Cloud concepts Flashcards
Scalability
It’s possible to adjust resources per demand, e.g. if there’s an increase in traffic and the systems get overloaded, more resources can be added to manage this. The same goes if there’s a decrease in traffic and resources can be cut down.
Scalability can be
- Vertical: add/remove processing power and RAM to a VM, needs downtime to update
- Horizontal: Add/remove resources, like VMs, either automatically or manually
Reliability
The ability for a system to recover from failures and continue to function. A decentralized cloud allows for deployment in multiple regions and enable applications to automatically shift to another server or data center if the one in use goes down.
Predictability
Performance and costs are highly predictable, thanks to autoscaling and load balancing, and the ability to track and monitor resource use in real time.
Governance
The ability for deployed resources to meet corporate standards and government regulatory requirements. Mainly used for IaaS and SaaS, and is ensured through set templates. Depending on operating model, software patches and updates can be automatically applied.
Security
Managing security for OSs and software, like patches and maintenance. Manually for IaaS, and automatically for PaaS and SaaS.
Manageability
There are two types of manageability:
- Management OF the cloud: managing resources like automatic scaling, preconfigured templates, monitor health of resource deployment, receive automatic alerts.
- Management IN the cloud: manage cloud environments and resources through a web portal, CLI, APIs or PowerShell.
Infrastructure as a Service (IaaS)
A cloud service model that only provides hardware, network connectivity, and physical security - it’s like renting space in a datacenter. Everything else, like OS installation, configuration, maintenance, network configuration, database, and storage configuration, software updates and patches, is up to the user.
Useful for:
- Lift-and-shift migration (create cloud resources similar to an on-premises datacenter)
- Testing and development (establish replicated configurations for development and test environments)
Platform as a Service (PaaS)
A cloud service model that provides everything in an IaaS (hardware, network, physical security), but also OSs, middleware, development tools, and BI services.
Useful for:
- Development frameworks (cloud-based applications)
- Analytics or BI (thanks to the included tools for analyzing and mining data)
Software as a Service (SaaS)
Complete cloud service model with a fully developed application (like e-mail, financial software, messaging applications, e.g. Office 365).
It’s the least flexible model, but the easiest to get up and running, since it requires the least amount of technical knowledge.
Shared Responsibility Model
Applies to all service models, and divides the responsibility between provider and user.
The provider is always responsibly for the physical aspects like datacenter, network, and hosts, and the user for information and data, devices allowed to connect, and accounts and identities.
Depending on which service model is used the responsibility shifts:
- IaaS: places most on consumer (besides the physical), like software, networks etc.
- PaaS: shared evenly
- SaaS: most on provider (except information, devices, and accounts or identities)
Cloud models
There are different cloud models:
- Private cloud: either on-premises or an offsite datacenter, used by a single entity (company or organization). Provides greater control, but costs more (hardware, networks etc) and lacks some benefits that are provided in a public cloud.
- Public cloud: built, controlled, and maintained by a third-party cloud provider (like Microsoft, Amazon, or Google), whose services anyone can use.
- Hybrid cloud: an interconnected environment that uses both private and public clouds. Can be used to provide increased and temporary resources to a private cloud, an extra layer of security, or just share a flexible infrastructure with mixes resources.
- Multi-cloud: Using multiple cloud providers (like Azure, AWS, and GCP), where resources and security in more than one environment can be managed.
Consumption-based model
There are two types of expenses to consider when comparing IT infrastructures:
- Capital expenditure (CapEx): one-time, up-front costs to purchase or secure tangible resources (hardware, property etc)
- Operational expenditure (OpEx): spending money on services or products over time. This is the type of expenditure cloud computing falls under, since you only pay for the resources you use (no tangible assets).
Resource
The basic building block in clouds and is anything you create, provision, or deploy: VMs, VNs, DBs, services.
Resource groups
A group of resources. Can contain multiple resources, but a resource can only belong to one group. Can’t be nestled (i.e. can’t put a group in another group). It’s possible to apply actions, which will be applied to all resources in that resource group.
Subscription
A unit of management, billing, and scale. Logically similar to resource groups (in which you logically organize resources), since you can organize resource groups.
A subscription is needed to use Azure, and is linked to an Azure account, which provides authenticated and authorized access to products and services.
An account can have multiple subscriptions, with different billing models, access-management policies, defined boundaries.
