Virtual Machine Scale Sets

Video Activity
Join over 3 million cybersecurity professionals advancing their career
Sign up with
Required fields are marked with an *
or

Already have an account? Sign In »

Time
22 hours 25 minutes
Difficulty
Intermediate
CEU/CPE
24
Video Transcription
00:00
>> Hey everybody and welcome back.
00:00
In this lecture we're going to be talking
00:00
about scaling sets,
00:00
which is pretty interesting.
00:00
It's another way to
00:00
build availability and redundancy
00:00
and to your virtual machine environments.
00:00
In this lecture, we're going to be talking
00:00
about an introduction of what skill sets are.
00:00
Then we're going to be talking about
00:00
some pretty cool stuff
00:00
autoscaling versus scheduled scaling,
00:00
and then horizontal scaling and virtual scaling.
00:00
All this scaling fun stuff.
00:00
Let's go ahead and talk about what a scale set is.
00:00
There are going to be times when you need to have
00:00
a virtual machine that is able to
00:00
handle lots of requests,
00:00
or lots of traffic,
00:00
and there are times when maybe that
00:00
one virtual machine cannot handle that entire capacity.
00:00
Then you want to begin
00:00
distributing that traffic to multiple virtual machines.
00:00
One way that you can do that's
00:00
>> a way that's scalable and
00:00
>> efficient is part of our
00:00
>> high availability solution here,
00:00
>> is by leveraging a VM scale set.
00:00
Really what a VM scale set
00:00
does is it allows you to deploy and
00:00
manage many load balanced identical VMs.
00:00
Let's say you have a Linux web server,
00:00
and web server is going to be hosting websites.
00:00
But maybe you have an e-commerce site on
00:00
there and maybe it becomes an overnight success.
00:00
Well, typically when that happens,
00:00
you get a lot of traffic.
00:00
Maybe have a big sale and a bunch of
00:00
large websites start sending traffic your way.
00:00
You want to make sure that you're able to
00:00
handle that workload capacity.
00:00
One way to do that is to use a VM scale set to
00:00
deploy identical web hosting VMs,
00:00
just like the one that you
00:00
>> would be using for your core.
00:00
>> To diversify the traffic
00:00
to these various VMs so that everything
00:00
stays available so that you don't
00:00
pick your performance throttle on a single VM,
00:00
and you're able to maintain
00:00
the speed and customer satisfaction
00:00
on the actual application.
00:00
VM scale sets can be intelligent.
00:00
That's one of the beauty behind
00:00
this service is that it can
00:00
detect when your performance thresholds
00:00
start hitting a certain percentage.
00:00
You can also determine what you want them to
00:00
be before it actually spins
00:00
up and identical virtual machine.
00:00
For example, maybe you noticed that
00:00
your CPU usage hits 60 percent,
00:00
and you want to set a policy in
00:00
place so that when it does that,
00:00
another VM begins to get booted up and end
00:00
deployed so that it can start
00:00
scaling and distributing
00:00
>> that traffic across the two VMs.
00:00
>> You can do that. It is
00:00
intelligent and can automatically do that for you.
00:00
Even if you were in bed,
00:00
you're asleep, and
00:00
you didn't know that this was happening,
00:00
it would be deploying this for you and making sure that
00:00
your virtual machine does not go down,
00:00
your website doesn't go down,
00:00
or whatever it is that you're trying
00:00
to keep highly available.
00:00
To do this, it uses the load balancer to share
00:00
that traffic across
00:00
the multiple virtual machines that it's deploying.
00:00
It also leverages a health probe
00:00
to determine the availability,
00:00
the health of the actual virtual instance itself.
00:00
It does this by pinging
00:00
the VM consistently to make sure that it's staying up,
00:00
and it's not going down.
00:00
Scale sets do support both Linux and Windows.
00:00
You do have that flexibility as well.
00:00
Let's go ahead and start talking
00:00
about the different types of scaling.
00:00
We're going to have horizontal
00:00
scaling and vertical scaling.
00:00
Horizontal scaling is basically when
00:00
you are replicating the same virtual machine,
00:00
and you're scaling out or horizontally.
00:00
Maybe you start off with one,
00:00
and you notice that you need to start
00:00
distributing that traffic across multiple VMs.
00:00
You can start to scale horizontally
00:00
and deploy two or three or
00:00
>> four or however many you need.
00:00
>> You can set these policies in place for
00:00
scale sets when you set up a scale set.
00:00
Then you can also scale vertically,
00:00
which just means that you're adding
00:00
more resources to the VM.
00:00
In my mind, I always think of this as beefing
00:00
up the virtual machine or in other words,
00:00
just making it more powerful, more stronger.
00:00
You can add more, compute,
00:00
more memory, more storage,
00:00
whatever you need in order to
00:00
scale vertically and meet the demands.
00:00
In addition to this, you have the ability to scale
00:00
a scale sets by using scheduled scaling or autoscaling.
00:00
Now, what do these actually mean?
00:00
So schedule scaling is
00:00
when you're going to hit a
00:00
>> certain performance threshold,
00:00
>> maybe you don't need to have
00:00
these virtual machines running all the time,
00:00
and you want to make sure that
00:00
you're optimizing your bill.
00:00
You don't spend so much money
00:00
on redundant compute resources.
00:00
You set up a scheduled scale to make sure
00:00
that you're meeting the
00:00
>> demands for a particular window.
00:00
>> Maybe it's a time or date or
00:00
a certain set of days in a week
00:00
when you know that you're going to need multiple VMs,
00:00
you can set a schedule scale
00:00
to make sure that the VMs are there,
00:00
they're present, they're able to meet the demands,
00:00
and then you can turn them
00:00
>> off automatically at the end.
00:00
>> That's a really nice thing about
00:00
scale sets that you can automatically
00:00
deploy and automatically deprecate
00:00
the VMs that are going to be deployed.
00:00
This is greats, like if you
00:00
know that you're going to have a sale
00:00
going on for an e-commerce store, maybe, you know,
00:00
there's predictable traffic for
00:00
an e-commerce store when it's not on sale,
00:00
when the products are not actually on sale.
00:00
But then you know that you
00:00
can have a pretty good idea of the type of traffic
00:00
that you might be getting when
00:00
a store might be putting a product on sale,
00:00
a product that might be popular.
00:00
You go ahead and throw redundant VMs at it,
00:00
you use a scheduled scale to meet that demand.
00:00
In addition, you have the option to use auto-scaling.
00:00
This is really for those unpredictable workloads.
00:00
Maybe you don't know what to expect.
00:00
Maybe this is your first go
00:00
around with your e-commerce store,
00:00
and you're not quite sure
00:00
what the demand is going to be like.
00:00
You're setting that baseline still.
00:00
Autoscaling is a good way to do that.
00:00
It's a good way to make sure that your application,
00:00
your website is staying highly
00:00
available despite what goes on.
00:00
This is going to help you sleep at night
00:00
because actually I have a pretty funny story.
00:00
I actually did this.
00:00
I had a website,
00:00
it was on an e-commerce store.
00:00
This was before I got into the cloud and all that.
00:00
I didn't really know what I was doing.
00:00
I was dabbling into
00:00
Application Development for the first time.
00:00
I had websites,
00:00
I had a sale going on and actually was planning to
00:00
do quite a bit and revenue for that sale.
00:00
I was getting a lot of traffic because I had
00:00
run these ad campaigns for a long time.
00:00
I had social media influencers help me promote
00:00
this product, and funny enough,
00:00
the application actually shut down
00:00
the night that I launched it because I
00:00
didn't know how to architect auto-scaling.
00:00
Don't be like me. Make sure you schedule auto-scaling.
00:00
If you're not sure what your demands is going to be,
00:00
you don't know what your traffic's going to
00:00
be, just have it in place.
00:00
It doesn't cost a lot,
00:00
and you may actually end up saving your business,
00:00
saving your profit margin,
00:00
and doing better off
00:00
because you had that redundancy in place,
00:00
you have that availability in place
00:00
and you prepared for that by using auto-scaling.
00:00
Yes, less than of the day. Don't be like me.
00:00
Make sure you have auto-scaling.
00:00
To summarize this lecture,
00:00
we talked about skill sets,
00:00
we talked about what they are.
00:00
We also just talked about the difference between
00:00
horizontal and vertical scaling and the difference
00:00
between scheduled and auto-scaling.
00:00
That's it for this lecture.
00:00
I'll see you guys in the next one.
Up Next
Azure Infrastructure Solutions Lab
1h 45m
Manage Azure Resource Deployment by Using an Azure Resource Manager Template Lab
45m