Modern Business Technology: Data Centers Versus the Cloud | Greetly

Should your organization store your data in the cloud or company-managed data centers? A detailed review of the options.

Executive Summary

  • Data is rapidly moving from on-premise and company-owned data centers to the cloud.
  • The "cloud" is a broad term covering multiple segments of hosting and servicing data.
  • Security, speed, and cost are major considerations in determining whether your organization should store information in owned data centers or the cloud (or a combination).

Computing has changed a lot over the last 20 years, and nothing has changed as much as the move toward “the cloud.” Individuals and organizations have been migrating their software, their storage, and their infrastructure away from localized data centers to the cloud, though at vastly different rates.

What is a data center? What is the cloud? How are they different, and how do they work together?

Definitions in Debate

When reading about cloud computing, you are likely to encounter the words “cloud,” “data center,” “private cloud” and “public cloud.” It is incredibly important to recognize that different publications and organizations define these terms differently from one another.

While the definitions laid out in this article are fairly standard and accepted among tech folks, it is crucial to read all articles with a keen eye. Resources written and published inside organizations and even on trusted internet publications may use these words differently. Take the time to determine what the article means when using these terms.

What Is a Data Center?

Two employees reviewing a checklist in a data center

A data center is an actual infrastructure that houses computing power. A data center is comprised of:

  • Servers
  • Storage devices
  • Networking necessities
  • Cables to connect everything
  • Power (which may include backup generators and/or batteries)
  • Environmental controls (cooling devices and fans)
  • The facility itself house everything

Some people use the term data center to strictly mean on-premises computing, i.e company-owned computers that support a single organization’s technical and security needs. (Another, more descriptive and accurate term for this is “enterprise data center” but even this can be used in multiple ways, with some using it for large-scale or hyperscale facilities.)

The fact is, data centers support all of the world’s computing and inter-connectivity. Data centers can be solely owned and used by one company or owned by the mega-companies that serve organizations and individuals all over the world. Amazon Web Services advertises 22 regions of availability, with likely one or more data centers in each one.

To summarize: a data center is a collection of all the physical computing needs used to serve an organization via some form of network.

Types of Clouds: Public and Private

Employees at a tech company with a modern office

No, not cirrus and cumulus, etc. Yet if you thought the data center terminology is confusing, cloud computing definitions are shifting almost as much as the stratus clouds literally floating over our heads.

At its essence, cloud computing refers to accessing files, programs, and information from a source other than the computer’s hard drive. Many organizations use internal networks (intranets) that allow users to access a centralized digital storage and computing area. This on-premise, localized data center is technically a form of cloud computing, though it is not what people think of when they refer to the cloud.

The most accepted and widely-used definition of the cloud is the internet. Generally, when most people talk about cloud computing, they are referring to outsourcing a portion or all computing needs to a third party and accessing it via the internet as opposed to a closed network. (As a reminder, when an organization outsources their computing to the cloud, it is simply being shipped off to one or more data centers owned by other companies.)

To get even murkier, private and public clouds (and hybrid clouds and multi-clouds) get added to the mix.

Basically, a private cloud is utilized by a single entity, with appropriate security clearance needed to access it. There may also be some responsibility for maintaining the infrastructure. A company-owned, on-premises data center is not usually considered a private cloud, but if a company contracts for dedicated servers from a third party, that would be a private cloud.

A public cloud, on the other hand, is what most people use when they use Amazon Web Services. The same servers, software, and data centers are being used by multiple entities. A small business in New York could be using the same server as a blogger in Australia. While individual accounts and organizational data are guarded by passwords, the hardware and bandwidth are not dedicated to one organization.

Visitor registration app welcomes visitors to a high tech company's reception areaDifferent Types of Cloud Services

Outsourcing to the cloud comes in many forms. Aside from the public, private and hybrid clouds, there are different ways organizations can use the cloud for different parts of their IT process.

  • Infrastructure as a Service (IaaS): IaaS is sometimes also known as Hardware as a Service (HaaS). This service provides the computers, servers, and networking power on which companies can install and manage their own software. Microsoft’s Azure and Amazon Web Services are examples of IaaS, though they often also go into the next step of Platform as a Service. Think of this as a third party providing a data center, but the organization providing all the rest.
  • Platform as a Service (PaaS): This type of service offers both the hardware necessary for computing, but also the next level of basic needs, like operating systems and databases. They may also provide technical support, like collaboration, application design, and development, and security.
  • Software as a Service (SaaS): SaaS is a model by which program developers offer application licensing through a subscription model. The software itself is accessed via the internet instead of being installed on a local computer/hard drive. Examples of SaaS include Slack, Dropbox, and Greetly.

Even if a company determines that a local data center is the best way to go, SaaS may still offer a great way to outsource certain operations. Some companies feel more comfortable outsourcing email or word processing programs than crucial applications.

Local Data Center Versus the Cloud

It is not one-size-fits-all when it comes to choosing whether cloud computing or an on-premises data center is the right solution for an organization.

Cloud computing offers several benefits, some of which make the cloud particularly attractive for small companies, or those just starting out.

  • Affordable starting costs: There is no need to outlay large sums for hardware or software upfront.
  • Cost-effective: Large providers have an economy of scale advantage.
  • Easy to scale up or down: Unlimited storage is available. There is no need to purchase more servers if a company grows, or to be stuck with too many machines if it shrinks.
  • Security and safety: A good service provider will have expert security personnel taking care of cyber and physical threats to the information and infrastructure.
  • Backup in case of disaster: Having servers and data off-site adds an extra layer of backup protection. If a fire or flood destroys a local facility, having all the data and computing power off-site can allow an organization to get up and running again more quickly. Larger providers are also likely to have backups for the backups, so if something awful were to occur at one of their data centers, their customers would be covered.

On the other hand, there are still times when localized data centers are preferred by organizations. In spite of predictions, companies are still building and maintaining their own data centers.

Local data centers do have certain advantages over outsourcing to the cloud.

  • Customization: A local data center gives an organization full control over all their computing needs. It is more cost-prohibitive to purchase outright all the hardware required, but the company doesn’t have to worry about a third party using sub-standard equipment. Or having to share it with anyone else. For organizations running complex programs that require a lot of bandwidth and fast speeds, a localized data center might be preferable.
  • Security: Some organizations feel far less comfortable trusting third-party cloud computing providers with their data. Highly confidential data may simply not be as secure if it is being stored by someone else. It is also possible that the data is in multiple data centers, and the company may have no way of knowing how secure each one is. A local data center ensures that only those with company credentials can access the data.
    (The major caveat to this is that it requires a certain level of competency from IT. An organization’s data may be at higher risk locally if it does not have knowledgeable talent on staff.
  • Speed: Nothing is faster than simply reaching into your own backyard. If an organization is in a location with slow or spotty internet connections, cloud computing may just slow it down. Lightning-fast speeds are ensured if a data center is local.

Conclusion - What Is the Best Option for You?

Determining the best option for any given organization requires a lot of thought. Some questions to ask:

  • Is the money available to purchase what would be needed locally?
  • Is a lot of customization required?
  • What is the talent level of IT personnel to build, maintain and secure a local data center?
  • Will there be a need to scale up quickly?
  • What are the appropriate data backup plans?
  • Is reliable, fast internet available, and is it sufficient for the computing needs?

In the end, it might not be an “either/or” outcome. It might be best to outsource certain software needs, like email, or visitor management systems, but keep crucial software and hardware on-site. The number of different sizes and types of organizations requires many custom options, and one size is certainly not going to fit all. However, the possibilities for the perfect on-premises or outsourced model are out there waiting to be explored.

Similar posts