A Primer on Cloud Computing

Colin Baird
6 min readAug 25, 2019

--

Cloud computing is defined as:

“a general term for anything that involves delivering hosted services over the Internet. These services are broadly divided into three categories: Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS) and Software-as-a-Service (SaaS). The name cloud computing was inspired by the cloud symbol that’s often used to represent the Internet in flowcharts and diagrams.”

Source

Practically speaking, it’s the Amazon’s, Microsoft’s, Google’s, and Alibaba’s of the world running warehouses full of powerful computers as efficiently as possible by leveraging their economies of scale, dividing them up into ‘virtual containers’ of varying sizes, and then offering them to others as a service for profit. If you’re thinking this sounds a lot like the concept of a rental business then you’re spot on, because that’s exactly what this is!

“Why buy… When you can rent?!”

Let’s assume for example that you own an e-commerce store that has a high level of variability in demand from your customers. Perhaps you see high amounts of traffic during daytime, and lower amounts in the evenings, or maybe it varies month to month, with surges around holiday seasons. Before ‘the cloud’ existed, your only choice would have been to purchase enough compute power to meet those peak loads. The issue here being that peak loads generally only last for brief periods of time, after which you may have a vast majority of your compute resources just sitting idle. Added to which, purchasing these physical machines incurs additional time and cost for setting up, securing and maintaining. As can be expected, the overhead for addressing scaling in this manner is high, and the sunk costs for your hardware leads to inefficiencies whereby you have a lowered pool of available capital to allocate to other areas of your business.

In cases such as this where workloads are variable, the cloud offers flexibility to both deploy and scale quickly at a fraction of the cost. This becomes especially true as advances in quantum computing continue to coalesce. It will be more economical to rent quantum compute power from a host provider far sooner than it will be to own one outright. It is for these reasons that cloud computing has become such a large industry in a relatively short period of time.

As noted in the definition at the beginning, there isn’t just one flavor of cloud computing however. The offerings are nuanced in a variety of ways so as to appeal to the broadest possible range of the market, from mom & pop’s online web store, to a medical insurance giant’s sensitive customer data.

Public Cloud represents a majority of the current market share and is, as it sounds, cloud hosting that is openly available to the public for use. Anyone from individuals to multi-national businesses can utilize these offerings. The catch here is that multiple customers could potentially be sharing the same physical hardware at the data-center level.

Private Cloud functions similarly to public, however the resources are dedicated and isolated per customer. Ranging from the hardware to sometimes even the physical placement of the servers in question. This solution is generally ideal for larger and more security conscious organizations.

Hybrid Cloud is an amalgamation of both public and private solutions. Some businesses may have a desire to operate their non-critical or sensitive operations to a cheaper public hosting solution, while also leveraging a private solution for say their sensitive customer information and proprietary data.

Infrastructure-as-a-Service (Iaas) is when you rent the computer architecture but provide everything else like the operating system, drivers and applications etc.

Platform-as-a-Service (PaaS) takes the rental model one step further, whereby you now only need to provide the applications and data, not the underlying operating systems etc.

Software-as-a-Service (SaaS) is when you rent the entire application stack so that you never have to worry about anything on the back-end, such as updates, patches, or server maintenance. Examples of this being Office365, Zoom, Slack, and Salesforce etc.

Source

As with anything, there are considerations one must weight before using these services as a solution to hopefully increase both organizational efficiency, as well as margins. A few of the primary pro’s and con’s to cloud hosting are:

Pros

  • Cheaper than buying expensive computers up front
  • Increased flexibility
  • Improved disaster recovery
  • Scalability

Cons

  • Requires constant (and stable) internet connection
  • Costs are potentially unbounded per the subscription model
  • Access and security concerns for highly sensitive data
  • Requires additional trust (aka added risk)

Given the list above, you might now have surmised (and rightfully so) that many government organizations do not often leverage public cloud solutions for their needs, due to the obvious concerns around security. The types of organizations that generally seek to pursue risk minimization are far less likely to deploy in a public cloud environment than say perhaps, the Uber’s and AirBnB’s of the world. Perhaps one day advances in encryption and distributed trust will be made that can scale to mitigate such risks.

Past, Present and Future

Cloud computing hit the mainstream with Amazon’s launch of AWS in 2006. Since then, other technology firms have followed Amazon’s lead and begun offering hosting services to consumers. At the time of this writing, Amazon currently dominates at 33% market share, followed more distantly by Microsoft at 18%, and Google at 8%. This is a great real-world example of the powers of the first-mover advantage.

Source: Statista

Over the past 13-years, since Amazon lead the charge to mainstream adoption for cloud computing with AWS, many businesses have primarily transitioned through three phases. Initially, not knowing what cloud hosting was, then gradually learning of the concept and advantages, and now finally, slowly migrating some, if not all business-facing operations to hosted environments.

If we are to extrapolate forward over the next decade, as the question transitions from…

‘How do I migrate my infrastructure and services to a hosted solution?’… to — ‘What else can I do with this environment to further optimize the use of my data?’…

We begin to see that the focus may at some point shift more towards a focus on the data-centric tools and services that specific hosting solutions might provide. This is where functions like machine learning, data analysis, and artificial intelligence are likely to begin to see a more rapid uptake as they are increasingly put to use against real world information. This is a large part of the reason why, if I had to wager on any horse in the race towards market dominance, it would likely be Google’s cloud offering (GCP).

Although markedly late to the game, and without the bootstraps that aided Microsoft; namely, having a massive pre-existing software infrastructure base that they migrated from perpetual licensing to a SaaS model — ex. Office 365. Google has the advantage when it comes to training software models to interface with a nearly endless variety of data. From the more mundane applications like image, voice and text recognition, to the science fiction like augmented reality and self-driving cars, Google’s early foundations as the gate-keeper to the internet has given them a marked advantage in the realm where the software we build meets the data we are surrounded by.

One need only compare Google’s ‘Assistant’ to Apple’s ‘Siri’ and Amazon’s ‘Alexa’ to observe the differences in maturity. Google’s Assistant is far better equipped to handle human questions, partially because we humans also tend to ‘Google’ questions manually in a similar way to how we would ask them verbally. These years of ingesting more human interactions than say Amazon for example, are a big part of the reason why Google is better equipped to meet growing market demands for intelligent insights on a wide array of generalized data.

While we can never fully anticipate the changes that may come with time, it is clear that hosted computing is here to stay and we are only just beginning to scratch the surface of the benefits that it can provide.

--

--