Today, cloud architecture takes center stage in most discussions relating to IT, whether we’re looking at Multi-Access Edge Computing (MEC) or the Internet of Things, considering IT for health care or another industry. As an essential part of modern IT infrastructure, it’s worthwhile to understand the roots of cloud computing and the role different cloud architectures play in today’s IT environment.

A Brief History of Cloud Computing

“Cloud computing” is not a new phenomenon. Its roots date back to the 1950’s when institutions and organizations shared CPU time. In fact, during my college years in the1970’s, we shared compute time with a main frame housed at SUNY Albany. This early time-share model formed the basis of today’s cloud concepts, allowing virtual terminals to run different routines and even different operating systems from shared nodes.

In the late 1990’s, the time-share model expanded to power large development projects for companies like Amazon, Microsoft and Google. By pooling resources using “private” cloud architecture, these companies leveraged proprietary operations and workloads that enabled them to scale efficiently and economically. Eventually, this architecture paved the way for a new model: public cloud.

By2010, the idea of a public cloud offering computing services through third-party providers over the public Internet started to make business sense. This model would enable businesses, and even individuals, to take advantage of the architecture tech giants had already used for years to reduce their investment in on-premises IT resources, efficiently scale to meet workload and user demands, and manage their costs by paying for only what they used.

Today, the public cloud gives thousands of businesses, developers, and individuals access to computing and storage resources that accelerate productivity. In fact, public cloud computing powers remote work and ecommerce, which we all have a greater appreciation for in the context of the COVID-19 pandemic.

In addition to public cloud, increased demand for computing resources and unique use cases have given rise to new cloud computing architectures.

Private Cloud: What’s Old is New Again

The increased demand for shared resources has inspired innovation in the use of cloud platforms. One of these innovations has made what’s old—private cloud systems—new again. In fact, today’s private clouds look a lot like their predecessors from the 1970’s. These networks serve a specific company or group of organizations. Typically designed to meet a particular infrastructure need or provide a solution for a particular use case, the private cloud is simply a system of shared resources that is hosted either on-premises or in a colocation data center and is not generally accessible to the public.

Use cases for private cloud include:

Low latency – In some cases, distance matters and tends to slow access to applications and compute power. When low latency is essential, private cloud networks offer an elegant solution.

Cost– When workloads are constant and intense, private cloud networks can help an organization realize significant cost savings and achieve efficiency.

Data governance – The EU has enacted GDPR, which limits the storage of sensitive data to the EU. California has recently passed a similar position with CCPA. To accommodate these governances, a private cloud can confine sensitive data to a specific geographic region.

Other considerations – Certain configurations may be easier and more efficient to manage in a private cloud. This is especially true if a workload requires a specific operating system that may not be available in a public cloud. In some cases, CPU or RAM requirements make a private cloud setting more appropriate.

Hybrid Cloud: The Best of Both Worlds

Another popular cloud architecture, hybrid cloud is characterized by workloads distributed across a public cloud and a private cloud and managed using a single access point. A hybrid cloud architecture might be used when running a workload is prescribed to operate in a private cloud, while development work is more effectively done in a public cloud environment.

There are several reasons to consider a hybrid cloud strategy:

Disaster recovery & redundancy – When a private cloud or public cloud is used for primary workload operations, it can easily be backed up using the counterpart for fail-over or redundancy. This is often more cost effective than other redundancy methods, like connecting two regionally disparate cloud platforms.

Data governance – With a growing need to confine sensitive data to specific regions, a hybrid cloud approach increases speed and efficiency. For example, a business can run public, non-sensitive operations in a public cloud while housing sensitive data in a private cloud housed in a specific region.

Network services routing – In many circumstances, network security prevents some end users from accessing a network. In these cases, a hybrid model gives all users access to a public network while secure operations remain available on the private network only.

Capacity – When a workload has a momentary or permanent capacity that isn’t available in a private cloud, a hybrid architecture allows some of that workload to be offloaded to a public cloud.

Development, testing, & production – A hybrid model allows for separate development, testing, and production environments. This makes it possible to stage development, execute roll-backs, and push code to production safely.

All of this brings us to the multi-cloud strategy.

Multi-Cloud: The Power of Multiple Public Cloud Services

Today, multi-cloud services are used by many network architects to accomplish their workload goals. In a multi-cloud model, multiple public cloud services from different vendors are combined to fulfill operational goals.

Although managing a multi-cloud strategy might seem complex, several platforms available today make it easier by providing a single access point for managing multiple cloud environments.

There are many use cases for multi-cloud architecture, including:

Platform efficiency – In many cases, certain platforms offer greater efficiency for specific tasks. Using a multi-cloud strategy, tasks can be distributed across platforms that offer the greatest efficiency and the data can be combined to achieve an overarching goal.

Redundancy – By hosting workloads on platforms located in different geographic areas, organizations can ensure data and operational redundancy.

Capacity limitations – When capacity in a private cloud is limited, part of the workload can be assigned to other cloud systems.

High availability architecture – For many businesses, downtime can be very costly. To ensure availability of data and services, these businesses may choose to operate high availability architectures by spreading their services across multiple providers.

Big data – Big data is not just a trend, but a significant drain on computing resources. In many cases, multiple public cloud services can be used to manage analytic clusters that require resources unavailable in a single cloud environment. These workloads can be distributed, and the results aggregated for use.

Development, testing, and production – Much like hybrid cloud models, multi-cloud strategies make it possible to separate development, testing, and production across platforms.

Other considerations – In some cases, it may make sense to divide workloads by performance, cost, and storage back-up requirements to the most appropriate cloud platform for each need.

Limited dependency – By distributing tasks across cloud services providers, a multi-cloud approach limits an organization’s dependency on a single provider.

Clustering – The nature of clustering requires multiple compute platforms, especially when critical data is being stored. Using a multi-cloud strategy, data can be shared and backed up even if a single platform or operational workload in the cluster fails.

These are just some of the use cases for private, hybrid, and multi-cloud models. Now it’s your turn, which model will your organization use to leverage cloud computing?  

‍