Characteristics and Types of Clouds

Cloud computing is also convenient because of its characteristics. These characteristics provide multiple benefits to the user and significantly reduce time spent by DevOps personnel. As a result, less human effort is needed.

In this lesson, you will learn about those characteristics and the different types of cloud deployment models provided by cloud providers.

Characteristics of cloud

The below diagram will give you a brief overview of this lesson.

Scalability and flexibility

As cloud computing is based on the logical separation and virtualization of underlying physical hardware, almost every component can be controlled using a piece of software. This eases the process of scaling resources on demand.

On-demand scaling works not only upwards but also downwards. You can upscale and downscale according to the need, without any manual intervention or continuous usage monitoring. The “auto-scaling” buzzword is a result of this scalability.

Flexibility in Cloud Computing allows you to change your deployment architecture without much effort. In addition, cloud computing will enable you to switch between different architectures very easily. For example, you may need to change your architecture to adjust to the cost, the latest security practices, and so on.

Broad network access

“Broad network access” is an inherent property of any cloud service. With broad network access, you can manage your cloud infrastructure from anywhere worldwide, provided you have a laptop and internet.

In simple words, the control panel for all the cloud infrastructure is provided with the help of a Graphical User Interface (GUI) using a website. The website can also be accessed publicly over the internet from anywhere.

Live migration

Live migration is a vital service. It reduces the risk of bigger downtimes due to system update rollouts. For example, if you have a peak time for your business and suddenly your host machine restarts or informs you to install the critical update, you will go nuts.

Google Cloud provides live migration for the Compute Engine, where the server is shifted to other compute machines in case any security updates or maintenance needs to be carried out by Google Cloud on the host machine.

Virtualized portability

In our day-to-day life, we change different service providers. For example, we continually upgrade to the latest phones. If the X brand offers more features and provides more value for the money, you immediately switch to the X-brand.

The same goes for cloud providers. Migration between different cloud providers is simple. Whenever you feel like switching to a cloud provider, you can change it. But usually, every operator looks for long-term relationships. Hence they will make you feel privileged so that you stay. Sometimes, they also provide huge discounts and minimize any unexpected costs.

Measured service

Cloud services can be measured and billed according to usage. The “pay as you go” model is based on this characteristic. Organizations pay for resource usage only and not for the whole service. The cost of the service can also be estimated using the price calculator, which most clouds provide free of charge.

We can also set a quota limit or spending limit on services and set reminders once the quota reaches near full utilization. Of course, crossing the quota limits does not stop the services we are running, but knowing how much money is being spent to make an informed decision and stop unnecessary services upfront is good.

High availability

Cloud services are highly available because so many users use the same infrastructure in a logical separation. Providers use the “multi-tenancy” model, which means multiple users have access to the same physical infrastructure with logical partition.

Multi-tenancy is the reason providers provide robust infrastructure. There are Service Level Agreements (SLAs) for the downtime of every service they provide, which is almost 99.9% for every service. Additionally, clients are compensated for any breach in the SLA. This is good because clients get paid if the infrastructure is down due to unforeseen circumstances. However, the clauses may vary per user basis.

To provide high availability, cloud providers maintain enough “pool of resources” to accommodate any sudden increase in demand or any unexpected load experienced by a user.

Let’s discuss how the network infrastructure is deployed on the cloud. There are different models as per the requirement. You’ll learn about each of them one by one.

Types of cloud deployment

Private cloud

Whenever an organization needs infrastructure that can be accessed within an organization only and no public exposure is required, a private cloud is a way to go for security purposes.

Let’s break this down with an example. Company X is a research company. It needs to host its website to track all its research studies. However, research data is highly confidential, and they don’t want to expose the website publicly. Furthermore, they want to restrict access to company employees only with a private network.

There are two ways to accomplish this. The first way is the traditional method: having their infrastructure set up for Intranet or private networks. Companies like Infosys, Wipro, and other old service-based companies use this approach.

The second way is using Virtual Private Cloud (VPC). In this case, they still have a private cloud, but it’s over the public Internet with the help of cloud providers.

In this case, a company will host its website with the help of a private cloud where no public IP is assigned to any resource. Employees will use a VPN to connect to the site.

At present, Google is the only company with a global private infrastructure; not a single part of their Google cloud infrastructure goes through the public Internet.

Public cloud

The public cloud is the most used type of cloud deployment. Suppose a company offers services that need to be publicly accessed; they need to host their services on the public cloud. Resources are assigned public IPs, which can be used over the Internet.

Examples of this deployment are amazon.com, google.com, and all websites we can access over the Internet. These websites are deployed under public clouds.

Public cloud services are more prone to attacks if proper security measures are not taken. Hackers target these websites to extract data or harm the owner’s reputation. So, security is a concern for publicly-hosted companies.

Hybrid cloud

In a hybrid cloud deployment model, companies deploy their company-specific internal network on a private cloud and public-facing services on the public cloud. Take the example of a research company again. The company hosts all its websites that track research data on the private cloud.

However, the company will host its website on a public server that is open to all. So, the server that hosts the company’s static brand website will be a part of the public cloud; however, all its internal websites that track research data will be hosted on the private cloud.

Multi cloud

Multi-cloud is the latest approach used by companies like Intel to provide better support for their clients. In this approach, customers are free to choose any cloud provider for a specific requirement and another cloud provider for different needs.

For example, Company X wants to opt for Google Cloud for better pricing for its CPU-intensive tasks and use AWS for long-term data storage or to host static files in different geographic locations. Combining resources from various cloud providers is called the “Multi-cloud” approach.