Cloud computing is a computing model where data is stored on remote servers accessed from the Internet. This remote computing platform is known as a cloud. When users access this cloud, computing resources such as processing power, information and storage are handled offsite at data centers that a user’s local machine or device connects to. These data centers are often massive in size and scale, with an infrastructure capable of offering computing resources far beyond those that can be achieved on a local machine for a comparable price. And instead of having to build a costly and extensive on-premises, hardware-based storage solution, organizations can merely pay for the amount of cloud storage they need and just pay a flat monthly fee. This flexibility in pricing and storage space is a major incentive for many small-to-medium-sized businesses (SMBs) to move to the cloud.
From the cloud provider’s perspective, there is no cap on the size of cloud storage, allowing it to scale indefinitely—provided storage capacity hardware is available. In lieu of adding space, data can also be moved, replicated and stored across multiples zones, allowing for greater protection against data loss in a way that could not be achieved with physical storage.
Cloud computing platforms can differ greatly; the most common implementations are public and private clouds, all of which will be discussed further in this article. Regardless of which cloud is being used, the key advantages include simplified scalability, near-instant provisioning, virtualized processing resources and speedy server base expansion.
While cloud computing does have staggering advantages that are evolving the ways we interact with data, there are some tradeoffs as well. At the present time, low-latency connections via cloud-based platforms are not comparable to on-premises systems; however, the need for low-latency is niche and typically not a factor in adoption decisions. Another thing to consider is that pricing, terms of service and security standards are not standardized across all companies, so the consumer must be meticulous in their decision-making process to ensure their cloud provider meets their needs.
Key Terms and Definitions
The Cloud—A type of computing that is not done locally on your desktop computer, servers, or small devices (phones, tablets, etc.).
Public Cloud—A multi-tenant cloud computing platform delivered by a service provider; makes applications, storage and other resources available to the general public over the Internet.
Private Cloud—A single-tenant, configurable cloud platform, typically maintained in-house, where hardware, storage and network resources are reserved for a single organization.
On-Premises Storage—A non-cloud computing environment that runs on computers on the physical location of the person or organization using that software.
Hybrid—A computing environment that uses both on-prem and cloud platforms.
Virtualization—More specifically, hardware virtualization, refers to the process of interacting with a host computer using another desktop computer or a mobile device by means of a network connection, such as through the Internet. Host computers are typically server computers, capable of hosting multiple independent virtual machines simultaneously for multiple users. Virtualization can take many forms, including memory virtualization, storage virtualization, data virtualization and much more.
Latency—The time it takes for electronic information to physically travel between two or more locations. All electric connections experience some form of latency, from the smallest nanocircuits to the largest bandwidth data networks.
Encryption—A complex algorithm that’s used to encode data for the purpose of privacy, security, and compliance concerns. To decode the encrypted files, a user needs the encryption key, and to crack encrypted data requires massive computer processing power.
Authentication Processes—Requirements in a network for a user to create a username and password. Multi-factor authentication is becoming more popular to reduce exposure to account compromise, by necessitating another level of authorization via a code sent to an additional device tied to that user.
Public clouds are based on the standard cloud computing model, where a service provider makes applications, storage and other resources available to the general public over the Internet. They are considered multi-tenant computing environments, meaning that the user purchases a “server slice” in an environment shared with other users (also known as clients or tenants). This environment is often geographically spread far and wide, allowing for multiple redundancy points for user data, promoting a secure and stable environment for business continuity.
Public cloud services are either offered free of charge or on a pay-per-usage pricing model. They are typically inexpensive for the end user and easy to set up, because all hardware, bandwidth and development costs are paid for and maintained by the provider. Pricing is predictable, because users only pay for what they use, typically by the hour for computing resources used. And, as with all cloud platforms, public clouds are readily scalable, allowing for the spinning up and tearing down of virtualized computing resources as demanded by end-user need.
Currently, the largest public clouds are owned by technology giants such as Amazon, Google and IBM, and because of their sheer scale, private cloud platforms just cannot compete on price and resources, not to mention their spend on stability, security and development into advancing technology. These providers have top security experts in their employ, and consider security to be a core component of their cloud platform. Despite the open connotations of its name, the public cloud is anything but unsecure.
For MSPs, a public cloud can help them provide a more flexible pricing model because it is a cost-effective solution for their needs. Effectively, the MSP has the manpower, resources and storage of the public cloud vendor they choose to deploy.
It’s important to remember that despite skepticism in the media, many compliance requirements no longer require a private cloud. In fact, the largest public cloud providers are leading in terms of healthcare and financial cloud adoption, assuring compliance regulations and security are maintained for these respective verticals to the highest degree.
A private cloud is a model of cloud computing that has IT services available over an IT infrastructure entirely dedicated for the sole use of one organization. Private clouds are typically managed by an organization’s internal team. Instead of a server slice, they are single-tenant environments, and all resources are dedicated and available to the organization that owns that cloud. This is achieved by implementing in-house storage servers accessed via cloud computing technology running on infrastructure within an internal data center. Private clouds offer similar scalability as public clouds, however the costs for infrastructure must be factored in. Additionally, a private cloud can be configured to maximize performance in-house as well. Private clouds can often use older technology than public cloud; technology that, once in place, is not as frequently updated.
Privacy, security and compliance mandates are main drivers of private clouds, because many companies believe these can be configured and maintained in-house at a more secure level than a public cloud. However this adherence comes at a price as well, and significant resources, development, and labor must be invested to keep a private cloud on par and/or more secure than the current public cloud landscape. Despite these factors, many internal and external compliance controls across the healthcare, financial, and retail industries demand private clouds for data storage.
On-Premise Platforms & Hybrid Solutions
Instead of using resources at a server farm or in the cloud, on-premises software, often abbreviated as "on-prem," is a computing environment that is installed and runs on computers on the physical location of the person or organization using that software.
On-premises platforms are well established, well understood, and very costly, but they are fast. If you are looking for video editing or transaction processing applications, the fastest speed will be through some version of an on-premises appliance, because they are able to shift vast amounts of data in a short amount of time, given the low-latency connection. Additionally, certain security regulations, particularly in Europe, may require organizations to keep their data in-house for the present time, which would necessitate on-prem platforms. However, regulations involving the cloud are changing rapidly, and in short time this may not be the case. On-premises typically have substantial costs upfront, and require additional expenses to maintain over time. The total cost of ownership, including software licenses, ongoing annual hardware and software maintenance, electricity bills and labor to manage the system can be exorbitant, and these systems can become outdated quickly, depreciating over time. As such, the resources, security and functionality of an on-prem system can lag far behind that of a cloud platform if proper funding, expertise and personnel are not secured.
Some organizations use a hybrid approach to platforms, utilizing on-prem and cloud solutions to achieve distinct goals. For example, active and private data may be kept on location, while inactive and older data will be moved to the cloud, where cheaper, more expansive storage is available.