Skip to content

Latest commit

 

History

History
79 lines (68 loc) · 4.87 KB

Cloud Computing.md

File metadata and controls

79 lines (68 loc) · 4.87 KB
Raw

Cloud Computing

Scalability in Cloud

  • Cloud Computing allows you to easily add and remove resources as you need them
  • There are two types of scaling:
    • Vertical scaling allows you to change the power of your existing instance. You could get a more powerful server for example
    • With horizontal scaling, you are simply increasing the number of instances you're using, sharing the workload across multiple devices

Cloud Services Models

  • Infrastructure as a Service (IaaS) - offers cloud infrastructure
  • Platform as a Service (PaaS) - offers infrastructure and software for application development
  • Software as a Service (SaaS) - offers ready-to-use applications in the cloud

Cloud Services Models

  • FaaS (Function as a Service) is a variation on SaaS, except that now you're only concerned with a single function, or part of the software
  • For example, you can run identity authentication, or payment transactions
  • FaaS often uses a billing model called "serverless"
  • This is slightly misleading because the computations are still run on a server - what it means is that you pay for the service rather than paying to rent servers
  • Serverless architecture is where managing the server and allocating resources is handled by the cloud

The Cloud Pyramid: Abstraction vs Control

Cloud Pyramid

Cloud Deployment Models

Private Cloud

  • The private cloud model is private by design and designated for exclusive use by its tenants
  • This means to access a private cloud, you need to connect to a network link, which means special network access needs to be set up by IT
  • An organization may hire a third party to host their private cloud infrastructure or host it themselves
  • Private cloud is different from on-premise in the following ways:
    • The infrastructure in this model adheres to cloud principles, meaning it uses virtualization that allows on-demand compute resources
    • Private cloud can also be located off-premises

Public Cloud

  • The cloud infrastructure is shared and open for use by the general public
  • The infrastructure is owned and managed by a cloud service provider, like AWS or Azure
  • Public clouds are Internet accessible, hence organizations can get started quickly with minimal investment

Hybrid Cloud

  • This is when an organization uses a combination of two or more distinct models
  • The different models interact with each other via a network link and can share data and services
  • It's more of a question of where data and services are physically stored
  • For example, you could store sensitive patient data on a private cloud for security reasons and use an application on the public cloud, like a business intelligence tool, to process it
  • Hybrid clouds are useful in the case of cloud bursting. This is when a private cloud is overwhelmed by demand and hits capacity
  • To avoid disruption of service to users, traffic is moved to a public cloud instance. This allows organizations to cost-effectively handle periodic spikes with pay-per-use pricing

Cloud Providers

1. AWS

  • AWS launched in 2006, two years before Google Cloud (2008) and four years before Microsoft Azure (2010); hence is the market leader
  • For Personal Cloud, AWS offers Amazon drive and Amazon photos
  • For Professional Cloud, AWS offers:
    • Simple Storage Service (S3) for file storage
    • Elastic Compute Cloud, (EC2) for computation
    • Relational Database Service (RDS) for databases
    • Redshift - analytics and data warehousing
    • Kinesis - real time data movement and analytics
    • SageMaker - predictive analytics and machine learning

2. MS Azure

  • For Personal Cloud, MS offers OneDrive
  • For Professional Cloud, Azure offers:
    • Azure Blob Storage for file storage
    • Azure Virtual Machines for computation
    • Azure SQL Database for databases
    • Data Lake Storage - store data before cleaning
    • Stream Analytics - real time analytics
    • Azure Machine Learning - train and deploy ML models

3. Google Cloud

  • For Personal Cloud, Google offers Google Drive and Google Photos
  • For Professional Cloud, Google offers:
    • Google Cloud Storage for file storage
    • Google Cloud Compute Engine for computation
    • Google Cloud SQL for databases
    • Big Query - data warehouse
    • Dataflow - batch and stream data processing
    • AutoML - ML model training and development
  • Google Cloud acknowledges that multicloud - which is a combination of different cloud provider services - is the future and that consumers don't want to be locked in
  • In 2019, Google launched Google Cloud Anthos, which lets you deploy and run hybrid multicloud solutions, combining on-premise servers and several cloud providers, all in one place