getting
really tired of saying the word service, but it's that's
just the term used. So there's storage as a service,
that's what sounds like. That's your online document databases or
photo albums or file storage or whatever. There's also video
as a service. There's compute as a service. There's probably
half a dozen more, but it all comes down to
(12:15):
the concept of someone else running a process on lots
of computers. You pay that someone a fee to make
use of those assets, and it frees you up from
having to build out your own massive computer system. Hey guys,
before we go any further, I need to take a
quick break to thank our sponsor. Let's get back to
(12:40):
net Ledger and Evan Goldberg. The company ran accounting software
that had a web enabled user interface, so customers could
pay to use the service over the Internet rather than
purchase a full software package and installed on their own machines.
And one of the selling points of this approaches there's
never a need to upgrade your software. And by that
I mean if you buy a software package, like a
(13:02):
conventional traditional software package, chances are sooner or later there's
going to be an updated version coming along, which creates
a lot of pressure on customers. So do you upgrade
to the latest version? If so, you probably have to
spend more money to do it. If you don't upgrade
to the latest version, you might find that some of
your data you work with becomes incompatible with your older
(13:24):
legacy systems because it's meant for the newer versions of
the software. So there could be features that the newer
version has that your old version does not have and
it doesn't support, and so you start running into compatibility issues.
But with software as a service, the provider makes all
the upgrades to the service on their end on the
(13:45):
on the back end, So you're accessing the software through
the Internet, whether it's through the web or through an app,
whatever it may be. You don't have to worry about
installing upgrades or patches necessarily. I mean sometimes you do,
but that's what the app world. If you're talking about
web browsers, you typically don't. So all you're doing is
just accessing the service. All of the upgrades are being
(14:09):
done behind the scenes on the provider side. So one
of the benefits of this model is that you always
have access to the latest version of the service as
long as you remain a customer. Net Ledgers features grew
the company evolved into net Suite, which better indicated that
the company was offering up multiple internet based services. A
short time after Goldberg found a net Suite, there was
(14:31):
another cloud based services company that popped up called Salesforce
dot Com. The founders of salesforce dot Com included one
former Oracle executive named Mark bennie Off and three software
developers named Frank Dominguez, Parker Harris, and Dave Mollenhoff, who
had come from Left Coast Software. Their first product was
(14:53):
a type of sales automation software, and Salesforce rapidly gained
attention as it was offering up enterprise level services over
a relatively simple web based interface. The company became incredibly
successful and was able to weather the dot com crash
and grow into a multibillion dollar global company. Net Suite
also made it through the crash and became a much
(15:15):
larger organization over time, and in two thousand and sixteen,
the company was acquired by Oracle. I'll talk more about
both net Suite and Oracle in upcoming episodes. But in
two thousand two, Amazon made a move to maximize company efficiency.
The standard practice at the time was to have enough capacity,
like computer capacity, to do your work while only using
(15:36):
ten of your capacity. But that's not very efficient. It
means that nine cent of your computer capacity goes unused,
and so Amazon began to explore other options that would
allow them to leverage that. The answer turned out to
be cloud storage and cloud computing, and in two thousand
six they expanded this by introducing Amazon Web Services, which
(15:59):
had a whole bunch of different web based services that
companies can use, including mechanical turk so quick side note
about Mechanical Turk because I just love this story. So
Amazon markets it as a tool that allows customers to
leverage human intelligence for specific tasks because humans tend to
be better at certain things than computers. And by better,
I mean we can do that particular type of work
(16:22):
much more quickly and reliably. But it can be expensive
to hire humans to do those tasks, especially if the
tasks are very simple, and if you only need people
for a short amount of time, then hiring a huge workforce,
a temporary workforce, that's that's an an enormous expense, and
not just money, but also in time. And so Amazon
(16:44):
Mechanical Turk is an on demand workforce service. It's kind
of like a group of people employed by Amazon to
do whatever it is you need them to do with
your human task issues. And then then once that's, once
that projects done, they go on to do something else.
And it sounds a little creepy if I'm being totally honest,
(17:05):
But the reason why I wanted to talk about them
was because of what mechanical Turk is a reference to.
It's actually referencing an old piece of clockwork chicanery. The
original device was called simply the turk, and it appeared
to be a clockwork automaton that had the figure of
a turk sitting at a chessboard. It was made by
(17:27):
a guy named wolf Gang Fawn Kimberland. He built it
in the late seventeen hundreds, like in the seventeen seventies,
and upon casual glance you would think it was a
robot of ingenious design, capable of defeating even skilled players
in chess. But in reality, there was a human being
hidden within the cabinet of the machine who was guiding
(17:49):
the turk's movements, and a playing chess against a human opponents.
So really it was just two humans playing chess, only
one of the humans was hidden out of sight. Just
I love that story, so I wanted to tell it quickly.
Around the same time the Amazon was introducing web services,
Google debuted Google Docs, which was actually based off two
earlier products. One of them was Google Spreadsheets, which Google acquired.
(18:12):
They bought it from a company called two Web Technologies,
and they also acquired a company called Rightly that ended
up being the basis for the Google Docs part, the
actual document offering of the product. How those moves began
to bring cloud computing into the world of the home
user for the first time. Earlier implementations were almost completely
(18:34):
focused on business to business operations, which means most of
us never see it right when we talk about business
to business stuff, it's interesting because it tells us how
the stuff we interact with, how that happens, how it
gets done in the background, but we don't see it directly.
That's why when we talk about companies like IBM, most
(18:55):
of us have very little experience working directly with IBM
products unless it's within the realm of the office, because
IBM doesn't really make products for the home consumer. The
same sort of thing here was that cloud computing for
the longest time was not for the home consumer. It
was for businesses. But with the emergence of these kind
(19:19):
of applications, we were starting to see people get access
to that sort of stuff for the first time in
a big way. And that's when it became necessary to
figure out how to explain that technology to people. And
that's when cloud computing really became a buzzword, and also
how I came to write sixteen articles about the stuff.
(19:39):
These days, many companies are involved with cloud services, along
with net Suite and Salesforce there's Amazon, Google, Microsoft has
its Azure platform, IBM is a big one, and there's
tons of others. And the services being offered by these
companies have grown significantly over time as well. For example,
(19:59):
i EM offers you the chance to work on a
quantum computer over the cloud, which still blows my mind.
We got just a bit more to talk about with
cloud computing, but before I get into it, let's take
another quick break to thank our sponsor. One thing I
(20:20):
feel we should talk about is the different types of
cloud computing models. And I almost said different types of clouds,
but I need to be more serious. That's a cloud joke.
One type is the public cloud model. Now, these are
services in which all the assets are run by a
third party. So I talked about Google Docs. That's a
great example of a public cloud computer model. So if
(20:41):
your company relies on Google Suite for all of its processes, really,
then you'll be relying on a public cloud. Then you've
got private clouds and those models. The company itself, whatever
company it is that's running the processes, it actually owns
all the equipment and the services that run on that equipment.
(21:03):
So let's say you work for a company ABC, The
private cloud is also owned by ABC, and it's got
all these different data centers and they're running all these processes.
So you're still accessing apps and programs and storage that
exists on other computers. It's just that you work for
(21:24):
the same company that owns those computers. It's not owned
by a third party. Now, you would want to use
the private cloud approach if you had mission critical applications
that you wanted under your full control and you didn't
want to entrust that to a third party. So if
you're handling really sensitive information or the processes you rely upon,
or really big trade secrets, that might be the way
(21:44):
you want to go. Or if your needs are just
so particular that there's not really a provider out there
that can meet your needs because they're so different from
what everyone else needs, this would be the way you
would go. There's a third model called the hybrid cloud
that merges those two. You have both public and private
cloud entities. So a company with a hybrid cloud approach
(22:07):
would have a public cloud stuff to do to handle
certain tasks, private cloud to handle other tasks, and as
you might imagine, the more critical elements to the company's
operations would probably run on the private cloud, while more
mundane processes might be pushed to the public cloud service,
and there's some sort of layer of communication and automation
(22:28):
that allows for the exchange of data between those two
clouds in a controlled manner so that the output from
one can work with the information on the other. In fact,
without that communication channel between the two clouds, you do
not actually have a hybrid cloud. You would instead have
what is called a multi cloud approach. You would have
two distinct cloud services that do not communicate with each other.
(22:52):
And there are valid reasons to go with a multi
cloud approach where there is no communication between clouds. So again,
let's talk about a really big company. Really big companies
could have departments. They're so large and have such specific
needs they do not need to communicate directly with other
departments to get their their work done. Their their work
(23:14):
may not relate directly with other departments, so in those
cases it might make more sense to have a separate
public or private cloud available to those departments either, so
you can have multi cloud approach that uses either private
or public or both. The important element here is that
there isn't that communication channel between them. So as companies
(23:35):
grow they often find that the departments inside them end
up being larger than the original company was when it
first got started, and at that size, finding processes that
work and scale is critical. Now, at the top of
the show, I mentioned that one of the big concerns
about cloud computing was security, and with so many companies
in trusting data and processes to third parties, security is
(23:57):
an absolute necessity without a company would go out of business.
Third parties would collapse if they were found to be insecure,
and like I said, that could create a domino effect,
a disastrous result among the third party's customer base. So
typically cloud computing applications rely upon profiles that require authentication
through some means, most commonly through a user name and password.
(24:20):
There are cloud based services that actually managed this as well.
It's called identity as a Service or i d a
a s NOW. Those services and not only managed user
log in information, but can designate different levels of access.
So for example, the head of a department might need
administrator level access to that department's data, while an employee
(24:41):
further down the organizational chain might only require limited access.
So you need to have a way of denoting that
so that you don't have some low level employee suddenly
have access to say, everyone's pace stub. That would be bad.
Cloud providers take security seriously and it shows because they
(25:02):
might prove to be attempting target to hackers. But the
big providers of technically been more resilient to attacks than
private enterprise centers have been in the past. The weak
spot tends not to be the data centers or the
providers or the security around them, but rather that tenuous
link between provider and customer. It's that old problem where
(25:25):
you find out that the weakest link in the security
system is the people. It's always the people, because people
might choose really bad passwords, which creates a security vulnerability
that's difficult to protect against. You might use something like
two factor authentication to help fight against that problem. That
can help. That's where a person needs not only a
(25:45):
password but some other form of authentication like a physical
token in order to access the services. And at that
stage and an authorized person would need to get hold
not just to the password, but also of some sort
of physical object a token or a cell phone or
something that the target owns. So it doesn't eliminate risk,
(26:05):
but it cuts down on it significantly. There are a
lot of other buzzword like topics that tie in with
cloud computing. There's big data that's all about how to
leverage the huge amounts of information that's being mined on
a daily basis. How do you do that in a
way that's meaningful and efficient. That's frequently associated with cloud computing.
Artificial intelligence and machine learning also get thrown into the mix.
(26:28):
There are services that use machine learning to sort through
big data on the cloud, for example. That way you
get a big, heaping handful of buzzwords all at once,
that's all in an effort to produce a particular result.
Sometimes you use machine learning and big data for legitimate
research purposes that can further our scientific understanding. Sometimes it's
(26:49):
for less lofty reasons. It might be to find the
best way to advertise to different groups of people in
order to more effectively sell products to them. Sometimes you
can get downright creepy, which as the recent Cambridge Analytical story.
Have to do a full episode about what that was
all about. Why raised such a fuss at some point,
But that's that's for another time. It's a complicated issue
(27:11):
that I can't just if I ran into it here,
we would easily go another hour, and UH actually have
to be at the conference keynote in a little bit,
so I can't do that now. At the top of
the show, I also mentioned that there can be an
issue about ownership, like who owns the data on these services,
and usually there's some pretty complicated terms of service that
lays all this out, where you're essentially granting a license
(27:35):
to the third party to have a sort of physical
possession of all that information. And the reason that's necessary
is because the way cloud computing works, the way you
can access it through different devices and in different places,
you have to give the company permission to be able
to show that information. So if I store a file
(27:59):
on my personal computer and I'm the only one who
has physical access to my personal computer, I'm reasonably assured
that I'm the only person who can see it. There's
no real permissions that need to be made or anything
like that. If I'm storing my data on someone else's
machine and I want it to be clear that still
my data even though it's sitting on someone else's machine,
(28:22):
then we have to draw up these agreements. And one
of those agreements might be well, if you request to
see and work with this data, I need to have
the permission to actually serve it to you. If you
choose to share it with someone, I have to have
the permission that allows me to share it so that
I'm not legally responsible If you later on say, oh
(28:44):
I didn't I don't want that person to have it. Well,
if you chose to share it, by the nature of
the cloud computing model, then you know that that's what happened,
that it got shared. So it's largely to protect the
third party providers, but it's also just this very complicated
language that gives them the permission they need to do
the business as they do it. So it can get
(29:07):
pretty complicated. On casual glance, it looks like you're signing
over all your data to another party, um, and you're
not really doing that, at least not in most agreements.
Any ethical agreement would not include that sort of information
in it. But that's pretty much the overview of cloud computing.
Will conclude our story on that. It's a fascinating model
(29:29):
that has its roots all the way back in the
early days of of of mainframe computers. I hope you
found this episode helpful and understanding what the heck that
buzzword meant in the first place. If you have suggestions
for future episodes of tech Stuff, whether it is to
explain something within tech, to talk about a specific technology
or a company or a person in technology. Maybe you
(29:51):
have someone in mind that I should interview or have
on as a guest. Reach out and let me know.
You can send me and email the addresses tech Stuff
at how stuff works dot com or draw me a
line on Facebook or Twitter. The handle for both of
those is text stuff hs W. Follow us on Instagram.
We've got an Instagram account that we post to regularly,
(30:12):
so check that out. And remember, on normal weeks, I
record this show on Wednesdays and Fridays. You can actually
watch me record it live in the studio. You can
just go to the link Twitch dot tv slash tech Stuff.
You'll see the schedule there, and there's a chat room
and everything. You can join in be part of the crowd.
And I look forward to seeing you and I'll talk
(30:32):
to you again really soon. For more on this and
bouthands of other topics, is it how staff works dot
com
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.