Cloud computing is a model for empowering users to access a useful, network-accessible, shared pool of resources that allows for quick changes to settings and configurations with minimal intervention from a service provider in an easy-to-maneuver self-service management interface.
Our Managed Cloud service uses several virtual resources to create and drive a combination of components. These components work together to create our cloud environment. This framework meets the National Institute of Standards and Technology's1 (NIST) five basic cloud computing platform criteria:
- Networking Capabilities
- Resource Pooling
- On-demand Self-Service
- Rapid Elasticity
- Measured Service
Additionally, three major framework models typically employ these services. They include:
- Cloud Software-as-a-Service (SaaS)
- Cloud Platform-as-a-Service (PaaS)
- Cloud Infrastructure-as-a-Service (IaaS)
There are many other framework variants defined as XaaS or Anything-as-a-Service. The number of “as-a-Service” models has increased to the point where a growing list of over 51 service platform types is now available.
Furthermore, these services share numerous deployment modes, including:
- Private cloud
- Community cloud
- Public cloud
- Hybrid cloud
Other key technology drivers that propel this growth include:
- Ultrafast wide-area networks (WANs)
- Potent and economical servers
- High-performance, often open-source virtualization software
As we can see, the vast array and arrangement of cloud-based services continue to develop to meet an ever-widening assortment of more specific use cases. This trend only continues to gain momentum as companies look to move entire sectors to cloud platform-based environments. We, however, will limit our focus of this article to the five basic tenets of cloud computing models.
Cloud computing platforms should have the capability to be accessed over a network utilizing a standard communication method. This method encompasses cell phones, tablets, desktop workstation computers, laptops, or other hardware devices. It does not, however, limit network communications to public clouds. It also applies to all other deployment models and frameworks.
Being able to reach our cloud platform is paramount. In some instances, microservice controllers can automatically, command and regulate entire container-orchestration systems autonomously. The ability for systems like Kubernetes to operate and communicate independently of human interplay only hastens the speed at which this technology is progressing.
Almost all public cloud computing systems are multi-tenant models by design. The meaning of tenant, in this case, refers to a user or customer. To specifically denote a system is multi-tenant is to say that it is engineered and crafted for multiple users to take advantage of. This is where the concept of resource pooling comes into play. Examples of multi-tenant software are:
The term multi-tenancy implies that a single cloud infrastructure platform supports multiple customers while maintaining data isolation. The ability of systems like these pool resources in a framework like SaaS enables users of a single application to employ millions of connections. At the same time, connected platforms share or pool resources to optimize performance.
By definition, all cloud computing platforms should have the ability to be managed by a user. This includes the potential to create, destroy, resize, or modify services on the platform. Well-rounded Managed Cloud services also allow users to add secondary supplemental add-ons to assist or augment existing services. These services can include disk storage, backups, security-related modules like firewalls or VPNs, network options like load balancers, and a host of other hardware-based alternatives.
Typically, these self-service capabilities are presented to the client through a graphical user interface (GUI). These interfaces allow a client to add services in an on-demand manner without utilizing programmatic interfaces like APIs or terminals. However, most major providers supply a method to use those options.
A cloud computing platform’s capabilities should have the means to provide both a pliable and scalable provisioning scheme. It should also have a method to release or renew the functionality of a system easily. This includes, in some cases, features that provide for the automatic expansion of the system proportionally demand. The just-in-time service method can save a sizable amount of capital expenditures over a year to almost cover the cost of a typical always-on type of system.
This elasticity is also needed to save on expensive cloud computing resources during periods of the lowest utilization.If a client knows their prime consumption times are from 8:00 am to 4:00 pm EST, on Tuesdays, Wednesdays, and Thursdays, they can spin up additional servers during those times and once the rush passes, decommission any extra servers to accommodate the reduced consumption.
In comparison, if a baseline of X number of servers is required during the week, and demand is significantly lower on the weekend, servers can be decommissioned over the weekend to reduce costs. The capabilities for provisioning can appear unlimited and should be apportionable in any quantity at any time. However, hardware limitations of the underlying systems on some vendors can come into play, limiting the availability of possible scaling. Users should always inquire about the capability or limitations of the underlying system to ensure that a true cloud platform exists.
Often, cloud computing resource usage is measured or metered, allowing companies to only pay for what they have used. This type of gauged resource utilization is optimized by leveraging only services which apportion capabilities on a charge-per-use payment scheme, implying that services that employ this type of service will monitor, measure, report, and charge based on the computing time used. Some services break this payment scheme into chargers per hour or minute based on the actual consumption of services used.
The costing model is primarily the preferred method for apportioning services, especially if the provisioning mechanism allows for a just-in-time or automated provisioning scheme. This provides a secondary level of transparency in pricing, especially when margins are low.
While these criteria are beneficial to clients in many situations, there do exist cases where the use of a cloud computing platform is detrimental or prohibitive in practice.
Shared HIPAA Responsibility
For example, if a client wishes to use a cloud computing platform to store HIPAA-related data2, the cloud storage provider would be held accountable the same as the client is. They would then both need to abide by the same laws and safeguards as the client alone would. In this case, the cloud provider is also a business associate under HIPAA regulations, the same as the covered health provider or entity. Typically, cloud providers only provide the physical safety and infrastructure needs to be set forth in their terms of service.
Lack of Control
Another aspect a client should be aware of is that they are no longer entirely controlling their service. When a service is contracted for from a host or provider, the client essentially hands over all of their data to a third party. While safeguards exist, this loss of control can be detrimental in some circumstances.
Variable Provider Options
Other shortcomings can include limited feature sets and cost structures from cloud providers. Varying levels of support, options, and services are provided as each carry its own service tiers. All of this should be reviewed when researching cloud providers.
Additional Hidden Costs
The client should also review the cost of add-ons and extras when choosing a new provider. The original (lower) rate may skyrocket with the extras needed to accomplish the service’s task, including bandwidth or customization charges.
Lack of Backups/Disaster Recovery Options
Cloud platforms are not a viable backup methodology. While disaster recovery should always be a part of every client’s strategy, selecting a cloud-based platform does not automatically assume or imply that backups are a component of every service. Backups and disaster recovery options should be one of the key functions of every cloud-based strategy.
Lastly, we look at management capabilities. If choosing a cloud-based platform because of an autoscaling function, which upon initial review will save a significant amount yearly over traditional server-based hosting, you should ensure that the client has the infrastructure and staffing power in place to support that feature.The questions below should be analyzed and addressed before purchasing a package from any cloud platform provider:
- Does the autoscaling option need to be configured by you (the client) or does the cloud host address that issue?
- If they do not provide support for autoscaling, do you have the capability to absorb the cost of hiring a full-time cloud-certified administrator to manage the service?
- If you plan on outsourcing the management of your cloud, have you considered the fees associated with using a third-party service provider?
- Is there an estimated minimum cost per month for that service?
Utilizing a cloud platform carries many advantages for a business today. Accessing data from anywhere, adding services on the fly, having the ability to scale, or even scale automatically is a definite benefit.Liquid Web’s Managed Cloud offers a predictable pricing model and several pre-configured solutions to meet almost every need. If your firm needs a more advanced cloud platform deployment, our knowledgeable hosting advisors can assist in selecting or help you design the perfect scalable environment from the ground up.
As a bonus, our Cloud Servers are fully managed by Liquid Web's Most Helpful Humans in Hosting™, meaning you avoid the learning curve and complexities typically involved in managing a public cloud platform. We will walk you through the process of deploying your cloud solution and then manage it for you!
1, 2 "Final Version of NIST Cloud Computing Definition Published | NIST." 25 Oct. 2011, https://www.nist.gov/news-events/news/2011/10/final-version-nist-cloud-computing-definition-published. Accessed 19 Feb. 2021.
Our Sales and Support teams are available 24 hours by phone or e-mail to assist.