Public, private, hybrid, and community clouds

- [Instructor] We want to talk about some cloud service models now as defined by NIST Special Publication 500-322. This special publication outlines our different cloud service models and deployment strategies. Cloud models are on the rise due to many factors including the fact that they have distributed storage for our data protection, they're very easily scaled out. They give us automated management capabilities and they give us access from any location among other features. The essential characteristics which NIST outlines for cloud computing are as follows. First, we have on demand self service. This means that users can request more resources or services as they need those and there'll be automatically processed by the cloud infrastructure without the need for any human intervention. If you're familiar with Amazon Web Services, as an example that's exactly how this works. Once you create an account on their platform, you can use their elastic cloud computing service to create a virtual machine as an example where you can choose the amount of resources and storage that you want to use and then that's provisioned for you without any need for intervention by an employee at Amazon. If you contrast this with traditional deployment using physical servers and the manual configuration, this is a much faster and more flexible solution. The second characteristic is broad network access. This means the cloud hosted resources are available on a wide range of devices such as PCs, Macs, tablets and smartphones. This also means that these resources are available from any location wherever there's access to the public internet. The next characteristic is resource pooling. This means that a provider's computing resources are working together in a pool to service multiple customers. This includes not only their storage space but also processor, memory and network infrastructure pooling. It's not feasible to have separate hardware for each customer if you're a provider and the cost savings that is offered by resource pooling is passed down to us as customers and that's one of the reasons that cloud computing is such an affordable solution when we're using a third party. Next is something called rapid elasticity. Just as we can provision cloud resources very easily and quickly, we can also tear down those instances easily. If you think about the elasticity of a rubber band, how that rubber band can stretch out as needed to fit around something and then snap back to its original form. That's exactly the same concept here. We have the ability to very quickly commission or decommission services as they are required. And the final characteristic outlined by NIST is measured service. This means that the cloud provider is measuring or monitoring the services that you're using for various reasons. Some plans may only be billed for the resources that you're actually using. So this would be necessary for billing reasons. It's also necessary to make sure that you have the resources that you need without running out. Cloud providers can keep tabs on your CPU, memory and network bandwidth to make sure that you're receiving the level of service that you're expecting and paying for. Now that we understand the five essential characteristics of cloud computing, let's talk about the four deployment models. First is the private cloud. In a private cloud, the cloud resources are used exclusively by one client or business. The private cloud can be hosted onsite in a company data center or that can be provided by a third party cloud service provider. The significant advantage of using a private cloud is the level of security and privacy that you have under your control. If you're familiar with some of the popular compliance models, such as SOX, PCI or HIPAA, it's possible to use a private cloud and still maintain compliance with these regulations. Private clouds give us more control and privacy in general. The downside to private clouds include the fact that they're more expensive to deploy because you have to purchase and maintain your own infrastructure. And you're also responsible for all of the management of those cloud resources and the underlying infrastructure. You'll see private clouds commonly used with places like government agencies or financial institutions basically anywhere that values security and privacy over the cost of benefits and the ease of use that we would find with other cloud models. Next is a public cloud. This is a cloud infrastructure that is provisioned for open use and this is the most common type of cloud deployment. This is probably what comes to mind when you think of cloud computing where we have resources owned and operated by a third party and those are shared over the internet. This would include services such as AWS, Azure and Dropbox just to name a few. Although public cloud models are extremely cost-effective and they give us great flexibility, it probably goes without saying that the main disadvantage is security and privacy. For the most part, these considerations are out of the hands of the customers and security and privacy is handled by the cloud service provider. Compliance regulation for things such as HIPAA and PCI that we mentioned earlier that is impossible with a public cloud. That's because these types of policies would need to be enforced globally for every user in the cloud. And of course, we know that we have multiple users so that wouldn't be possible and we would be unable to meet those regulations. Community cloud deployments are a new variation of the private cloud model that has some specific use cases. As an example of this, there are certain cloud providers out there who cater solely to government agencies allowing them to use their service. And because of that these sensitive organizations can still reap the benefits of shared cloud computing. They can still save money and have the ease of use while they do still have the assurance that all users are sharing the same privacy, auditing and security principles. That's something that would not be possible when using a public cloud. The disadvantage here is that there are currently no broad standards or best practices outlined specifically for community cloud models. This is a relatively new type of cloud that is still evolving. So there are still some details that need to be nailed down before we expect to see this widely adopted. And finally we have hybrid cloud models. Hybrid cloud models combine two or more of the models that we've already mentioned, private, community or public. This can take advantage of the benefits offered by all of those different combined models. For example, we might have sensitive applications or we might have applications with a high workload that we would want to deploy in an onsite private cloud. And we might, at the same time, use a public cloud for business applications that are less critical. The initial upfront cost is going to be higher than a public cloud model and the implementation can be somewhat complex. It requires that we make sure we choose private cloud hardware that is easily compatible with whichever public cloud service provider we choose. So that's a look at the essential characteristics of cloud computing, along with the four deployment models public, private, hybrid and community clouds.