Ace the Google Cloud Digital Leader Exam: Your Ultimate Guide!

[Music] hi this is Esther and I'm here to help you pass the cloud digital leader exam by Google Cloud let's talk about why

Cloud why Cloud all right so first to understand what's Cloud let's talk about data on premise all right so as you can see in the left there are a lot of

servers all right and these are my servers that I own in my organization and I use them to hold customer data all

right customer data oh God I think my pen isn't working so I just want dra for now so I use this to store customer data

and for example all right now I have um 100 customers so I only need one server let's say for

example but suddenly okay I went big and a lot of people wants to use my website now I'm now I would used to be a small business with 100 people using my

website now a lot of people want to use my website so I bought one more server here and now I'm waiting for it to come but while I'm waiting for it to come I

only can hold 100 customers data so it's very hard for me to scale now let's say um now it's Christmas period right

suddenly a lot of people want to use my website to buy Christmas gifts for their family member so I buy a few more servers to accom accommodate that right

I buy two more servers right okay so now Christmas period is gone Christmas period is

gone these two servers here are in Idol there's nothing there there's nothing there because basically it's money sitting there and not being utilized it's such a

waste of money because I don't actually need that much space at that time so now you can see the problem with owning your own asset right having your own servers

and buying all the servers and owning it it's it's quite problematic and is worst about this is that you have to pay for staff right you

these servers Don't just run on their own you need air conditioning to cool them down right you need to get a man or a woman to like take care of all the

servers and maintain them right if you own them you have to configure them you responsible for everything about them and it's not scalable for example if my

customer size sub grow right I need to buy new servers and I need to wait for them to arrive that takes time that takes

time that takes time and that takes money and hiring this guy takes money right buying this server takes money and

H having I those servers that are not I'm not using it is really wasteful that's why people are moving to the

cloud cloud platform such as Google Cloud right so let's kind of compare what's the difference between on premise

and public Cloud okay on premise versus Google Cloud all right so going on Google cloud is like renting instead of buying a server you rent other people's

infrastructure other people's server so you don't have to hire somebody maintain it cuz it it's own by Google Google hires people to maintain it so it's kind

of like renting versus owning right if you use public cloud or Google Cloud you're like renting you're renting a room right but let's say you have your

own on premise on premise you're you're kind of owning owning so you spend a lot of money building the house right you build the house you spend a lot money

buying the server which are not skillable at all right and then you also have the higher people so move to the public CL we move

to operational expenditure because you're paying either a monthly fee based on a contract or whatever it is all right so that's how you see it so on

premise you spend way more money than if you rent right I mean uh we all know that if you buy your own house it's more expensive than renting a house right

it's it's that concept buying versus renting it's cheaper to rent if you don't want to live there anymore you pack your back and move but if you buy a

house you you have to take care of those assets that's capital and expenditure capital expenditure okay so what is the benefits of a cloud

first of all a scalable this picture illustrates scalability very well so it it grows if it's doing Christmas period or

a new year shopping period right or start of school year kids are buying school supplies and a lot of people on are on your website buying school supplies like

amazon.com you can scale right cuz Google has all these servers All Around the World in different regions and zones right so you can just use their

infrastructure and skill you don't have to buy some and then they take time to come right that's a lot of money and after the shopping season is over right

you have Idol servers that's sitting there doing nothing that cost thousands and thousands of dollar right so that's scalability that's the benefit the other

is you pay as much as you uh you only you sorry you only pay as much as you use but that's the different difference from on premise you you got to buy your stuff right you could be using you could

be paying for things you're not even using okay that's the difference abstraction is you don't have to worry about the finer details abstraction for

example a good example of abstraction is this vending machine you don't have to understand how this vending machine but you just need to know how to

press the buttons you just need to know okay I want Coke press Coke okay now I got to put my cash in all right I put my cash in okay I just need to press this

button then the coke going to drop right down out right that's all you need to know you don't have to know the intricate details of how the vending machine work in order to use it you don't have to know those intricate

details abstraction is don't have to worry about the finer details and for Google Cloud you have encryption right it's automatic

encryption on Transit and at rest you have security these public Cloud infrastructure right right Google have a lot of security guards and cameras and

all that kind of stuff they don't just let anybody go in like a playground okay it's highly

secure okay and lower total cost of of ownership or TCL alert alert this will come out in your exam it's very

very important total cost of ownership why is it lower you don't have to buy servers okay you don't have to hire guy to maintain all your service in your

company so on premise you're buying service for yourself you own it right owning versus renting remember owning your own house and building it from

scretch it's a lot of work it's more expensive renting is cheaper all right so lower total cost of ownership and

that's the benefit of public C now let's go down and look at the different types of public Cloud that is

available right there's Azure there's AWS there's Google Cloud there's neighbor Cloud there's IBM Cloud there's even a Huawei cloud and on and on and on

I can go on a lot but the big tree are Azure AWS Google so if you want to learn forget about the rest focus on Azure AWS

Google most people use AWS okay in certain countries they only use aour so depending on where you live pick your focus now the other concept alert alert

that will definitely come out in the exam is the concept of multic cloud and hybrid Cloud which I will explain to you

simply multic cloud is when you use more than one public Cloud okay so for example Aster's comp

decides to use Google Cloud plus AWS if I use these two I am a multicloud right

if I only use Google Cloud no I'm not multicloud I'm not multicloud if I use only azur sorry I'm not multi Cloud I

have to use more than one public cloud in order to be mul Cloud so I have to use more than one of

one of these sorry one of these one of these public cloud service either Azure AWS Google Cloud nav Cloud yeah what's

hybrid Cloud hybrid cloud is when you use public Cloud plus private Cloud slon premise okay so for example I have my I

own and rent at the same time basically you can think of it like that I'm owning and renting at the same time that's hybrid Cloud for example I have my own servers and

infrastructure in my own company on premise and I'm also a customer of Google that makes me a hybrid Cloud all

right so this will definitely definitely be be tested make sure you understand it all right so another concept that you have to have to understand for the exam

is Regions and zones so if you look here these are the regions that Google has Sydney melbour Jakarta

Singapore and the ones the Circle Blue Dot are current regions with tree zones so zones is like a

place and regions is like a place to but the zones is inside the region all right so Google has 40 reasion one two

one zones 12 21 zones 187 Network Edge locations and 200 plus countries and

territories so if you look at here murn Australia apnc is a region it's a region

and inside this region Melbourne there are three zones a b b c Australia Southeast 2 a

BC so using Australia as an example okay using Australia as example this is a region region okay this a

region and these zones are inside this region okay so

this is Zone one zone two zone three okay so this is the data center in zone one

this is the data center in zone two this is the data center in zone three okay

now let's say an alien comes and this alien takes this Zone away what happens

my data in zone one is all gone it's all gone so in order to protect my uh

data what I should do is I should put my data here and here okay so I should replicate my

data in both of these two zones or even if I can replicated to the third Zone you'll be even better so in that case if an alien takes

away this data center at least my data is still here right and here but let's say if the alien comes

and the alien takes away my data also here and here so what happened my data is gone my gone my data is gone forever

right it went to alien land so what I should do is perhaps I can put my

data in a different region for example Warsaw Poland okay then I could put in the data

center here right so if the alien comes and takes away the data center here takes away the data center here takes away the data center here and least I

still have my data in war South Poland and I can still function as an organization right and

so this is talking about you know if I'm protecting myself from zono failures I'm being Zone failures

okay but I also should protect myself from Regional failures so this whole what if the whole of melburn fails right

so that's the things that business need to think about because now I'm using alien as an example but in real life and

world right sometimes there war sometimes there's fire you know and you know sad things happen so we need to be

prepared and we need to continue as a business to have Disaster Recovery plans okay so you now you really love the cloud and you want to migrate to the

cloud just like birds are migrating to another country during the summer or the winter so there are many ways to migrate the

cloud to the cloud right first you can either rehost so when you rehost this is the minimal time and effort right you

don't have to use much time and effort but then you reap very little business value from it then the other approach is rep platform

for this you have to use more time and effort but then you read more business value which is the diamond

okay so this is the rehost we platform ref

factor and bring imagine and all right IP is like the system died you're not going to use it

already it's rip rest in peace okay you rip and replace this is something you definitely definitely have to remember in the exam

it will definitely be tested and it's under the category of migration all right and then if the exam question says I don't have a lot of time

and I don't have a lot of it stuff then you shouldn't be picking refactor as the answer or reimagine as the answer because the answer should be rehost lift

and shift because we rehost lift and shift takes the least amount of time all right so take note of that during the

exam but the problem with rehost lift and shift is that you you can see that it doesn't have a lot of value right it

doesn't take full advantage of the cloud so for example if you're using SQL on prise you just move the cloud SQL you're

basically using the similar service right from on premise to Cloud now when you refactor you actually change the code that's why it takes more

time and more effort but you reap more benefits of the cloud all right reimagine is more more like

the strategy the vision of things all right and when you reimagine things you benefit most from

the cloud you benefit most from the cloud all right so when you rehost which basically lift and shift you transfer with with no

modification you don't change anything at all and this is good for a legacy application when you re

platform you have minimal changes so you don't change the architecture you don't change the architecture okay and when you refactor you actually

change the architecture that is the difference okay between this two and when you reector you rewrite the

code rewrite the code where needed and when you reimagine it's about the cloud Strat can you think of ideas okay what will my

cloud be like in the future should I use AI should I use more this so that I can give more business

value to my business all right so to end off and review there are the migration types are re

huls which is lift and ship minimal effort you don't spend that much time on it but not that much

business value okay we platform you're not changing the architecture yet you don't spend that much time yet and then there's more

business value when you refactor you change the code so you go on to the computer like this guy over here and you're actually change the code

that takes time right so you'll have more time used but you'll reach more business value now for reimagine it's the

strategy part the strategy part all right rip and replace the name is uh quite obvious

you're saying bye-bye to it you don't want to use it anymore all right remember to memorize this for the exam and you'll do

well migration to the cloud think of it like moving services so you're moving from one home to another home and you need people to help

you move your things yeah think of that for example you're moving from your on premise data center to Google Cloud then then you need to call Moving Service to

move all your things there don't you right so just imagine this as a moving service these guys are helping you move all your stuff to Google

Cloud so database migration service is these movers helping you move your relational database migration to Google

Cloud so this in includes my SQL postr SQL SQL server and Oracle database so what are relational database

relational database are like think of it like human human we have relationship right we have mother grandmother

grandfather and and things like that so relational database could be um this is

cars yeah this database is cars then the other database that conect to it could be like Ferrari or Jeep or

like Mercedes right they relationship they have relationship to each other they're relational okay that's why we call relational

database and it's very structured there's a fixed formet you have to follow the other way to tr transfer is you literally put it in this box you put

your data in here and Google gets it and they bring it to their Data Center and they

transfer the data you only do this when you have too much data that you need a physical Drive migration some customers just cannot do online transfer so they

say get this service and this is a temper resistant rocker guide Ruger D case and they put their data inside of this case and this case goes

to Google data center all right so this is a very um interesting way of transferring data in certain situations the other way of transferring

to Google cloud is storage transfer service storage transfer service storage transfer service automates the transfer of data to from and between object and

file storage systems including Google Cloud Storage Amazon s tree Azure storage on peris data and more so it's multicloud you can

multicloud remember what multicloud mean multicloud means more than one public Cloud right so I can do with more than one BL I can do with AWS I can do with

azour right and this stands for on premise this one okay and you don't have to be worried cuz data is encrypted

what does data is encrypted mean so for example my data is 1 12 three for for instance then I don't want anybody to know my data is one to3 for example is

my super secret secret secret secret password is one to tree and I cannot let anybody see that all right so it has to be encrypted so let's say somebody

steals goes into um my company and steals my data if my my data is encrypted instead of seeing one two three actually this is

what they see uh if it's encrypted maybe they see like X Y one 2

a or they do you see some weird character that's not even closely related to one two Tre so it's encrypted in that sense all right

the other option is cloud interconnect Cloud interconnect Cloud interconnect is a direct physical connection between

your on premise Network to Google Cloud platform short form gcp all right so there's two option when you want to use cloud interconnect

option one dedicated interconnect option two partner interconnect so what is the difference between dedicated interconnect and partner

interconnect the difference is for dedicated you go to a co- location facility while your partner in

interconnect you go to Google's third party so what it happens is that you have a direct um direct physical

connection from your on PR Data Center and the data flow fls here it flows to the code location facility then it flows

to Google cloud data center okay it flows in that

way through a lot of you know wires and stuff like that okay now for partner connect data starts from on previse then

it flows to the partner then it flows to Google Cloud all

right so note usually people only use third party when they cannot access Co facility they usually only use a third

party when they cannot assess a CO location facility so having Co location facility is definitely the best option

pick that so what what is a CO location facility a CO location facility is often referred to as a data center it's a

space where business can rent physical space to house their servers and other it equipment they have cooling system security High internet speed and

more let's talk about the Google resource hierarchy Google resource hierarchy what's that

now it's how things are organized just like your computer Fone right you could have a hierarchy of file for example your mother file is YouTube videos and

then you can have sub fils inside those vales right so that's a hierarchy itself so here the label

for hierarchy is org note org note that's your organization Noe organization not that's why I put this picture here and within the organization

you would you could have different folders that's level treat you could have different folders one folder for maybe like HR usually you might not do that it like that but just an example

then you can have another folder for like um uh business and another folder for like uh management yeah let's say so you could have different folders to

organize your stuff to organize your stuff and within those folders you can have projects different projects not just one you could put multiple projects

inside right and within the projects you could have resources yeah like buckets and

storage okay or your spanner or your virtual machines all the service right you can put it here your Cloud SQL things like that all your resources

you put in level one and that's how deep it gets all right here's the important part you need to know what the exam

alert alert let's say let's use this boy Sammy as an example okay Sam why

can let's say I I gave Sammy assess to this file she will automatically have

assess the projects inside the file and the resources inside the projects inside the file so he has everything inside

this F all right the the assess is inherited

from from the the the top levels right now if I says Johnny you cannot

touch let's say this vow then I I mean Sam if I tell Sam if I put in my i a i a am that Sam cannot touch this vow then

he cannot touch everything inside this vow everything okay I hope that's clear right so it's

orote folders projects resources resources are inside projects

projects are inside folders okay so during the exam they will ask you questions like oh you know I gave Sam

access to this folder would he be able to access this project and this resource Google's site reliability engineering Sr

framework for golden signals for monitoring the reliability of a service number one latency number two traffic

number three errors number four saturation and let's go one by one first for traffic think of your traffic jam

you know if there's a lot of cars on the road there's high traffic if there's not a lot of cars on the road there's low traffic right right right okay so it's

the same for your website your system or whatever if there's a lot of people accessing your system then the demand of your system is high right high traffic

that's the demand your system request per second so for example during Christmas a lot of people will go to amazon.com and buy gifts for their friends right oh I got to buy a gift for

Johnny I got to buy for a gift for I got to buy a gift for Mama so there will be high traffic during salesp Spirit there will be high demand to go

on Amazon right so it's how many requests that reach your system all right now let's go on to saturation what

is saturation saturation is how full the services how much capacity left as capacity is Rich performance will be

degraded capacity of resources utilization like CPU memory and this let's imagine this is your oh what

happened this is your CPU okay you this is your max max okay and you already used this

much so as you reach the max your performance will be degraded so this is something you got to watch out

for now now let's look at latency what's latency think of latency as speed latency is how fast data travels right when we talk

about Ben withd we were talking about how much data could be carried within a certain period of time latency is how

fast how fast how fast okay this white guy he's data Mr Data okay Mr Data Mr Data Mr Data is running so how

fast can Mr Data run and it's often measured in milliseconds okay High latency by the way it's bad it's not a

good thing low latency is good high latency is something you wouldn't be happy with when you're gaming right it's it's super slow low latency is good it's just like you're racing right does it is

it better if you take 1 hour to rise something or you take like 5 Seconds to rise something of course the lower the better right okay

speed now in the principle of latency um there's a there's something called hop okay so imagine this girl is hopping

right now if Mr Data had all this hops here right where he had to hop hop over will he be faster or he be slower he'll

be slower right so each hop increases latency it'll slow him down it'll slow Mr Data down all right

so a hop refers to the journey a data packet takes from one network to another such as from one router to the next each

hop introduces a small delay known as latency due to the time it takes from the router to process and forward the

packet okay to help you better understand let's see this Delivery Man okay he's carrying this data okay delivery man is carrying

data and his ultimate destination is this house okay this house is your network device all right so imagine the

other houses as other network deices that the delivery men have to stop by before they go to your house which is the ultimate destination okay and you're

playing games in this house so you really want this delivery man to bring the data real fast don't you but the problem is he has to go to this house

first okay that's one hop then he has to go to this house first that's not hop another Network device then it reaches

you so it's slower more hops slower all right I think you get it by this point what's a house a house is a network to device all

right now down to the last one error what is error I mean the you guys all know right failed request such as

HTTP 500 errors or timeout system errors all right so to review the Google sites reliability engineering SRE framework for Golden signals for monitoring the

reliability of a service number one latency number two traffic number three errors number four saturation all right what is redundancy redundancy is

having multiple instance of a critical component or system to ensure avail ability and reliability okay so for example to help

you understand for example I have a document this is my homework for my teacher so in order to prevent accident let's say my dog ate my homework and

this copy got destroyed my dog ate it I I have no more hand homework to hand in to my teacher what am I going to do for class tomorrow I'm in trouble so here's the thing this is where redundancy come

in the picture what I do is I make copy iies of my homework right just in case my dogs dogs eat it I make copy one and two

here so it's like having multiple exam cheat sheet so if you lose one you'll still have backups to rely on it's

having multiple instance of critical components replicating your critical components so if one critical component goes down right at least you have the

two or the other three or four or whatever it is it prevents downtime and data lost so what is bandwidth bandwidth

bandwidth is how much data can travel in a given amount of time how much data can travel in a given amount of time so if your bandwidth is very low you cannot

download large files large file down s require higher bandwidth to function smoothly else it will not function

smoothly your internet service provider will tell you the bandwidth of your internet right so higher the bandwidth the more data you can download so think

of bwith like this this guy is the delivery man and he's carrying many boxes of data if he has high bandwidth

he can carry more boxes of data if he has the low bandwidth he can carry less that less packages of data so what is

this delivery man's bandwidth or in other words how much data can this delivery man handle over the network

connection now bandwidth is often measured by bits per second kilobits per second megabits per second or gigabits

per second this is the measurement of bandwidth so again bandwidth is how much data can travel in a given amount of

time and this will affect your download speed of large FS what are the Google

observability first there's Cloud logs Cloud logs is where you store logs I mean the I know but um what are logs

think of logs as that keeps track of everything when and what happened this way if something goes wrong we can look at the logs to understand what happened

and fix it it's for example you writing down um okay I changed my pants at 10:00 a.m. I brush my teeth at 7:00 a.m. I

a.m. I brush my teeth at 7:00 a.m. I

went to school at 8:00 p.m. for example

things like that and so if something goes wrong you know exactly where it went wrong for example if you lost your pencil and you're like okay when did I

when did I take my pencil okay that might be where I went wrong those when and what you can actually future it based on the date for example if I only want to see things that happen on the

27th of July I can look at that if I only want to see things that happen at 9:00 a.m. I can look at that all

9:00 a.m. I can look at that all right you go here when things go wrong and you want to know what time and where what happened you can search for it f

and search for it so what is cloud monitoring Cloud monitoring is like think of it of a like a dashboard with the overall

heal and set you can set alerts with a dashboard yeah dashboard here okay what's Cloud profile Cloud profile is

the all about the CPU power and memory in the exam if you see like CPU power memory usually the answer is CL

Cloud profile Cloud profile all right next Cloud Trace Cloud Trace is all about tracing what tracing latency it's

all about tracing latency latency all right you have reports about latency you have API for tracing latency okay let's

talk about the difference between horizontal scaling and vertical scaling okay so vertical scaling is I I have the

server and I would increase the capacity of the server with CPU Ram or storage so if this this one server right I increase

it I increase it so I make the server more powerful so um imagine I have a pizza and I need to

feed more friends so what I do is I would um make this pizza bigger so that I can feed more

friends now for horizontal scaling instead of making this pizza bigger I would just buy more pizza of the same

size right in order to feed my friends so this is horizontal scale if I have this server let's for example This Server is Tre

gigabytes okay then I'll get two more of 3 GB servers so that this pool of servers would

distribute the load together all right so the difference is vertical scaling it increases the capacity the CPU the RAM and the storage of the

server so I'm making this bigger and better that's vertical scaling horizontal scaling is I'm taking one server and I'm adding more of the

servers horizontal scaling is more flexible and it can handle more traffic but requires coordination between multiple service so you need it you need to

manage um you're adding Pizza you have more friends to feed you add more pizza you order more pizza you order more large

pizza here is I have a mediumsized pizza but I have so many friends coming so I'll make it a XL sidence Pizza

instead in order to feed my friends that's the difference okay okay so now we're going to talk about the concept of infrastructure as a

service platform as a service software as a system so software as a service container as a service and function as a service okay so these are the different

types of serves that you can have in Google right so this BR is the Google services then the second row is what you

are responsible for then the third row is what Google manages or in other words what Google is responsible for and the last one is the remarks that I wrote

okay first infrastructure as a Serv infrastructure a service includes services such as be met engine which or

being V this you are responsible for the data and configuration application code scaling runtime and Osos which stands

for operating system but you don't have to be responsible for the virtualization and Hardware because Google takes care of the infrastructure and the how the virtual machine

functions and things like that for platform as a service it includes app engine Cloud

run you are responsible for data configuration application code scalings so you have to know how to

code then Google is responsible for runtime operating system virtualization and Hardware so for both infrastructure as a

service and platform as a service know how to code especially infrastructure as a service right you need like a lot of good IT staff who know their stuff well

software as a Serv service is good for people who are not very technical They Don't Really

code so Google services are like Gmail workspace yeah Google workplace like Google doc Google Drive all these

software as a system and it's very common in the internet uh Salesforce for example monday.com service now that all software as a service you probably use

the software as a service Outlook these are all software as a service what are you responsible for your data and configuration on Gmail I don't have

the code but I'm responsible for my settings I'm responsible for my contacts I'm responsible for whatever data I put in right

well Google manages your the application code the scaling the runtime the OS and virtualization Hardware is other thing

so for software as a system you don't really manage a lot of things right it's the easiest you don't have to code to

now moving on the container as a service it includes PK and Cloud run what are you responsible for you're responsible for the data and

configuration application code scaling and runtime Google manages the operating system virtualization hardware and you

have the code for this for function as a service there's Cloud run functions it was previously Cloud functions now Cloud run functions they put it inside Cloud run you are

responsible for the data and configuration and application fold Google manages the scaling runtime operating system virtualization hardware

and you have to code CU Cloud function is easy but what happens here is that there is a trigger and this trigger this event it could be you

deleting something or you adding something the storage this is an example but there are many other possibilities but this trigger of you deleting

something or adding something will lead to a small piece of code being run right and it does something so this is what we

call function as a service what about sustainability of Google data centers first one Google data center is 24/7 carbon free

energy by 2030 okay Google has been carbon neutral since 2007 and continues to invest in carbon upset projects and it's designed

to be highly energy efficient and it just cares for the environment the turtles the bird and it just loves nature and reducing carbon emission okay

and here's the important part tip TI for McQ questions about how do your company be more sustainable the answer is usually choose Google data centers to help organizations meet their

sustainability Target all right every company would want to promote themsel right so the answer is always should be Google Cloud how to save money on Google

discounts let's talk about how to save money on Google yeah we all want to save money don't we so how first there's sustain use discount and committed use

discount committed use discount is like okay go go I promise you I'm going to use your service for one to threee so

you get a one to threee contract and then you get lower price for signing this one to treye contract there's no upfront cost but the thing is

you have to pay for it regardless of whether you use it or not for example the contract is for example you paid for 100 gab of something but you only use 80

gb it doesn't matter you have to pay for the 20 gab that you didn't use 100 minus 80 20 you have to pay for that all right therefore you should only use this when

you can predict your workload right it only makes sense for example I always use 100 gigabyte or this much space in my virtual machine right compute engines

then I should go for committed use discount because I know how much I'm going to use this has higher discount rate than S which stands

for sustained use discount however this requires specific term now let's go to sustain use discount

okay sustain use discount is if you keep using it'll be cheaper keep using it'll be cheaper all right you don't have to

commit a specific term right they won't say oh one year one to two years or whatever it is it's not like committed use you don't have to commit to a specific

term there's lower prices for continuous use okay the more you use the cheaper it is the more you use the cheaper is sustained use discount the more you use

the cheaper is it's good for unpredictable VI usage pass patterns when you cannot predict it use sustain use discount okay when you can predict

it use committed use discount and committed use discount is saying okay I promise I'm going to use this for one to

two year all right I hope it's clear when to use sustain use discount and when to use committed use discount okay so now you know how to save money on

Google Cloud don't you so let's talk about what SLI SL are metrics so this could be

latency this could be how available it is right it could be a durability so for example your boss is trying to measure your

performance so he could be looking at how well you do with customers he could be looking at your speed you could be looking at your friendliness so these

are metrics to measure how well something is functioning right metrix so we use these metric or

SLI to make our slos now slos are like go they goals they objectives they're things that we want

to meet for example I want to make sure the latency is always less than or equal 300

Mill seconds so these could be values or range of values where I could say I want it in between 300 to 400

milliseconds so just in case you don't know latency is like how fast the results return to you right if you're playing a game and it's taking forever

to load it has high latency okay so something with low latency is better and faster then you would use your

slos and put them all into an SLA which is a contract that is legally

binding and your contractor is obliged to meet all the service level agreement that you agreed

on now having all these will help our system to just be more efficient

and we can measure how well it's doing now just remember SLA is basically

BL of slos okay it's just a bunch of SL O's together that is agreed upon and a contract and

signed okay let's talk about Google customer care there are four types basic standard enhanced premium Now basic is free

anyone everyone can use it and it has documentation building support active assist recommendation giving you insights on

recommendation all right so for standard it's only English and it's 5 days a

week so if you need other languages you have to get enhance more premium right standard is only

English this has uh dender has unlimited tax support and you can have recommended for workload under development you have can have Cloud support

API all right so for enhance it's 24/7 and there's more languages more languages right if definitely you're paying more therefore

you have faster response time and they will offer technical support escalations and de parting Tech Support to help resolve your uh vendor issues for

premium this is the most expensive one you get technical account manager right and this is something you get if you

have critical work critical work modes and then this technical account manager will help you right in these

critical workloads you will also get credit for Google Cloud boost training because some things you have to pay for that upgade and then you have event

management service for you know planning pick events let's talk about billing

building so if you go to the building section in Google this is probably what you're going to see on the left hand side and this will be tested in the exam

so do look out for it they have reports they have cost table and they have cost breakdown and budget alerts so

what is reports reports are the forecasted total cost cost if you want to find out how much you will spend what are your cost Trends you got to go to

reports if you want to see what are the things driving your cost causing you to spend a lot of money you have to go to reports right if you want to see your

daily cost goodbye select the Futures you can go to reports how about cost table what do you do in cost table cost table is more for

like invoices and stuff like that right it's for invoices so you can see like okay these are my invoices for September

okay these are my invoices for July it has a tabular view of your monthly cost for a selected invoice or

statement so in the exam always always remember that if the question talks about invoices and things like that it's

definitely cost table okay invoices now what's cost breakdown can you see the logo of cost breakdown it kind of looks like a waterfall doesn't it it

looks like a waterfall the water is Flowing down and it's breaking it down what it has it's it

has at a glance waterfall view of how much you would have spent at the on demand price for your Google Cloud usage

how credit saved you money on your invoices it also tells you how much money you save okay so how about budget

alerts budget alerts are places where you can set your budget for example I only want to spend $10 this month on Google Cloud I don't want to spend more

than that on Google CL then you can click um go to the other third section and you can even set percentages and amount and when to

trigger do you want to trigger on the actual or do you want to trigger um when you're reaching do you want to

trigger when you're reaching or when you have already reached the budget basically that's what it's saying all

right do you want me to warn you or do you want to tell me when it actually happens okay that's budget and alerts so if the exam question asks you which one

gives me alerts it's budget alerts I mean duh yeah you'll definitely get this right I know for sure when they when the question asks you where can I see an

overall view you know go to reports when the question asks you where can I see a waterfall waterfall view it's cost breakdown cost breakdown waterfall view

cost breakdown waterfall View cost breakdown waterfall view all right and if question asks you oh how do I find out uh where's my invoices per month uh

I want to find out what's my invoice in September I want to find out what's my invoice in October cost breakdown cost breakdown that's cost breakdown and if the question asks you I

want a tabular view of my monthly costs and view invoice by month right then it's cost table cost table

cost table all right let's talk about data sovereignty and data residency so data sovereignty is about data is governed by the laws and

regulations of the country where it is stored for example if I'm storing my data in Switzerland then my data is

subjected to the law of Switzerland if I'm storing my data in China then it's subjected to the law of China if my data

is tored in Finland it's subjected to the law of Finland it's subjected to the law of where it is stored okay now data

residency data residency is about the physical location of where the data is stored data data residency controls

along specifying the geographical location for data storage for example XYZ country May mandate that certain data must be stored within the country's

borders uh you wouldn't want your citizens passport number and permanent resident card information to be stored outside of

the country that's a security risk so some countries mandate that my citizens personal data or personal identifiable

information you should not be able to store up outside of our country I am am I am am identity and

assess management I a for short it's a service that helps you manage assess for your Google Cloud resources you use it to set policies to prevent

misconfiguration you use it to give roles who can do what let's use Sally as an example so let's say this lady here she's Sally

okay let's say that and uh I I use I am to find out I no no to give Sally rolls I can give Sally

rolls which is shown in this picture here so that I can tell the cloud what Sally is allowed to do here okay so for example I can give

Sally the role of vertex AI service agent um or I can make her the editor or I can make her the owner right or maybe I can make her the cloud PB sub service

that means if I give her this vertex AI service agent that means she can only assess vertex AI service agent she can't do the rest of the stuff right so this

is about controlling who can see what and who can do what controling who can see what and who can do what can this person just delete this file is he or she allowed to do that is this person

allowed to go into this project or this resource yeah you can grant all the asss I am all right what is beyond Corp Beyond

Corp started out as Google's own internal initiative of implementing a zero trust security model but then they uh publish it and it's something that

you can follow using the Beyond cor Enterprise service in Google Cloud so remember Beyond Corp is different from Beyond Corp Enterprise Beyond cor

Enterprise is a user interface designed to help you meet the Beyond Corp all right so what's the idea of Beyond Corp Beyond Corp is trust no one verify

everything it's it's a kind of mindset think of it as a mindset that people have right um do I just let anybody come in my house or do I just like uh first check who they are right you would check

who they are so it's about trusting no one data governance what is data governance you know how in human countries we have government and

government will tell you you can't do this you can't do that and if you murder someone you needs to go to jail yeah because governments go governance is

very important to make sure a society runs in order right so the same thing for data data has to be governed too you can't just let the data run crazy and do

anything you can have the same person with different birthdays and different database you can't have that kind of thing happen you can't let the data run

crazy and then you can't have like someone's birthday or name spelled wrongly and things like that right so we should have rules surrounding data

governments so what is data governance data governance is a framework policies and processes and standards that ensure the proper management

data proper management quality and security of an organization's data so this also means that you should ensure

the consistency of data make sure data doesn't contradict itself so for example the same person should not have different birth dates and different

database right that does not make sense then the other thing is accuracy right you shouldn't have people's birth date wrongly or you shouldn't speak be spelling people's name wrongly data

should be accurate and data should be reliable all right so data governments imagine data governments as like a group

of people here this governing And discussing how we should govern data also data is subjected to law yes data

has to be compliant with law for example in Europe right they have very strong

laws regarding privacy so for example I have the right to be forgotten in Europe I have the right to be forgotten that means if I ask you hey please

remove my address birthday name and your database you have to do that for me in

your right so your data of your company has to be compliant with the law else you're going to in trouble so

we need um there there are also many other standards and that you have to be complied with so how do we manage all that it can go crazy crazy crazy right

we can manage that with Google cloud data Plex Google cloud data Plex is a data management and governance platform all

right then the other thing we can use to manage data is use identity and assess management I am to control who has

access to which resource you know just like humans we have Africans Asians you know westerners different type of people

right data there's different type of data too for structured data it's has rows and columns it's like SQL and relational database so it's very

structured you have to follow the structure what kind of services you can use spinner CL SQL these are all for

structured data all right structured data big table fire store therefore semi structured data therefore semi-structured data semi-structured

data that falls in between structured data and unstructured data there's somewhat in between all right so they kind of have structure and they kind of don't have structure they're kind of

mess but they're kind of Tidy too something like that so some structure like tags or markers like

Json and XML these are considered semi-structured data all right now in f fire store of

course you can store structured data you can store a table in inside of fire store which is a structure data but the

thing is it's more designed for semi structured data okay now moving on to unstructured data so they have no

predefined format it could be images videos audios emails Word

documents so they're not tables and they could be your favorite MP3 F they could be your favorite movie they could be your favorite picture of your dog or

your favorite picture of your family these you put in cloud storage of Google you put in cloud storage of Google okay

they Google Google Cloud Storage is designed for unstructured data you can put like video there you can put your MP Tree there you can put your favorite

music there all kind of unstructured stuff right they don't have uh defined format they don't have a defined format

[Music] all right let's talk about ABY apigy the

first tree word is API API what is API API stands for application programming interface it's a set of rules and protocols that allows different software

applications to communicate with each other all right that might sounds sound a bit complicated but I promise I'll explain exp this in a simple manner

imagine API as two friends in a party these two friends in a party want to talk with each other this friend might be asking the other friend what did you

eat yesterday the other friend could reply Pizza I ate pizza yesterday they're sharing data and communicating

with each other now software components will also communicate with each other like friends in a party they want to Che

and share information for example there could be one website and this website is sending a request to a weather API or

this we website could be sending a request to a API to get cute pictures of cat that could be two system Software

System component talking to each other now another analogy to help you understand is a way analogy for example

you are a customer at a restaurant and you want to order your favorite strawberry milkshake for example you ask

the waiter to go to the kitchen to get the strawberry milkshake okay so imagine the food is the data so the waiter goes to the kitchen get your strawberry

milkshake sends it back to you right so this is what we call an API request and then then you receive the request right so the data travels from the kitchen to

you and you request the data from the kitchen which is your milkshake so in terms of API there are

producers and consumers producers and consumer the cook who is making your milkshake is the producer while you who

are drinking the milkshake is the consumer right and so ape is a like a food bizaar

Marketplace but for the consumers and the producers to come together and they can buy each other's food for example there's one star selling milkshake

another star selling pizza another star selling chicken nuggets I'm already getting hungry seeing all this but yeah so it's like a Marketplace for API it

helps you make your API more discoverable it helps you build manage and secure your apis all

right so consumers they want to know how to use your API so that's also what AB does it provides good documentation to

how to use your API let me show you what an API is this is an API that gets cute pictures of cats all right so this

URL is a request and usually if you have to pay for the API there's a key behind this URL that you have to put

in in order for this response to even come out now this response has an ID key

value pairs this is the key this is the value and URL which is the cat PNG then the width and the height of the photo and if you look this is the photo of the

cat that I got so this is an example of a cat API when I'm when I'm sending a request this URL as the

consumer I'm re the producer sends me this image of a cat back so this image is like the milkshake that I ordered

from the kitchen and the producer sends it back there are also other apis that you must pay for for example weather API right so API

is a business it can be a business right see this company is charging is charging for weather apis okay so let's talk about the

machine learning options on Google Cloud first of all you have the big query this big query is

good for data people good for people who use SQL all right big query is very very Dynamic it can do a lot of things it's a

data Warehouse you you can use SQL on it you can query it you can also use SQL to do machine learning on it so it's very very

Dynamic and most data analysts they will know how to use SQL and what what type of SQL it is it's all quite similar to each

other then the easiest form of it is pre-train apis I showed you the cat API where it generates a random image of a

cat pre-ra apis is somewhat similar to that and it's less customizable but faster you faster to implement because Google did everything

you just use what Google already made so what kind of apis are there of course this is uh not all the apis I didn't list all thems one is cloud

Vision okay Cloud vision is replicating the human function of seeing for image analysis and recognition so when you see an image and

you see a cow on it you know there's a cow on it right so that's a robot replicating the human function of seeing things and

identifying what are in the image and things like that all right next Cloud natural language API

Now leral language is what is it it's conversational it's a robot

replicating human conversational ability and human emotional ability so this includes sentiment analysis what is

sentiment analysis sentiment analysis is a robot trying to find out whether you feel unhappy or happy in this sentence or

neutral but it's not right all the time because at the the end of the day it's a robot and currently sentiment analysis is

not is still not a developed science and robots often get it wrong okay just to let you know Cloud speech detects and text of

speech API for example when you're typing something right and it converts it to speech that's text to speech speech to text is

if you're talking you're talking and then it transcribe it's like your note taker uh it helps you take uh minutes sometimes people use

this to take minutes right so people are talking person a person B is talking and then the cloud API converts their

conversation into text so that a human doesn't have to take minutes well this is just one of the many many kind of applications that is possible with uh

Cloud speech text all right all right the other option is auto ml it's a no code using vertex AI then the last option is the hardest

which you custom your training now sometimes what Google offers you is not enough so you have to custom it that so that it's specific to you right specific

to your industry oh if the exam asks you questions like okay I I want to do machine learning I want to do machine learning on Google Cloud but

here's the thing I don't have amazing data scientists I just have data analysts right so what's the option the option is Big query pick big

query other possible uh exam questions are for example my staff has absolutely no uh

data science knowledge and um I don't have time to train my own data then don't pick custom Training custom Training it takes time what's the fastest is pre-train API then this

should be the answer because it's easy and it's fast Google did all the training already and they tested it they make sure it's somewhat reliable or more than

reliable then they publish it so you're using something that's already built upon hence it's faster and easier all right okay so last talk about the AI

twos in Google there's the contact center AI so contact center AI it gives you insight to customer like the sentiment and

topics for example a customer calls you say I'm upset with this TV it broke immediately like I just bought it and it just broke so it'll give you insights

that this customer is angry all right and it also you can like create this robot cust customer service person right

because maybe you can't afford to hire so many C customer agents so you can take advantage of Google contact center Ai and create

robot um customer agent so it Ed generative AI Power contact center experience and it also has like Asian assist a robot that

assist the call basically a robot doing that okay so this is about all you need to know you just need to know generally for the exam all right

the other ai2 is document AI it takes unstructured data from documents and transform it into structured data right so all your receipt your driving

driver's license it can take that and put it in the table for example I can take all the receipts that you

have and then take out the the item and amount I can just extract the item and amount can you imagine how convenient

this would be if I'm an accountant or um I'm a company where employees can claim receipts of maybe what they ate overseas okay they took a picture of their

receipt they sent it back and they are supposed to claim it I can just use document AI to keep hard copy uh to not keep hard copies of the receipt I'll just have a

soft copy of the receipt and then transfer that into a table so that I can like calculate that's that's just one example of the use case it could be many

many use cases okay now the other concept you have to understand is what is virtual machine virtual machine in a very simplistic way

is like a computer inside a computer right used to be one server only do one thing so let's pretend this is a server it only can do one

thing but what happened with VM is that we found a way to put a machine

inside a machine so that things are compartmentalized okay it's a software program that emulates physical computers

they allow you to run multiple operating system on a single physical machine so virtual machine is kind of like

computers but you can have multiple computers inside computer kind of like that and these computers these virtual machines right they run independently

from each other so they don't really affect each other and yeah so therefore it's good for isolating things and

testing different software in different environments and the other interesting thing about VM is it can emulate the

hardware components like CPU memory and storage to make it be like a computer inside a computer basically

and who who is the one who's making this possible it's the hypervisor right a hypervisor is a software that manages BM

it helps to alloc your resource Ure each VM runs smoothly right there are two types which this will not be tested so I

will not go into details I only cover things that will be tested but this understanding will help you for the other part which will be

tested which is blog storage for vabs so there are two types and do know the difference there are local SSD and persistent

this so what is the major difference local SSD SSD is physically attached to this to the VM it's physically attached

so when the VM dies the local SSC dies of the VM and the data is lost so when you want to use local SSD right you should only use it for temporary

storage but because it's physically connected to the VM it's faster than the persistent this

okay and for persistent this it's not physically connected the VM it's um Network connected

okay so when the vmip dies or is terminated the persistent disc the

information and persistent this persists persists that's why it call it's called persistent disc because you do um is you

can put it you can put um things that are not for temporary use if you're it's for temporary use use local SSD

okay you when do you use persistent this is when you need extra storage in your VM then you attach it online so by

default each uh VM will already have one position this but let's say you need more you can you can buy one from Google all

right compute engine compute engine are for virtual machines so when you want to step set up an instance of a virtual machine you use compute engine and for

compute engine there's two option that you need to know for the exam it's spot VM and preemptive VM so the thing with

preemptive VM it it expires after 24 hours yeah it it can be terminated any time after 24 hours or sooner if Google

needs the resource that's preempted for you so why would you use it why would I use something that expires in 24 hours

well uh you would use it to save money right for Thought tolerant and flexible workloads for example if you're doing batch processing if the VM is suddenly

gone right it wouldn't really affect you because maybe you're only back matching every week or whatever it is it wouldn't

affect it doesn't have to be on all the time so therefore it's F tolerant then the other option is spot VM so these two are both like economical

option when you want to save money for fought tolerant workload so they can afford to stop for a

while the sport VM is um I do for things like batch processing data analysis or testing okay this is what it's ideal for

and they are very cheap compared to regular VM cheaper than regular VM the but it's not reliable because Google can

terminate spot lium anytime right if the resource is needed elsewhere so you should only use spot VM and preemptive VM for workload that are not time

sensitive and can handle interruptions only for workloads that are not time sensitive all right now the benefits of

computer engine it can Auto scale increase decrease that's Auto scale load balance so it can move the

traffic around distribute the traffic and and you can pick TPU if you're doing things with machine learning this

Hardware is designed by Google inhouse and it's um something they're very proud of hence they like to test it in the exam it's a hardware that's good for

machine learning so it looks like this it's a piece of Hardware that is designed for machine learning that's if

you're doing machine learning pick TPU there's also gpus you can also have standard VM more expensive it's a this

Compu engine is a infrastructure as a service so you're responsible for all the configuration and everything inside here we're responsible for a lot of stuff compared to Pas platform as a

service compared to SAS right it's more work so you only pick compute engine when you have the IT people

right you can choose the predefined machine type size and configuration for your computer engine is customized ible

let's say if you want really really large VM really really powerful VM then it's more expensive right the smaller

the size the cheaper it is and depending on what type it is the prices change right what is vertex AI vertex AI is a

machine learning platform where you can develop and deploy manage ml models Okay so it's the whole cycle right you

develop it you deploy it deploy it means you make it available for people to use your or yourself to use now let's go down to bigquery

bigquery is very interesting it's serverless means you don't have to care about the servers it doesn't mean there's no server everything has server

here B ciry use it's like a data warehouse so think of you know how you store boxes and deliveries in a

warehouse well think of all these boxes as data right big query is a place to store data data in a warehouse we need a

place to store data too just like uh parcels and deliveries need a place to be stored the interesting thing about bitc

is that the storage and also does analytics and it also can do machine learning so it can do a lot it can do a

lot you can use it for machine learning SE search geospatial analysis and business intelligence

okay you can work with both structured and unstructured data you can use SQ on it what's SQL SQL is basically something

like this right I can select something from a table and sum it up and show this sum as total SS and from this

table right this is you don't need this for the exam but just in case just to give you the good the good big picture so that you understand stuff but you really don't need this for the exam all

right you just uh I I just thought it's better to show you what SQL is and and not just

have you memorized blindly so we use SQL to per perform complex ques and analytics on large data sets leveraging its powerful skl

capability so this is quering this is me selecting from the table and this is another look at mury

the other storage is cloud storage cloud storage uses buckets right so they could be buckets and buckets you know how you store water in buckets just think the

water is data right you're storing data in buckets right and uh inside the buckets you could have folders inside

the folders you could have subfolders all right so cloud storage is just like for storing things and storing objects and unstructured data okay you

can store like uh what's unstructured data unstructured data is like images

videos pictures your favorite you know music your favorite pop music your favorite country music those are unstructed

data now for cloud storage they have this four storage class which will be tested in the exam alert

alert the first is standard standard is most expensive and you put things there when you use it frequently you put things there when you use it

frequently nline are things you use once a month once a month Cod

line are things you use once 90 days archive are things you use once a year and it's cheaper right so you have

to strategize by the way this kind of cod line standard nearline archive this

concept is in Azure blob stor and AWS as well this concept so once you learn it you understand it you'll be good to go for other Cloud as well they have

different naming but the concept is the same you have to know where to put your stuff if you don't use something frequently you

should put it in an archive else you'll spend a lot of money so you're like oh but what if I it's it's such a hassle for me to move things around then

there's there is an autoc class Fe function autoc class function so it moves things around for you if you find it a hassle for example um if I use a

lot of cat photo this month standard so so I need to assess cat photos frequently I'll use standard but suddenly I don't like cats anymore I

like dogs so maybe autoc class will help me move the cats to Archive and the dogs to standard that's kind of what autoc

class do then it saves money you know that's that's the thing so when you think of the storage classes think of your fridge for example your meat right in order to preserve

your meat you'll put it in the freezer to freeze it up to preserve it so that it doesn't you know rot because you won't use it that often so think of that as cold line it's cold it's cold in the

freezer it's very cold that's things use once in 90 days all right so once you're ready to use your meat or eat your meat you're going to put it down to the uh

Chiller to defrost it right so think of the chiller as standard and nearline those are the chiller you're ready to eat it so it's things that you frequently use so

you can notice standard is frequently used and their line is dirty so 30 * 390 so 3090 one year 3090 365 that's how you remember and this will definitely be

tested in the exam what is looker look look let's look let's look at graphs let's look at bar charts let's look at pie charts

looker look it's for data visualization business intell Ence and graphs it's a third visual to and it's the easiest to remember for

the exam so let me give you an example of how I can use looker in Google for example I'm a company that stores

transactional data in bigquery for example I sell chocolate milk right so I'll store all my sales in

bigquery and then I can take the information from bigquery send it to looker and looker will do a data

visualization to help me see the reports that analyze customer buying patterns over time so this is good for management to see the data

visualizations let's talk about big table big table big table is designed for iot so for example I want to collect

the temperature every single second then I would have a lot of data right so it has to be able to handle a lot a lot

a lot of data so large large amount of data it can be even terabytes or even petabytes of data huge amount of data

it's great for large analytical and operational workloads it's great for time serious data like temperature for the second temperature for the next second temperature for the next next

second it's primarily designed for semi-structured data no SQL so what is iot let me explain what is

iot iot stands for Internet of Things internet of things now Internet of Things is like random objects communicating with each other for

example your cup your cup could be sending signals your computer in the app to tell you what's the temperature of your coffee right what's the temperature of your

chocolate or it could be your airon uh connected to your cell phone so you can turn on your airor even before you go home so that your room is all

clean when you're back home it could be your piano that could talk to the internet and send information to the internet right it could also be your car

right car connected to Internet so all these random objects are can send information to the in the cloud and then

maybe your phone gets the information from the cloud and service and display that information to you let's talk about spanner spanner is

great for Real Time gaming it speaks SQL it's a structured data it has zero downtime and Ure data consistency and integrity it scales horizontally there's

global distribution and and it auto replicates so if I have different players in different countries right right it's it's able to handle that

because it's globally distribution is powerful now what's SQL is something like this all right and to learn more about spanner there's actually a lab

that you can do which is get started with gaming on cloud spanner where you can find it it's Cloud skew boost search

Cloud boost. gooogle then you'll be able

Cloud boost. gooogle then you'll be able to find uh a lot of courses I want to build a website would you build a website with

me what do I use to build a website do you want to build a website yes I want to build a

website build you use app engine to build your website and this is platform as a service so you don't have to worry about the infrastructure right you can

build your app ation with your selected coding language you know human we have languages like for example Korean Japanese English French Italian Spanish

computers also have language of communications for example Java python PHP the go Ruby

C these are languages where you can build websites uh and then you you have to code it but then they will manage

so you Cod you code you put in an app engine then they will it will be deployed now yeah the good part of ab tion is that

you can host different version of website for example uh making some changes on this website but I don't want it to the users to see all my changes

and when I make changes there could be problems right there could be downtime so downtime is like us can't touch it users can't use it so what would happen

is that I give the user one version they're they're using it then I work on the other version right and so when the other version is ready I'll push that version up so that they use the new

version and I'll take the old version down so this is a a way to reduce downtime and improve the user experience all

right moving on to container as a service container as a service these are the two container as a

service Cloud run like this boy is running and Google coer n's engine let's start with Cloud run now in

Cloud run you have three options first you can deploy one revision from an existing container image so you upload an image that you

you already created then the second option is you continuously deployed from a repository like GitHub third option is you use cloud run

function and this one Cloud run function is function as a service okay not container as a service this one is function as a

service so what is function as a service function as a service is small pieces of C right this is a small piece of

cod once something happens like there's a trigger it trigger this small piece of code to run for

example um the trigger can be anything like I add something in my Clon scale or I delete something for my spinner those are all potential

triggers right if I delete something or I add something these are all po triggers for example I can have a pub sub trigger I can have a cloud storage

trigger I can have a fire store trigger I can have other event tar trigger right for adding and deleting and this when I for example add something to my pups up

or do something there and pops up it triggers this code and this code is doing something right it's an event all right this Cloud run function is

previously known as Cloud functions but now they put it inside this uh inside this Cloud run thing so it's became merged together so Cloud run functions

all right for cloud run functions is Ser this it doesn't mean there's no server it just means you don't have to worry about the server it's a function as a

service you can even have multi event trigger Management on functions so to help you understand trigger for example you're allergic to

flowers so when you smell the flower the flower is the trigger and it triggers you to S sneeze okay this smelling the flower is an

event is a trigger then it leads to an event which is us which is a action a verb a verb happens right so this this

Cod is the verb that's happening a small pieces of cat that does something the verb that happens from the trigger all right so that's Cloud run

function now what's Cloud run Cloud run is container as a service when you're doing with images Docker Hub right

you're dealing with containers that's what you're dealing with you can use any language and run your containerized application with Val without managing any

infrastructure it does the load balancing for you and it deploys the website for you so you can run front end backend Services bad jobs host llms and

Q processing workloads okay so this is container as a service

Cloud run is container as a service okay because you're using doer and artifact registry now the other container as a

service is Google kubernetes engine Google kubernetes engine all right now in order to

understand Google kubernetes engine you have to understand containers what is containers containers is like a box that holds everything an application needs to

run imagine it like a lunchbox for software lunchbox for software container is like a lunchbox for software you have your chopsticks you have your uh spoon

you have your fork you have your food right in a lunchbox so inside a container there's Cod which

is the heart of the application the instructions the computer should follow and there's the dependencies which are special tools and how burst the app

relies on and there are the settings which are the the configuration files that tell the app how to behave so that's a container container

is like a box that holds everything an application needs to run so that there's no um problem where it works on your computer but it doesn't work on my

computer because you're using a different um version as me right the container will tell me which version to

use so everybody use the same version all right this is not suited for legacy system you would need to reector Alliance so it's not

possible it enables rapid application deployment testing fully automated cluster life

cycle management pod and cluster autoscaling now for container a container is a cloud neutral right so you can take this container

here uh take this image and put it into different for example I can take an image from majour and use it in Google I can take an image in Google and use it in ABS so it's

Cloud neutral in that sense all right so let's now talk about security security in the cloud how do we protect ourselves and be secure on the cloud first there's

security Command Center and there's Google Cloud armor Google Cloud armor is for dos attack and application attacks like cross site scripting

xss and SQL injection so what is Dos attack it's like this evil guy okay he gets multiple machines and this is he manipulates

these machine these machines are like his puppets he gets his puppets to attack your server with many many many requests

so basically it's like maybe your server can only

take 100 people right so what this bad guy will do he will overload your server with so so so many requests that your server stops to work and the service is

no longer available right so you could like for example websites right if um 7 billion people use a

website right it could break down so what this bad guy is doing is it's it's just sending man man request until the silver cannot take it so the silver shuts down all

right now what's security Command Center security Command Center is like there's vulnerability dashboard and it helps you

to identify vulnerability and trat reporting service it also it can also assess the virtual machines and containers in your Cloud environment for

vulnerabilities vulnerabilities vulnerabilities are like weakness weakness that the attacker can take advantage of and get into your

system and do bad stuff those are vulner vulnerability so there's the premium and Enterprise tier of security Command Center include

the security posture service right so if you pay more money you can get the security posture service it also can scan for

misconfiguration or you really don't want to do misconfiguration then it can manage compliance with industry

standards so if the exam question talks about D do attack which is denial a service attack because the web page gets

overloaded with request that it doesn't work anymore so it's a denial service attack then you should pick Cloud armor Google Cloud

armor but if the exam question is talking about detecting vulnerabilities and trats and having a vulnerability

dashboard security Command Center is the correct answer all right let's talk about data capturing data processing data store and analyze

data so think of data as this apple right I'm picking apples when I'm picking apples I'm capturing data like this aune she's picking apples right

she's capturing the data after you capture the data you have to process the data process the data so think of

processing the data as washing the a apple chopping the Apple in the pieces so that you can store it in the fridge store and analyze it you can

analyze that b and see how much sugar there is how much vitamin C there is things like that store and analyze it

okay or the other way is you can think of data as water right you're capturing

the water right you you gather all the water then you put it in a pipeline and the water

moves in the pipeline and moves and now you have to process the water you need to clean the water so maybe you do like reverse on

osmosis or you use ultra ultra violet Rays to cure the bacteria inside you filter it you clean the water make sure it's

usable then after that the water flows and it's stored as Ice Cube in the fridge that's another way to to look at data

right and so we need to capture data so this is the the time we're capturing data what do we use to capture data we can use

pub/sub now Pub slub you can just think of it as YouTube right think of it as YouTube Right a

YouTuber has is a publisher and a YouTuber has subscribers like you so the

subscribers will whenever the YouTuber publishes any videos they will receive an alert and then they can watch the video all

right so think of that Pub and sub Pub stands for publisher sub stands for subscriber so it's a messaging service

right publisher sends a message subscriber sees it the other subscriber which is subscribed to this service sees the messager messages that the publisher

sent okay so this is a way of capturing data okay imagine the this message is data

now after the data flows through the pipeline it can be it has to be processed right we don't want things to be messy messy we want it to be clean and tidy

so that we can store it we don't want to store messy things afterwards we use cloud data flow to to kind of do the ETL

which is extract transform load it executes apach Beam pipel lights that ingest data into Data Warehouse

such as bqu right it job is to ingest the data to bigquery or other place so the of data flow is take from here

and bring to here take from here and bring to here so it reads from one or more sources and transform the data and write the data to the destination it can

do B stream made uh batch and streaming data processing so what's batch batch is

like I take eight apples or I think some apples at a time maybe like 20 apples or 30 apples I think 30 apples at a time and I

will I will carry those dirty apples and bring it back home to be washed that's batch process processing I I take a group of them and and wash it and wash

it but stream processing is I take one EP bring it to wash store I take one EP bring it to wash store I do one at a

time so you typically stream data processing is used with like like uh

time series things for example um the temperature at a.m. was like 20 Celsius then I'll send that in right I can't

wait to batch processing usually usually okay usually that kind of stuff it will be streaming data processing because I need to know the temperature right I

want to know the temperature 1 second AG go right now then it has to stream right if I want to know the temperature of 5 minutes ago it got to be it got to

string because I need real time data so it has to stram I cannot wait for you to um gather uh like maybe like 5 years of temp weather

temperature and then bring it over or like uh usually it's not 5 years shorter maybe like 1 month c 1 month Cent all

right now cloud data proc is something similar to cloud data flow but the thing is cloud data proc uses Hadoop and Spock

Hadoop and Spock okay cloud data flow uses AP beam pipeline that's different difference so in the exam if if the

question as Hado and Spark data processing pick uh cloud data proc okay uh a good way to

remember is that this Hado has a p and data proc has as a P2 so these two comes together okay basically both cloud data

proc and cloud data flow does ETL right but they just use kind of like a different technology cloud data Pro uses

Hadoop spark while cloud data flow uses uh Apache Bean pipelines okay so uh cloud data from takes from here brings

to Big query okay and what does big query does big does storage and Analysis it's a data warehouse think of think of

this right think of data as boxes and you're keeping it somewhere keeping it in a warehouse okay and then this guy is

going to analyze each of the boxes all right so this is the typical flow from an external system you'll go to Pub sub

a mastering service publisher subscriber that you'll go to data flow

right data flow or data block right then it'll go to big query and looker what's

looker the first four words look look it lets you know that something very visual right so it's for data visualization so you can see pie charts gra you can

customize whatever you want to see you can just um click on it it's quite easy easy to use let's do a recap of what we learned first of all block storage for

Block storage there's persistent this and local SSD local SSD is for temporary files while persistent this is for

non-temporary files it is not physically attached to the VM like local SSD what is VM VM is like a computer inside a

computer compute engine compute engine lets you pick your MVM the machine tipe the size and configuration Google kubernetes engine

is a container as a service and it helps to orchestrate manage scale containerized application on kubernetes Google called automar helps

to protect you from dos Attack app engine helps you build website Apple helps you manage your

application and manage your apis and make your apis more discoverable and what is an API API is like two software components like

friends at a party they want to chat and share data it is also like a waiter who goes to the kitchen which is like one software component and goes back and

brings the data information to the consumer the customer the food is like the data and the food travels from one software component to another software

component we also talked about looker which is a data visualization platform it has pie charts bar charts and anything you can think [Music]

of we talked about security Command Center which helps to detect vulnerability and trat we talked about the AI twos where there

is document AI document AI basically takes your documents such as receipt driver license and forms take those unstructured data and put it into a

table right which is structured structured data they also have contact center AI which uses generative AI

powered contact center so you can use AI to assist your human agents or you can use robot agents to assist your

humans we also talked about machine learning options which is Big query great for data analysts who know SQL pre-train apis which is good when you don't have

your data to train and is less customizable but faster and easier Auto ml which you can use vertex Ai and no

code custom Training you need to code on your own and you need data scientists to do this we also talked about the types of

migration which is rehost lifting shift minimal time and effort but minimal business value and you benefit very little from the cloud there also rep

platform refactor reimagine repin replace we also talked about the different type of service there's infrastructure of the

service there's platform as a service there's software as a service there's container as a service there's function as a

service these are the different type of service with infrastructure as a service being the service that you're most responsible

for and software as a service as being a service that you're least responsible for things and you need last technical skills because you don't need

a code at all for software as [Music] system we also talked about the different types of data we have structured data semi-structured data and structured

data spanner is good for structure data cloudsql is good for structure data big table fire fire store is good for semi structured data cloud storage is good for

unstructured data like your favorite music your favorite movie we also talked about data residency which means a country might

say I don't want you to put my Citizen's permanent residence card in another country because that would be a security compromise that would be

dangerous data sovereignty is that data is governed by the laws and regulations of a country where it is stored for security model you have

privileged assess which is dangerous means basically super user this privilege means you only are given asss that you need for your job zero trust AC

cure means you authenticate every device you trust nobody we also talked about the need for data governments a framework of policies

and processes and standards that ensure proper management quality and security of an organization's data we also talk about the idea of

Beyond Corp which is a trust no one verify everything is Google's implementation of the zero trust security model

we also talked about the IM I am is us to set policies to prevent misconfiguration it's about what kind of assess am I going to give Sally am I

going to allow her to be a Vertex area service agent or I'm going to allow her to go into Cloud SQL or am I going to allow her to use spanner I can give her

different roles and different assess based on what she need and I should only give her the minimal level of privileges least

privilege we also talked about big query which is great for your data analysts who are familiar with esm it does

storage it also does analytics and it also does machine learning it's very Dynamic and it's a absolutely unique offering by

Google we talked about big table which is great for I because with iot you have all these sensors sending data constantly you will have very very large

volume of data terabytes or even par petabytes of data we talked about file store which is good for sharing

files it's great for mobile and games it's like a network file storage system a file level storage architecture that allows multiple users users and devices

to access a store file over in the network and it's almost real time means if I add a file inside you'll be able to see it almost

immediately we also talked about vertex AI which is a unified platform for developing deploying and managing ml models is a food thing

we also talked about container as a service which is cloud run and a function as a service which is cloud run

functions and how Cloud run functions is work is that there's a trigger that runs small pieces of code and the small pieces of code does something all

right we also talk about the concept of regions and zones zones are inside

regions for example Melbourne Australia APAC is a region and Australia Southeast 2A Australia selfies to B Australia

selfies to C are zones and this Tre zones are inside the region of melbour ladies and gentlemen we have reached Al Final Destination thank you

for joining me on this Cloud journey I look forward to seeing you again on our next adventure safe [Music]

troubles like comment and subscribe [Music]