With more people becoming digital citizens, the ability for an application to explode in popularity has all but rendered obsolete the traditional IT hosting mindset of discrete servers performing discrete tasks. Instead, organizations are looking to gain greater flexibility in cloud computing.

The primary issue with traditional hosting methods is procuring and provisioning all the gear for the expected usage — if adoption were to suddenly climb, it would be difficult to bring online the necessary infrastructure to scale. This leads to the second problem: If app usage decreases, you have hosting gear that goes unused, resulting in wasted capital expense. This is where cloud computing changes the game for organizations, allowing them to rapidly scale for high usage and roll back infrastructure when no longer needed. Let’s dig into three hosting models that help organizations achieve cloud flexibility.

The Rise of the Data Team

Public cloud: Flexible computing power without the headaches

When talking about cloud computing, the first thing most systems engineers think of is the public cloud model. One way to conceptualize the public cloud is like a utility, similar to water or electricity. A cloud provider, such as Amazon, Microsoft, or Google, has a server farm partitioned into virtualized servers where computational power can be consumed on demand. There is no notion which physical device is powering the application, in the same way that we don’t know which generator produced a given watt of electricity that powers your home.

Treating computational power as a utility, technical users no longer have to concern themselves with the exact device powering their application. Instead, they have the ability to programmatically create or destroy computational resources on demand by incorporating the cloud provider’s API into their app. This flexibility in cloud computing effectively lets an organization use exactly what it needs in one moment and eliminate it in the next. And with many companies interested in reducing their capital expenses, the public cloud converts that capital expense of procuring traditional hosting gear into an operational expense that recurs monthly, helping to free up their balance sheet.

While public cloud is most often considered multi-tenant (i.e., different groups using virtualized resources on a physical server), some providers offer a single-tenant option. This allows a cloud consumer to have the assurance that the only data on that gear is from their company. Public cloud also extends beyond computational power and includes cloud storage (object, block, and file) and cloud databases (relational and non-relational).

Private cloud: Privacy and power

Organizations that desire cloud flexibility need to abide by certain governmental, compliance, regulatory, and/or strategic requirements that preclude them from using the public cloud. This is the perfect use case for a private cloud, a collection of servers that are stood up either on- or off-premises and house the applications and data of a single organization.

While this may sound like a dedicated IT hosting environment from the ’90s, the more modern private cloud environment has two key differences. First, a single hosting device may be virtualized and provisioned for multiple uses throughout the organization — this helps increase the overall utilization of the hosting infrastructure, enabling the company to squeeze more value out of its capital expense. Second, private cloud technologies (such as Azure, VMware, or OpenStack) provide the same programmatic access via an API for an application to instantly provision and bring a resource online or spin it down after it is no longer needed. Gone are the days of calling up the IT department and waiting for them to provision gear for your particular project.

Hybrid cloud: Squaring the circle

For the ultimate flexibility in cloud computing while balancing compliance requirements, an organization may want to consider a hybrid cloud. This combination of a private and public cloud can be particularly helpful for companies that must adhere to stringent standards for certain parts of their online footprint but can take advantage of the public cloud in other areas.

Additionally, the hybrid cloud is ideal when an organization is willing to incur a capital expense on the hosting infrastructure for the standard, day-to-day traffic but wants to “rent” computational power in the public cloud for bursts of traffic. This is best seen in a seasonal e-commerce business: The company could build a private cloud for its standard online traffic, but during the middle of its season, it leverages the public cloud to serve that peak traffic.

In either instance, a company needs to be savvy in how it builds its application to take advantage of this hybrid cloud flexibility. Namely, it would have to ensure that the application was not monolithic, where everything was built as a single unit and could not be split up across hosting environments. Instead, the company would need to build the application by splitting off the different components into microservices, enabling certain parts of the application, such as the database, to be housed in separate places.

Cloud: The present and the future

Regardless of which cloud computing format a company adopts, cloud just offers more flexibility over a traditional hosting environment, full stop. This takes the form of maximizing utilization of computational resources via virtualized environments, having on-demand infrastructure that can be programmatically brought online as the application scales, and eliminating equipment that is no longer being used as traffic subsides.

Every company is becoming a data company, and the cloud is all about making data and the computing power to store and analyze that data available when and how a company or application needs it. Intelligence gleaned from these cloud sources can be seamlessly infused into in-house workflows or customer-facing apps to evolve businesses and unlock new revenue streams. The flexibility of the cloud has been so revolutionary that new companies are born entirely on the cloud — their data teams never touch an on-premises server, and all their code is hosted somewhere else. It’s a new world for businesses of all kinds. Today and tomorrow are all about the cloud.

Building Data Models to Power Independent Exploration

Tristan Nunnally is a product manager for code-based analytics at Sisense. He’s passionate about people, ideas, and technology, in that order.

Tags: | | |