[Newbie] Instances, Containers and Swarms
 

Ok i am way over my head as not only am i a newbie to linux after reading up on cloud computing, the use of containers and container (docker) swarms i am still not 100% i understand fully or at least to a level i am satisfied with lol Feel free to *really* dumb it down as i said am new to linux and cloud computing (but willing to learn and take in new stuff/concepts!)Here are some of my questions

#1 Suppose you install an "instance" of an OS for example Ubuntu 20.04 on a cloud server

#2 How many clients can run that single "instance" Clients in theory dun need a hdd/ssd (no full blown OS) just a thumbdrive with a client app to log into this "instance" am i correct to say this?

#3 Are clients allowed some minor customisation like change to a desktop wallpaper of their choice? Also are apps/software strictly pre defined by the Cloud OS server i.e only a set/fixed set of apps/software are allowed to be run by the clients?

#4 How do the server account for different hardware on clients? Maybe some integrated GPUs others AMD/Nvidia discrete GPUs? Drivers are on clients in separate folders?

#5 Are all apps/software installed on the server and clients merely have a shortcut/quick link to said apps/software only?

#6 Suppose the server is strictly to serve fat clients : since it removes the need to hand over virtual cpu cores to them does it mean the cloud server does not have to be some supercomputer beast with massive cores/specs? what about RAM?

#7 How exactly does one "containerise" an OS on the cloud? Server has docker server daemon installed while client has docker client installed?

#8 Swarm : i've read about this and it says the advantage is "high level of availability offered for applications" What does this mean exactly? Apps/software run faster or the more worker nodes joining the swarm the more efficient it becomes?

Hey,

I took this course & one on Kubernetes at Udemy: https://www.udemy.com/course/docker-mastery/

You will learn quickly what it's all about.

The term "client" is very vague. You seem to think of a hardware device, but "client" is used often for pieces of software. What kind of clients do you have in mind?

You can run any software on the cloud server. You just have to install it. I don't quite understand, however, how the software on that server affects the clients, and how clients are customized by the server.
Normally, the clients' architecture doesn't affect the server. Clients usually access servers using architecture-independent protocols. You seem to have a very specific scenario in mind, though. What scenario is it?
The functionality of an app can be entirely on the client, without the need of a server, or entirely on the server, where the client only displays information and collects input, or anything between the two extremes.
I don't know exactly what you mean. Servers don't hand CPU cores to clients, they process clients' requests. The fatter the client, the less load on the server, obviously. Whether the server has to be a supercomputer depends on the application.
Docker is not normally used to containerise entire operating systems. It's main purpose is to containerise applications, web sites, databases etc. Containerising an OS would be quite an effort, I guess.
To use a containerised OS, just access it over the network. Docker client is not needed.
"high availability" doesn't mean "fast". It means "always available, even when parts of the hardware or software break down". Normally this is achieved with redundancy.
However I don't think availability is the main purpose of Swarm. Instead, Swarm allows you to deploy containers on several servers, running several applications. It's the equivalent of Kubernetes (and Kubernetes is the de-facto standard for container orchestration, whereas Swarm doesn't seem to have a very bright future).