WEEK 5 — DIGITAL TRANSFORMATION

Cloud/Fog/Edge Computing

Agenda

  1. 1Introduce Cloud, Fog, and Edge computing
  2. 2Explain various cloud service models
  3. 3Case study involving the choice of service models
  4. 4Edge computing
  5. 5Why choose Edge over Cloud
  6. 6How Fog computing can bridge Edge and Cloud
  7. 7Conclusion
  8. 8Q&A
Dr. Davood WadiDigital Transformation
Three tiers of computing: at the bottom, edge devices (a smartwatch, a car and a traffic light) send data up to a fog router in the middle, which sends it up to the cloud on a framed screen at the topCloudFogEdge

CHAPTER 01

Cloud Computing

CASE STUDY

The Growing Pains of FreshBake Bakery

  • FreshBake Bakery is a small but rapidly expanding business with 10 physical stores in Vancouver.
  • They currently host their website, inventory system, and customer loyalty program on an aging on-premise server.
  • During peak holiday seasons, their website crashes due to traffic spikes, inventory updates lag, and IT maintenance costs are eating into profits.
  • The owner wants to scale operations, improve customer experience, and reduce IT headaches but lacks technical expertise.
  • Your group is a tech advisory team hired by FreshBake.
  • Analyze their challenges and propose solutions.
Ten bakery stores in an arc, each linked by a line to one aging on-premise server rack in the middle, which shows a red warning signOn-premise

Questions

  1. 1What are the key business outcomes negatively affected by the current digital infrastucture of the company?
  2. 2

    Propose a solution for the company outlining:

    1. 1.First steps
    2. 2.Transformation
    3. 3.Expected improvements in outcomes

DEFINITION

What is Cloud Computing?

Cloud computing delivers on-demand computing services:

  • storage,
  • software,
  • processing power

over the internet.

Instead of owning physical servers, companies rent resources from cloud providers like AWS, Microsoft Azure, or Google Cloud.

Top: a business owns a physical server, drawn in blue. Bottom: the same business links in red to three screens holding storage, software and processing power, rented over the internetOwnRent

CLOUD COMPUTING

Key Characteristics of Cloud Computing

1.Scalability

  • Resources can be scaled up or down based on demand.
  • E-commerce platforms handle traffic spikes during sales.

2.Cost-Effectiveness

  • Reduces capital expenses by eliminating the need for in-house hardware.
  • Pay only for what you use (“pay-as-you-go” model)

3.Accessibility

  • Services are accessible from anywhere with an internet connection.

4.Flexibility and Innovation

  • Supports rapid deployment of applications and services, fostering innovation.
SIMULATED. A year of weekly demand for an online store, with a spike at the holiday sales. On-premise: one fixed server capacity at 50 of 100. The site crashes in 4 weeks, and 32% of the capacity paid for sits idle.On-premiseCloudWeeks the site crashes4Paid for, unused32%SIMULATEDCapacityCrashesJanFebMarAprMayJunJulAugSepOctNovDecDemand
Exercise · Which one is it?

Each business below uses the cloud. Name the key characteristic of cloud computing that each case shows.

A ticket website adds servers for the hour a stadium concert goes on sale, then gives them back.

Which concept does this case show?

CHAPTER 02

Cloud Service Models

CASE STUDY

Airbnb

Scenario: Airbnb hosts millions of listings worldwide, requiring scalable infrastructure to handle fluctuating demand.

Solution: Airbnb uses AWS EC2 to:

  • Scale server capacity during peak travel seasons.
  • Reduce infrastructure costs by only paying for what they use.
  • Ensure high availability for its global user base.
Twelve monthly stacks of server units: three base units every month in black, and extra units in red added only in the summer and December peak travel seasonsPeak travel seasonsJFMAMJJASOND

SERVICE MODEL 1

Cloud Service Models

1.Infrastructure as a Service (IaaS):

  • Provides virtualized computing resources (e.g., servers, storage).
  • Example: AWS EC2.

Amazon Web Services (AWS) EC2

Amazon Elastic Compute Cloud (EC2) is a prominent IaaS offering that allows businesses to rent virtual servers for their computing needs.

Who runs each layer, by service model: On-premise: the business runs servers and storage, tools and platforms, and applications itself; IaaS: the provider supplies the bottom layer. Red blocks are rented from the provider; black blocks the business runs itself.Servers,storageTools andplatformsApplicationsOn-premiseYouYouYouIaaSProviderYouYou

CLASS DISCUSSION

Discussion Question

Imagine your group is advising a company considering migrating its on-premise IT infrastructure to an IaaS provider (e.g., AWS, Azure, Google Cloud). Each group will be assigned a specific industry:

healthcare,e-commerce,government,education.

  1. 1
    What are the top 3 factors (e.g., cost, scalability, security, compliance, latency) your assigned industry should prioritize when selecting an IaaS provider? Justify your choices.
  2. 2
    What unique challenges might this industry face in adopting IaaS, and how could they mitigate them?
  3. 3
    Draft a shortlist of best practices for ensuring a smooth transition to IaaS, tailored to your industry’s needs.
    • Each group will present their recommendations.
    • The class will then discuss:Are there universal IaaS adoption principles across industries, or do priorities fundamentally differ?
An on-premise server rack moves along a red arrow to the cloud on a screen; below, four industry badges: healthcare, e-commerce, government and education

CASE STUDY

Snapchat

An engineer in profile faces a screen of app code in red, with their back turned to two blue server racks behind them

Scenario: Snapchat needed a platform to support rapid user growth and enable its engineers to focus on app development (company's main focus) rather than server management.

Solution: Snapchat used Google App Engine to:

  • Deploy its app quickly without worrying about server provisioning.
  • Automatically scale resources as its user base expanded.
  • Simplify backend management with integrated tools and APIs.

SERVICE MODEL 2

Cloud Service Models

2.Platform as a Service (PaaS):

  • Offers tools and platforms for developers to build applications.
  • Google App Engine.

Google App Engine (GAE)

Google App Engine (GAE) is a PaaS that enables developers to build, deploy, and scale applications without managing the underlying infrastructure.

Who runs each layer, by service model: On-premise: the business runs servers and storage, tools and platforms, and applications itself; IaaS: the provider supplies the bottom layer; PaaS: the provider supplies the bottom two layers. Red blocks are rented from the provider; black blocks the business runs itself.Servers,storageTools andplatformsApplicationsOn-premiseYouYouYouIaaSProviderYouYouPaaSProviderProviderYou

SERVICE MODEL 3

Cloud Service Models

3.Software as a Service (SaaS):

  • Delivers fully functional applications over the internet.
    • Salesforce,
    • Microsoft Office 365.

Salesforce CRM

Salesforce is a leading SaaS platform offering Customer Relationship Management (CRM) tools that help businesses manage customer interactions, sales, and marketing activities.

Who runs each layer, by service model: On-premise: the business runs servers and storage, tools and platforms, and applications itself; IaaS: the provider supplies the bottom layer; PaaS: the provider supplies the bottom two layers; SaaS: the provider supplies the bottom three layers. Red blocks are rented from the provider; black blocks the business runs itself.Servers,storageTools andplatformsApplicationsOn-premiseYouYouYouIaaSProviderYouYouPaaSProviderProviderYouSaaSProviderProviderProvider
Exercise · Sort into groups

Sort each cloud service into the model it belongs to.

        Tap a card, then tap the group it belongs to.

        USE CASE

        Use Case: Coca-Cola Enterprises

        Scenario: Coca-Cola Enterprises needed a centralized solution to manage sales processes across its global workforce.

        Solution: By adopting Salesforce CRM, Coca-Cola was able to:

        • Provide sales teams with real-time access to customer data from any device.
        • Streamline customer engagement and improve collaboration across departments.
        • Analyze customer behavior and generate insights to optimize marketing strategies.
        A cloud on a screen at the top links to a phone, a laptop and a desktop screen below, each showing the same customer record in red

        CLOUD COMPUTING

        Benefits of Cloud Computing

        1.Agility and Speed

        • Companies can quickly adapt to changes and deploy solutions faster.

        2.Reduced IT Overhead

        • Outsourcing infrastructure management to cloud providers.

        Business Benefits of Cloud Computing

        3.Enhanced Collaboration

        • Everything is on the web
        • Teams can work together using shared tools and resources.
        • Develop using various devices and places

        4.Global Reach

        • Cloud data centers allow companies to operate globally with low latency.

        CLOUD COMPUTING

        Risks and Challenges

        1. Security Concerns:

          Data breaches and unauthorized access can occur if not managed properly.

        2. Dependence on Providers:

          Outages from providers can disrupt operations.

        3. Cost Management:

          Poorly managed usage can lead to unexpected expenses.

        Exercise · Which one is it?

        Each company below runs on the cloud. Name the risk or challenge each case shows.

        A storage folder left open to the public lets strangers download customer files.

        Which concept does this case show?

        CLASS DISCUSSION

        Discussion Question

        How has Netflix leveraged cloud computing to transform its global streaming service?

        Consider scalability, content delivery, and customer experience in your answer.

        A globe in the middle, with four screens around it, each playing a video

        CHAPTER 03

        Edge Computing

        CASE STUDY

        The Real-Time Dilemma at HealthGuard Hospital

        • HealthGuard Hospital uses IoT-enabled wearable devices to monitor patients’ vital signs (e.g., heart rate, oxygen levels) in real time.
        • Currently, all data from these devices is sent to a centralized cloud server for processing.
        • However, during emergencies, doctors experience delays in alerts, and network congestion during peak hours causes critical data to arrive too late.
        • The hospital also struggles with sky-high bandwidth costs and security concerns due to transmitting vast amounts of sensitive patient data.
        • They want to improve response times, reduce costs, and ensure reliable care without compromising patient privacy.
        A patient with a smartwatch on the left sends data on a long arrow up to the cloud on a screen; the alert travels back down to a doctor on the right, marked by a red hourglass for the delay

        CLASS DISCUSSION

        Discussion Question (Group Activity)

        • Your group is a tech innovation team tasked with advising HealthGuard.
        • Analyze their challenges and propose a solution to transform their operations.
        1. 1

          What are the key drawbacks of relying solely on a centralized cloud for real-time patient monitoring?

        2. 2

          Which 3 core benefits of edge computing (e.g., reduced latency, bandwidth optimization, data privacy, offline operation) would most directly address the hospital’s needs? Justify your choices.

        3. 3

          What practical steps should the hospital take to implement edge computing? Consider hardware/software needs, hybrid cloud-edge integration, and staff training.

        Two people in profile face each other, each with a speech bubble, in conversation

        DEFINITION

        What is Edge Computing?

        Edge computing processes data close to its source,

        • IoT devices
        • local servers

        instead of relying on centralized cloud data centers.

        • reduces latency,
        • minimizes bandwidth usage,
        • enables faster decision-making
        • for time-sensitive applications.
        SIMULATED. Network traffic at 30% of peak hours. Cloud lane: a smartwatch sends every reading over the network to the cloud before the alert sounds, a delay of 0.85 s. Edge lane: a chip beside the smartwatch decides on the spot, a delay of 0.06 s.Network trafficQuietPeak hoursSIMULATEDCloud0.85 sEdge0.06 s

        EDGE COMPUTING

        Edge Computing

        1.Proximity to Data Source

        • Data is processed at or near the device generating it, reducing delays.
        A car with a red processing chip right beside it; the cloud on a screen sits far off and unused

        2.Low Latency

        • Critical for real-time operations
          • autonomous vehicles
          • industrial automation.
        A car sends data a short way to a chip beside it, marked by a red lightning bolt for speed; the cloud sits far off

        3.Bandwidth Optimization

        • Less data is sent to the cloud,
        • saving costs
        • reducing internet congestion.
        Many data packets go from a car into the chip beside it, and only one red packet goes on up to the cloud

        4.Reliability

        • Systems can continue to function even with limited or no internet connectivity.
        A blue no-signal sign stands between the chip and the cloud, and the car keeps sending data to its red chip
        Exercise · Match up

        Match each characteristic of edge computing with the case that shows it.

        Characteristic

        Case

        Tap a term, then the example it goes with.

        USE CASE

        Edge Computing in Retail

        Company: Walmart

        Scenario: Walmart uses IoT sensors and edge computing in its stores to:

        • Monitor inventory in real-time.
        • Optimize energy usage for refrigeration and lighting systems.
        • Process customer data locally for personalized shopping experiences.
        A store at the top; beneath it a red processing chip in the store links to the shelves' packages, a snowflake for refrigeration, a light bulb for lighting and a shopper

        CHAPTER 04

        Fog Computing

        CLASS DISCUSSION

        Running a Factory

        • Imagine you are managing a factory.
        • It has dozens of machines operating simultaneously.
        • Each machine produces sensor data about its speed, components, and production
        • The data needs to be analyzed in real time to derive key business insights.

        Propose a computing model for the factory.

        Discuss the pros and cons of your model.

        A factory floor of twenty machines in four rows, each a gear sending out its own sensor signal

        DEFINITION

        Fog Computing

        Fog computing is a decentralized computing model that extends cloud computing by bringing data processing closer to the network's edge.

        It bridges the gap between cloud computing (centralized) and edge computing (localized), enabling more efficient and scalable data processing across distributed systems.

        Three machines at the edge on the left link to two red fog routers in the middle, and the routers link to the cloud on a screen on the rightEdge(localized)FogCloud(centralized)

        FOG COMPUTING

        Key Characteristics of Fog Computing

        1.Intermediate Layer

        • Sits between edge devices and the cloud, handling data locally but still connecting to the cloud for large-scale analysis.

        2.Distributed Processing

        • Data is processed across multiple fog nodes (e.g., routers, gateways) within the network (i.e. local network).

        3.Scalability

        • Can handle large volumes of data from multiple IoT devices without overwhelming the cloud.

        4.Supports Real-Time Applications

        • Offers low-latency processing for applications like smart cities and industrial IoT.
        SIMULATED. 24 IoT devices share 2 fog nodes. Sent straight to the cloud, their raw data is 24 units against a cloud capacity of 40; through the fog nodes, only 2.92 units reach the cloud.IoT devices24SIMULATEDCloud capacityStraight to the cloudThrough fog nodes
        Knowledge Check

        A city adds 5,000 parking sensors. Gateways in each neighbourhood process the sensors' readings first, so the city's cloud keeps up even though five times more devices are now sending data. Which characteristic of fog computing does this show best?

        USE CASE

        Fog Computing in Smart Cities

        Company: Cisco

        Scenario: Cisco enables smart cities to use fog computing for applications like:

        • Managing traffic lights based on real-time traffic conditions.
        • Processing surveillance camera data locally for faster threat detection.
        • Enhancing energy efficiency through real-time monitoring of utilities.
        A traffic light, a surveillance camera and a utility meter each send data down to one red fog router, which links on down to the cloud on a screen
        Exercise · Sort into groups

        Where is each piece of data processed? Sort each case into edge, fog or cloud.

              Tap a card, then tap the group it belongs to.

              CONCLUSION

              Conclusion

              KEY TAKEAWAYS

              1. 1Introduction to Cloud, Fog, and Edge computing
              2. 2Cloud service models (IaaS, PaaS, SaaS)
              3. 3IaaS in different industries
              4. 4Edge, Fog, and Cloud, and how to combine them effectively