We can make a classification into 4 large groups based on location and service offerings.
What is the Public Cloud and its characteristics
These are computing and storage services that are offered through a hyperscale and are available to the general public. They are offered through the Internet. This group includes clouds known as AWS, Azure, GCP and OCI, among others. These types of clouds offer IaaS, PaaS and SaaS services. At the infrastructure level, part of the administration and security is taken care of by the hyperscale itself, simplifying administration for IT teams.
Another great advantage of the public cloud is its pricing, as it is pay-per-use, i.e., if a service is not expected to be used, it can be shut down, saving costs, or, for example, if we need a database with little use, we can use a database with payment for queries, which would be cheaper than setting up a server in which we would have to pay for the operating system licence and the database engine, as well as the computation itself, for example.
Scalability is undoubtedly another of the great advantages of the public cloud, as we can grow and shrink in a simple way and without having to think about hardware resource capacity. Unlike other clouds, the public cloud generally operates in shared mode. That is, on the same physical infrastructure, there can be several clients, albeit all isolated from each other. For IaaS services in particular, at specific times of high demand (as happened during the pandemic, for example), this could lead to capacity problems and is not contractually assured.
In other words, if you have a server switched off, when you switch it on, it may give you an error indicating that there is no capacity in the physical data centre where it is located. These two possibilities can be solved by contracting dedicated infrastructure or reserving capacity to ensure that our server will always be able to switch on, although this entails a substantial additional cost. On the other hand, we also have the option of hiring servers that can be shut down if, for example, there is a peak demand and there is no capacity, with hardly any prior warning. These servers can be hired at a much lower price for this feature, for example for non-critical processes that can be interrupted and resumed, also producing significant savings on the bill.
What is the Private Cloud and its characteristics?
These are computing and storage services that are offered through a service provider and have a more private scope and are not for the general public. They are offered through the internet, although they are generally accessed through a private internal network. This group includes, for example, our proprietary Telefónica Virtual Data Center (VDC) solution. This type of cloud offers IaaS services, and sometimes PaaS, but is more focused on IaaS. As in the public cloud, part of the infrastructure and security is assumed by the service provider, simplifying administration for IT teams. In this case, the billing level would be fixed, whether or not the service is used. This cloud is very focused on loads that need to operate 24×7. A server running 24×7 usually costs less in the private cloud than in the public cloud.
In the private cloud, in general, work is also done with shared (and isolated) infrastructure with several clients, although the service provider can also give the option of working with dedicated infrastructure, which is also usually cheaper than in the public cloud.
Growth in the private cloud is less elastic and tends to be somewhat more constrained than in the public cloud, although the service provider plans for capacity and is looking at increases in demand for resources in order to anticipate this. The private cloud is also more attractive in terms of data privacy, so if you have a service that requires high confidentiality, the private cloud is a great option to consider.
What is the Hybrid Cloud and its characteristics?
It is a combination of both to take advantage of the best of each. We can have critical services in IaaS that require 24×7 in the private cloud and loads in PaaS services in the public cloud, and they can coexist with each other. In this type of cloud, connectivity between the public cloud and the private cloud comes into play. We can do this in several ways, although the most common are through a Site-to-Site (S2S) VPN, with which traffic travels over the public internet encrypted, or through a Cloud Connect, which is a private point-to-point service in which the private cloud network and the public cloud would be treated as ‘sites’ within the MPLS. It all depends on the availability of the service, the network throughput and the latency we need.
What is Multicloud and what are its characteristics?
It would be the same concept as the hybrid cloud, but involving several public or private clouds. For example, we could have a VDC for a company connected to Azure and AWS. In this case, several S2S or Cloud Connects would be needed to interconnect all the clouds, but the same philosophy would apply. Each cloud would be a ‘headquarters’ of the company. Here we could either set up one headquarters as a ‘hub’ or ‘central node’ that routes all the traffic between all the clouds, or connect each cloud independently to the other clouds. This second option is more expensive, but creates greater redundancy in terms of communications and lower latency.
