in this lesson we're going to describe the differences between the types of cloud computing if once again we go and check out the az 900 skills that are tested we're really focusing on these five skills that are assessed define cloud computing describe public cloud describe private cloud describe hybrid cloud and compare and contrast the three types of cloud computing so that's the focus for this particular lesson now i've talked a bit before about what is the cloud and the idea of the benefits of the cloud that elasticity that agility the high availability but i focused on cloud azure the public cloud but if i really think about the cloud there are particular properties that we think about that define the cloud there are key principles now a key principle is around pooling resources i.e i don't want islands i don't want hey this group of servers and this storage is for this particular group of people then another little island over here and another one over there it's very inefficient they might have busy times that are different from each other and they always have a spare amount of capacity so if i have hundreds of islands that spare capacity really is quite high there's a lot of wasted space and it limits any particular island from kind of peaks if i pull all the resources together instead of having a hundred islands i just bring all of that resource together now i have this massive amount of potential capacity and if one business group is having a quiet time another one can really leverage that so that pooling of resource gives us this fantastic elasticity that workflows can take advantage of then there's a whole idea of self-service i don't want to have to go to some administrator and say hey please go and create me a virtual machine i want the idea of self-service now self-service still needs controls so often we think about certain maybe quotas i give them a cloud of a certain size to control hey how much resource are they consuming there may be the idea of certain policies or standards or templates that they can actually leverage but the whole point here is is this on demand they go and provision things as they need to there are going to be things like showback i can show them what they're using maybe it's also chargeback is another important aspect there's also things like separating of networking there's a whole bunch of different aspects but we have some of these key kind of tenants i don't want islands i want to pour the resources that gives me this great agility between everyone so if we have this idea of the cloud and some of those key tenants well then i can offer that from different places now the obvious one we think about when i think about what is a cloud absolutely i think about the public cloud eg azure azure is an example of a public cloud this is typically the most complete offering again when we talk about the idea of a cloud there's a certain set of capabilities we expect now a lot of those capabilities are actually exposed via the management infrastructure that sits around it that exposes these concept of clouds of quotas of self-service of showback of chargeback of offering different types of service and if i think about a public cloud offering that's really going to have the most complete set of capabilities to meet all of those definitions of the cloud to have true software defined networking to have true rich policy capabilities and show back and charge back and all of those different role-based access control elements we would expect also for me as a company well this is going to be true opex i really am just paying for what i use remember also this is primarily accessed over the internet now as we talked about in previous lesson i can from my on premises have like a site-to-site vpn i'd have a point-to-site vpn for a particular user i can use things like express route for a private actual network connection not even encrypting a tunnel over the internet but that's a very common pattern we think about a public cloud we're offering it over the internet and with the public cloud we really think about it is just limitless we think about many regions i.e i can provision services in a lot of different places and many regions there are many services that i can pick from i'm not limited to vms or a container i can have artificial intelligence services i can have managed postgres or my sequel i have bought in the cloud cosmos db databases huge range of capability storage from hard disk drive based to nvme super low latency capabilities i have the ability to have services offered out to the internet offered privately geographically distributed just a fantastic set of capabilities available to me and again there's this very strong idea of governance policies role-based access control budgets to control actually what i can spend but i can also have private cloud absolutely on premises i have the idea of a private cloud i think about well i had those physical servers so i can think okay well we had those physical boxes that were running some kind of hypervisor a key definition here about that private cloud would then be the management infrastructure you put on top of it there's going to be a certain management set of software that i leverage that would expose to the end user portal that would enable those clouds of different quotas and different types of offering to actually be leveraged so i can still create this idea of a certain set of service i want to offer to different groups of the business so as a business unit person i'll see certain services offered to me i'll have a certain amount a quota but here's the thing although as a particular business unit i might see a particular set of surveys and as a business unit maybe i pay for what i use as a company this is capex i'm buying those servers i have to go pre-buy i have to buy the licenses i have to purchase all of those things and once again i have a certain fixed set of services i'm going to offer for a duration of years but i am running my own servers so i have full flexibility within the capabilities of whatever management stack i'm running that is providing that private cloud set of capabilities now one of the things we will see later on is azure for example but they actually have the ability to extend and enable services on premises you'll hear things for example about azure stack which has different types an azure arc which brings azure management capabilities from azure so a single pane of glass it can actually bring different types of azure services built on top of kubernetes could be app services could be data services could be ai services so i actually get a consistent not only management infrastructure but also consistent services between the public and the private cloud which brings us to the idea of hybrid and hybrid really just means we're using both we're using a private cloud and we're using a public cloud often seamlessly i might think about i'm offering a certain service from my on-premises private cloud but in busy times i'm actually going to burst up into the public cloud i might have some global load balancer to distribute my workload so it can normally point to on premises but in these busy times or if there's a failure well now i go and spin things up in the public cloud that consumption base so when you pay for it when i need it so i go and burst out into the public cloud maybe some services i might have an anchor it's all about anchors maybe i've got some mainframe workload i can't move to the cloud so some things i still need to run on premises but other services i can move out in the cloud so i'll have a combination of them but we think about i don't want to use completely different tooling this is where i think oh okay things like azure arc i can have the azure management the resource-based access control the policies the whole governance model and then bring the services as well to my on-premises so as a developer for example i don't care i work a certain way and it can really run anywhere i want that's the best type of hybrid cloud where i as a user of the service do not care if it's running in my private cloud or it's running in the public cloud so i get a lot of flexibility here i might have data requirements regulatory requirements that say you're not allowed to run in the cloud public cloud now that's a lot less common today but it may still happen or then i want to run a private cloud but i still want the features that i'm used to it might be i need to operate in a country where maybe there's not an azure region i can use something like azure stack i just like hub i just like edge i just like hci different maybe it's complete rack with azure stack hub maybe it's single boxes with azure stack edge maybe it's my own infrastructure running a special azure brand of the hypervisor and software defined networking and storage and other administrative controls with azure stack hci so i can run those where i need to run it on premises but other times i'll go and use the public cloud so those are really the different types of cloud a lot of companies want that consistent model consistent policy consistent everything else and i can get that through services like azure and azure stack and azure arc where probably my primary scenario will be running things in the public cloud but there might be times maybe because of the latency there's still a certain distance between maybe an on-premise location in the clouds data centers i want to run it on-premise maybe it's artificial intelligence scanning a work line looking for safety issues or sensors i need to react super quick i want to run those workloads at the edge very close hey i'll run it in my private cloud so i can run things using my infrastructure i can run things in the public cloud i can use that hybrid approach as well so those are really the three types of cloud computing we're going to see our big focus here obviously is going to be the public cloud and using azure