Service Models and Cloud Engineering

By Sarah M. Goulet

Although the "Cloud" generally refers to the umbrella term of remote data storage, it should come as no surprise that there are several different subtypes of Cloud computing based on usage and need.  These are respectively known as the Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service (SaaS) models.  All of these models share the same basic benefits: very low infrastructure investment, efficient resource utilization and speed, clearly defined storage capacities, high reliability (especially since these models prevent data loss), and global access (Gould).  Naturally, each flavor of cloud computing carries its own potential, although these services can (and do) overlap.

Infrastructure-as-a-Service (IaaS)

Infrastructure-as-a-Service models are best for companies and users that need to outsource computing and data storage.  For companies that need high amounts of computing power, don't have their own data center, or are heavily involved in data management, IaaS computing models tend to be the way to go.  Instead of having to pay high amounts to acquire all of the necessary hardware, companies that use this model only need to pay for what they use - which, as you might imagine, can cut costs drastically.  In short, this is a hosting service, and it's closest to the process of virtualization (Benzina 124) that we discussed in the History of the Cloud post.  It's highly affordable and accessible.  On the downside, however, the vendor has control of the programs you can access, and they tend to be slightly slower because they rely solely on internet access, thus using a lot of bandwidth.  This model tends to be mostly used by businesses and related enterprises (Benzina 127).

Image by Michal Jarmoluk from Pixabay 

Popular examples of IaaS include:

• Amazon Web Services,*
• DigitalOcean,
• Microsoft Azure,
• IBM SmartCloud, 
• Apache CloudStack (open-source).

*Strictly speaking, AWS can function as all three models.  

Platform-as-a-Service (PaaS)

Platform-as-a-Service models are ideal for companies and users that are involved with developmental processes, especially in terms of creating applications for both the web and mobile devices.  As with the other service models, PaaS offers plenty of storage and access to resources, but users also have access to a wide range of pre-built tools (Barabas).  This provides flexibility for developers, who can test, prototype, and release their applications more quickly with less worry about infrastructure.  This also allows teams of developers to work remotely!  On the flip side, however, switching vendors is very difficult because the services tend to be challenging to migrate, and elements of a company's infrastructure that are not cloud-compatible may need to be switched to get the full effect of the service (Paranet).  Nonetheless, the rapid developmental potential of PaaS services are favored by many.  

Some examples of PaaS services include:

• Amazon Elastic Beanstalk, 
• App42 PaaS,
• AppAgile,
• Apprenda.

Software-as-a-Service (SaaS)

This final deployment model takes ideas from the other two models and combines them; it is, as the name implies, a host for software programs that are accessed via the web. These programs and applications are available on demand and charge based on a subscription (although some services may be free) (Gould).  The service, as with the IaaS model, can scale to need and doesn't require the user to actively maintain and update the software, as this is automatically done by the provider.  This allows for extremely fast innovation; there is no waiting period to install or re-install software, since you can start using the services almost immediately from anywhere.  Moreover, often these kinds of services can be run directly from a web browser without any need to directly download applications (Benzina 125).  It's the simplest to deploy of the three models (and the most flexible), although they also tend to be the most costly and most difficult to monitor (Paranet).  Services that use this model are perhaps the most recognizable to the general user.  

Image by rawpixel from Pixabay 

Some examples of SaaS include:

• Salesforce,
• Microsoft,
• Adobe Creative Cloud,
• Google G Suite,
• Zendesk,
• Oracle,
• GitHub.

A Final Note

Applications of the Cloud are quite flexible these days, and the idea as a whole has moved in leaps and bounds.  It's quite possible that you may not even realize how much you benefit from Cloud services.  Streaming services such as Hulu and Netflix allow you access to hundreds of hours of video services, while services such as iTunes do the same with music.  Some of these services are even free, such as Google applications such as Docs and Drive.  Whether you're a student, a professional, or just a casual internet user, the ability to access applications and services remotely is one of the keystones of the Internet as we know it.  Who knows what technologies it may drive in the future!

Interested in learning more?  Check out this list of the top 21 Cloud Service Providers in 2019.

References: 

Barabas, J. (n.d.). An IBM perspective: IaaS vs. PaaS vs. SaaS.  Retrieved from https://www.ibm.com/cloud/learn/iaas-paas-saas.

Benzina, K. (2019). Cloud infrastructure-as-a-service as an essential facility: Market structure, competition, and the need for industry and regulatory solutions. Berkeley Technology Law Journal, 34(1), 119–142. Retrieved from https://doi.org/10.15779/Z38QV3C43D.

Gould, L. (May 2017). What's "cloud computing?" AD&P, 48-51. Retrieved from https://www.adandp.media/articles/whats-cloud-computing. 

Paranet. (October 2019). The three cloud computing service models. Retrieved from https://www.paranet.com/blog/bid/128267/the-three-types-of-cloud-computing-service-models.