[Cloud Architect] 4. Introduction to Design for Cost, Performance, & Scalability

Getting to Faster and Cheaper

 Four Cloud Models

• Private Cloud: Organization owns, operates, and governs their cloud computing resources.

• Community Cloud: Cloud resources provided for organizations and community groups.

• Public Cloud: Owned by the government, academic institutions, or a business (e.g. Amazon Web Services, Microsoft Azure and Google Cloud Platform).
• Hybrid Cloud : Combines public cloud and private cloud to allow data and resources to be shared between them.

Key Points

• Moving to the cloud doesn’t guarantee that your application will be faster and cost less to operate. In fact, without the proper baseline metrics, you won’t know where to start when planning your move.

• Most performance issues can be traced to the application itself and some applications will need to be redesigned to be cloud native in order to achieve truly optimized performance. Common causes of performance issues:

  • Poorly designed applications
  • Database design constraints
  • Inefficient network routes

• Moving to the cloud should improve network response times. Global edge cache locations can bring the server closer to the end user.

 

High Performance In the Cloud vs. High Performance Computing

Term	Definition
High Performance Computing (a.k.a. supercomputing)	Supercomputing harnesses the computing power of multiple computers and aggregates them to tackle complex problems in business and academia

Do more thing is much less time.

 

Key Points

• Supercomputing harnesses the computing power of multiple computers and aggregates them to tackle complex problems in business and academia.

• In supercomputing, computers are called nodes and nodes are organized into groups called clusters.

• Supercomputing is the ultimate in high performance computing in the cloud.

Additional Reading

Please follow the links below to learn more about Amazon’s managed database services, their on premises hybrid cloud offering, and high performance computing in the AWS cloud.

• AWS RDS
• AWS Outposts Brings AWS Cloud Hardware on Premises
• HPC on AWS

 

Defining Goals

Key Points

• Planning to move to the cloud is a process that requires pragmatism and neutrality to make the best decisions for your technology.

• The team is comprised of directors, managers, and engineers from Infrastructure, Operations, and Software Engineering teams

• It is important to define your migrations goals and base your migration plans on your actual metrics. Metrics provide the information you need to make data-driven decisions.

Examples Of Cloud Migration Goals:

• We are migrating to the cloud to reduce our infrastructure costs by 25%

• We believe that by hosting our application in AWS, we will be able to deliver download speeds that are 60% faster than our baseline

• We are going to duplicate our infrastructure in the cloud and maintain our AWS account as a warm backup site for disaster recovery

• We are consolidating our data centers and moving to the cloud because our AWS account will provide one centralized view into our environment with more visibility into how our compute spend is being utilized.

Additional Reading

The links below offer additional insight into the practice of cloud migration, and will offer you further details into the scope, procedures, and goals of a cloud migration project.

• The 5 “W's of Cloud Migration
• Cloud Migration