It is important to understand the differences between the service models, in order to figure out how the available software or cloud services work at their fundamentals and who will utilize them, and for what purposes.
Basic Service Models
There are four main service models:
1. Infrastructure as a Service (IaaS) : The Cloud Service Provider (Google) manages physicals servers, data storage, as well as networking between servers, and users. You will have access to compute and virtualization tools; for example BigQuery for storing data, Google Compute Engine (GCE) delivering configurable Virtual Machines (VMs), Google Networking Products for setting up remote connections.
2. Platform as a Service (PaaS): The Cloud Service Provider manages the infrastructure as well (including providing the OS ), and provides you the platform for your workflow. PaaS services from Google include Vertex AI, Google App Engine, which are intended for development.
3. Software as a Service (SaaS): The Cloud Service providers provides you the full software application, and you are the direct user of it, without having to manage anything. Examples here include Gmail, Chrome Web Browser, YouTube, Google Docs, etc.
4. On-Premise. This is definitely not recommended for startups aiming to get up to speed with AI quickly. This means that everything, including the physical servers are self-hosted. Having your own physical datacenter at your location means it is on-premise, as you are responsible for setting up every level of service. The benefit of on-premise is that you can full control of the dataflow and services you provide to keep all of it 100 % secured (assuming you know what you are doing). This is what typically government and non-profit organizations use to maintain their data locked and secure at specific locations. However, keeping things on-premise will be a cause for major headaches; including a big upfront cost in setting things up from the beginning, having to worry about keeping all the servers operational and the additional significant time overhead associated with getting up to speed. Also it is expensive and requires a wide range of skillsets, and hardware may need to be replaced/upgraded, adding to the costs.
Google provides all of these. The below table provides a description for each
|Service Models||Example Users||Services||Google Focuses On||You Focus On|
|IaaS||Cloud Administrators, Network Security Engineers||Virtual Machines (VMs), Networking Tools, BigQuery||Servers, Networking, Storage||Cloud Infrastructure|
|PaaS||Developers, Data Science and MLE practitioners||Vertex AI, Google APP Engine||Above + Cloud Infrastructure||Development|
|SaaS||Business Users, All org||Gmail, Chrome, Docs, Drive||Above + Development||Nothing|
|On-Premise||Datacenter Team||Physical servers||Nothing||Everything|
With this in mind, you know have equipped yourself with the background information necessary to understand the main google product offerings.