Unit 1: Computing paradigms, cloud fundamentals and migration
Cloud Computing notes · PTU syllabus (UGCA1936)
On this page
Unit summary
Cloud computing delivers computing resources such as servers, storage, databases and software over the internet, on demand, and you pay only for what you use. This unit traces how computing evolved from grid and cluster computing to the cloud, explains the cloud service and deployment models, compares cloud with cluster and grid computing, and shows how an organisation moves its applications into a cloud.
After this unit you can
- Compare grid, cluster, distributed, utility and cloud computing
- Define the cloud and explain its delivery models, deployment models, characteristics, benefits and challenges
- Distinguish cloud computing from cluster and grid computing
- Describe the broad approaches to migration and the seven-step model of migrating into a cloud
PTU syllabus topics
- Recent trends in computing: grid
- cluster
- distributed
- utility and cloud computing
- vision and definition of the cloud
- delivery models
- deployment models
- characteristics
- benefits and challenges
- cloud vs cluster vs grid computing
- migrating into a cloud
- broad approaches and the seven-step model of migration
Topic 1
Recent trends in computing
Computing has moved from one big computer shared by many users to resources that are rented like electricity. Each step below solved a limit of the one before.
Distributed computing
Many networked computers cooperate on one task and appear as a single system
Web search back ends, online banking
Cluster computing
A group of similar, tightly coupled computers in one location acting as one machine
A college server room running a batch job
Grid computing
Loosely coupled, geographically spread computers of different owners share their idle power
Scientific projects such as protein folding
Utility computing
Computing is metered and billed like water or electricity
Pay per hour of server time
Cloud computing
Utility computing delivered over the internet with self-service, elasticity and virtualised resources
Gmail, Google Drive, AWS
Example
From grid to cloud: a research lab joins a grid and borrows idle computers from other universities to run a simulation, but must wait for them to be free. Renting the same power from a cloud provider gives it in minutes, billed by the hour, with no waiting and no machines to maintain.
Topic 2
Defining a cloud
The NIST definition says cloud computing is a model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (networks, servers, storage, applications and services) that can be rapidly provisioned and released with minimal management effort. The vision of cloud computing is that computing becomes a utility: users do not buy or manage hardware and software, but consume them as a service, anywhere, on any device.
- On-demand self-service
- Provision resources yourself without calling the provider
- Broad network access
- Use services from any device over the network
- Resource pooling
- The provider serves many customers from a shared pool
- Rapid elasticity
- Scale up and down quickly as demand changes
- Measured service
- Usage is metered, so you pay only for what you use
Topic 3
Cloud delivery models
The delivery (service) models describe what you get.
- SaaS (Software as a Service)
Finished applications used through a browser: Gmail, Google Docs, Salesforce
- PaaS (Platform as a Service)
A platform to build and run applications without managing servers: Google App Engine, Azure App Service
- IaaS (Infrastructure as a Service)
Virtual machines, storage and networks you configure yourself: Amazon EC2, Azure Virtual Machines
- IaaS: you manage the operating system, middleware and applications; the provider manages the hardware.
- PaaS: you manage only your application and data; the provider manages the runtime and servers.
- SaaS: the provider manages everything; you simply use the software.
Exam tip
A quick memory aid: the higher the layer (IaaS → PaaS → SaaS) the less you manage and the less control you have.
Topic 4
Cloud deployment models
Public cloud
Many customers share the provider's infrastructure over the internet
Low cost and highly scalable, but less control
Private cloud
One organisation, hosted on-premises or by a provider
More control and security, but higher cost
Community cloud
Several organisations with shared concerns, such as universities or hospitals
Shared cost, shared compliance rules
Hybrid cloud
A mix of private and public clouds connected together
Keep sensitive data private and use the public cloud for peaks
Topic 5
Benefits and challenges
- Benefits: no upfront hardware cost (capital expense becomes operating expense), fast deployment, elastic scaling, high availability, automatic updates, access from anywhere and easy disaster recovery.
- Challenges ahead: data security and privacy, dependence on internet connectivity, vendor lock-in, compliance and data-location rules, performance variation, and unpredictable bills without cost control.
Example
A start-up's choice: a new online-notes start-up expects 500 users in the first month and 50,000 after an exam. Buying servers for 50,000 would waste money for most of the year. On a public cloud it runs two small servers normally and adds twenty during exam week, paying only for those days.
Topic 6
Cloud vs cluster vs grid computing
| Aspect | Cluster | Grid | Cloud |
|---|---|---|---|
| Ownership | One organisation | Many organisations share | A provider, rented to many |
| Location | One site | Geographically spread | Provider data centres worldwide |
| Coupling | Tightly coupled | Loosely coupled | Virtualised pools |
| Resource management | Central scheduler | Distributed scheduling | Self-service, on demand |
| Billing | Fixed cost | Usually free or shared | Pay per use |
Topic 7
Migrating into a cloud
Migration means moving an application, its data and sometimes its processes from local servers to a cloud. It must be planned, because a badly moved application can be slower and costlier than before. Broad approaches to migration:
- Rehost (lift and shift): move the application as it is to virtual machines in the cloud. Fastest, but gains little from the cloud.
- Replatform: make small changes, for example moving the database to a managed database service.
- Refactor (re-architect): redesign the application to use cloud features such as auto scaling.
- Replace: drop the old application and adopt a ready SaaS product.
- Retain or retire: keep what is not worth moving and switch off what is no longer needed.
The seven-step model of migration into a cloud
- 1
Conduct migration assessments
Business case, costs, risks and readiness
- 2
Isolate the dependencies
Find what the application depends on, such as databases and other systems
- 3
Map the messaging and environment
Match each component to a cloud service
- 4
Re-architect and implement lost functionality
Rebuild features the cloud does not provide
- 5
Leverage cloud functionalities
Use auto scaling, storage and monitoring services
- 6
Test the migration
Check functions, performance and security
- 7
Iterate and optimise
Tune cost and performance continuously
Example
Moving a college portal: step 1 finds the portal is slow only during admissions, so cloud elasticity helps. Step 2 notes it depends on a local database and a fee gateway. Step 3 maps the web server to cloud virtual machines and the database to a managed database. Step 4 rebuilds a feature that used a local printer. Step 5 turns on auto scaling for admission season. Step 6 tests with 5,000 simulated users, and step 7 trims unused servers after the season.
Exam tip
Learn the seven steps as a list; the exam asks for them in order. The first step is assessment and the last is iterate and optimise.
Key terms
- Cloud computing
- On-demand delivery of computing resources over the internet with pay-per-use billing
- SaaS
- Finished software delivered as a service
- PaaS
- A platform for building and running applications
- IaaS
- Virtual computing infrastructure rented from a provider
- Hybrid cloud
- A combination of private and public clouds
- Vendor lock-in
- Difficulty moving away from one provider
- Lift and shift
- Moving an application to the cloud without changing it
Quick revision
- Evolution: distributed → cluster → grid → utility → cloud.
- Five characteristics: on-demand self-service, broad network access, resource pooling, rapid elasticity, measured service.
- Delivery models: IaaS, PaaS, SaaS. Deployment models: public, private, community, hybrid.
- Cluster is in one site, grid is spread across owners, cloud is rented from a provider.
- Seven-step migration: assess, isolate, map, re-architect, leverage, test, iterate.
Important exam questions
Practice questions written to the PTU exam pattern for this unit's syllabus: short answers (Section A style) and long answers (Sections B and C style).
Short-answer questions
- Q1.Define cloud computing.
- Q2.What is utility computing?
- Q3.Name the three cloud delivery models.
- Q4.What is a hybrid cloud?
- Q5.List the essential characteristics of cloud computing.
- Q6.What is vendor lock-in?
- Q7.What is meant by lift and shift migration?
Long-answer questions
- Q1.Compare grid, cluster, distributed, utility and cloud computing.
- Q2.Explain the cloud delivery and deployment models with examples.
- Q3.Discuss the benefits and challenges of cloud computing.
- Q4.Differentiate cloud computing from cluster and grid computing.
- Q5.Explain the seven-step model of migration into a cloud.
Stuck on this unit?
Message SBS on WhatsApp for help with Cloud Computing, or to ask about studying BCA at Synetic.