☁️Virtualization & Cloud
VMs, containers and running in the cloud · 42 terms
Virtualization
8How one real computer becomes many virtual ones, and the hypervisors, images and vCPUs that make it work.
One real computer split into many fake ones. Software does the split.
The software that splits one machine into many. It keeps each part apart.
The kind of hypervisor that sits right on the bare machine. Nothing is under it.
The kind of hypervisor you run as an app. It sits on top of your own system.
A whole computer made in software. Lots of them can share one real box.
A virtual machine you rent from a cloud provider. You pay while it runs.
A saved disk that new virtual machines start from, all set up the same.
A slice of a real chip that a VM gets to use.
Cloud Basics
6What renting computing online means, and where in the world the rented machines sit.
Renting servers and space online. You pay just for what you use.
A place in the world where a cloud firm runs a group of data centres.
A part of a region with its own power and network. It fails on its own.
Servers you own and run in your own building, not rented in the cloud.
More servers when busy, fewer when quiet. It happens on its own.
The cloud provider guards the cloud. You guard what you put in it.
Cloud Deployment Models
5Who owns the machines, who shares them and whose law reaches them: public, private, hybrid, multi-cloud and sovereign.
Cloud servers that anyone can rent. Lots of users share the same machines.
Cloud servers for one firm only. No one else shares them.
Your own servers and a public cloud, linked to work as one.
Using two or more cloud firms at the same time.
A cloud where the data, the staff and the law in charge all stay in one place.
Cloud Service Models
5How much of the stack you rent and how much you run yourself, from bare servers to finished apps.
Renting bare servers, disks and networks. You run the rest yourself.
A cloud service where you hand over your code and it runs it for you.
A finished app you use online. You often pay for it each month.
Run code with no servers to look after. Pay only while it runs.
Upload one small function. The cloud runs it each time something happens.
Cloud Providers
10The companies that rent out cloud computing, and what each of the big ones is known for.
A firm that rents out servers, space and tools over the web.
Hyperscalers
3The US giants with regions on every continent and hundreds of services each.
The cloud from Amazon. It is the biggest, with hundreds of tools.
The cloud from Microsoft. Firms that use Microsoft tools often pick it.
The cloud from Google. It is known for data, AI and Kubernetes.
European Cloud Providers
6Cloud firms based in Europe, often picked to keep data and contracts under EU law.
A French cloud provider, one of Europe's biggest. It rents servers and cloud at low prices.
A German web host and cloud provider, once called 1&1. Small firms use it a lot.
A German cloud from the group that owns Lidl. Its data stays in Germany and Austria.
A French cloud provider owned by iliad. It is known for renting GPUs for AI.
A German hosting and cloud provider, known for cheap servers. It was founded in 1997.
A cloud run by Deutsche Telekom. Its old name was Open Telekom Cloud. Data stays in Europe.
Cloud Cost Management
8Keeping the cloud bill in check: who spent what, who pays for it, and how teams cut waste and buy cheaper.
A way to run cloud spending as a team. Engineers, finance and managers share the bill.
Cloud Billing
3The cloud bill as data: how it is written down, and how it is split between the teams that ran up the costs.
A report that shows each team its share of the IT bill. No money changes hands.
Billing each team for the IT it uses. The cost comes out of the team's own budget.
One shared format for cloud bills. Bills from many clouds then line up, column by column.
Cost Optimization
4Ways to pay less for the same work: the right server size, and the discounts a cloud sells for a promise or for spare machines.
Matching each cloud server's size to the work it really does. Most end up smaller.
Cheaper ways to rent the same server than the normal hourly price, each with its own catch.
A promise to use one server type for one or three years. In return, it costs far less.
A promise to spend a set amount each hour for one or three years. In return, prices drop.
A spare cloud server sold very cheap. The cloud can take it back at short notice.