Get startedGet started for free

Cloud deployment models

1. Cloud deployment models

Welcome back! I'm Lis and I'll be your instructor in this second chapter. Here we'll talk about cloud in practice, beginning with cloud deployment models.

2. Cloud deployment models

Choosing a cloud deployment model is one of the first choices to make when migrating to the cloud. It's informed by how much control you need over your cloud environment, which is defined by your business requirements and budget. There are several models, but the most common are public, private, and hybrid. Let's go through each of them and understand their unique value.

3. Private cloud

The private cloud model is private by design and designated for exclusive use by its tenants. This means to access a private cloud, you need to connect to a network link, which means special network access needs to be set up by IT. An organization may hire a third party to host their private cloud infrastructure or host it themselves. The private cloud model can be appealing because infrastructure is designed for exclusive use. This gives an organization direct control over the resources available, alongside with how data is stored. For example, you can specify what kind of operating system and the exact hardware you want. This is flexibility that does not come with other models. This is especially useful if there are specific security requirements or risks. But, it usually requires more investment in terms of time and capital expenditure compared to the other models. Consider the infrastructure that needs to be designed, hardware that needs to be bought, set up, and maintained. However, this decreases if a third-party, like IBM or RackSpace, provides and maintains the hardware. You maybe wondering how this is different from on-premise. If so, good question! The infrastructure in this model adheres to cloud principles, meaning it uses virtualization that allows on-demand compute resources. Private cloud can also be located off-premises.

4. Public cloud

The second model is the public cloud. The cloud infrastructure is shared and open for use by the general public. The infrastructure is owned and managed by a cloud service provider, like AWS or Azure. Public clouds are Internet accessible. This lets organizations get started quickly with minimal investment. They don't have to host or commit to an infrastructure, instead they can pick what services they need from a provider. They can also instantly buy more capacity making it easier to scale. If it's important to have direct access to the data centers or hardware than public cloud is not a good choice. Public cloud providers keep their data center locations secret, even to their clients, for security reasons.

5. Hybrid cloud

Why not both? Enter the third model: hybrid. This is when an organization uses a combination of two or more distinct models. The different models interact with each other via a network link and can share data and services. It's more of a question of where data and services are physically stored. For example, you could store sensitive patient data on a private cloud for security reasons and use an application on the public cloud, like a business intelligence tool, to process it. Hybrid clouds are useful in the case of cloud bursting. This is when a private cloud is overwhelmed by demand and hits capacity. To avoid disruption of service to users, traffic is moved to a public cloud instance. A classic example of this are seasonal spikes. For example, retail businesses can expect extraordinary traffic during sales periods like Black Friday or Boxing Day. This allows organizations to cost-effectively handle periodic spikes with pay-per-use pricing.

6. Other deployment models

There are other cloud deployment models to consider including multicloud. Multicloud combines different cloud provider services. For example, an organization could use Azure for backups, AWS for website hosting, and Google Cloud for analytics. Multicloud provides flexibility on choosing pricing plans and service offerings. It also reduces reliances on one vendor. Don't confuse this with the hybrid model which uses a combination of cloud deployment model, not services!

7. Other deployment models

Another model to consider is the community model. This is when the cloud's infrastructure is shared by a specific community for their exclusive use. Usually this community has a shared interest or concern, whether it be having the same security requirements or jurisdiction, or sharing a mission. This shared infrastructure allows for easier collaboration and sharing of data, whether it's research data between institutions or government data across agencies. The infrastructure itself can be managed and hosted internally or externally.

8. Let's practice!

Alright, let's check your understanding!

Create Your Free Account

or

By continuing, you accept our Terms of Use, our Privacy Policy and that your data is stored in the USA.