Video summary

Sesión 4

Main summary

Key takeaways

Educational

Main ideas and concepts

Course session context / logistics

  • This is Session 4 of the course and the second session focused on the term and concept of cloud computing.
  • Two main topics are covered:
    1. Cloud service providers
    2. Current trends in cloud computing
  • Communication/interaction notes:
    • A QR code is shown for YouTube sessions (past, current, next).
    • The forum is for specific questions about the session.
    • There is no roll call in sessions (including comments).
    • Viewers do not need to be connected to the forum during the live session to avoid distractions (comments may be requested at moments).

Topic 1: Cloud service providers

Learning objectives

By the end of the talk, participants should be able to:

  • Identify the main cloud service providers
  • Understand the basic infrastructure of cloud providers
  • Explain the main characteristics of these platforms
  • Explain current trends in cloud computing
  • Relate trends to organizational needs

What cloud service providers are (core explanation)

  • Cloud providers are companies that build and operate large data centers globally.
  • They invest in:
    • Data center buildings
    • Thousands to hundreds of thousands of servers
    • High-speed networks
    • Storage systems
    • Energy backup
    • Specialized air conditioning / cooling
    • Physical and logical security mechanisms
  • Instead of organizations buying/maintaining their own hardware:
    • They rent resources on demand
    • They pay for what they use (time/capacity)
  • Key operational shift vs. traditional on-prem:
    • Traditional: buy servers → ship → install → configure OS/software → maintain → replace (can take weeks/months)
    • Cloud: create virtual servers in minutes, scale up/down, delete when not needed

Major providers mentioned (dominant market)

  • Amazon (AWS)
  • Microsoft (Azure)
  • Google (Google Cloud)
  • Also referenced: IBM, Oracle, Alibaba (plus others later)

How to choose a provider (decision factors)

The speaker emphasizes that choosing only the “cheapest” option is not sufficient. Organizations should analyze multiple factors.

Provider evaluation checklist

  1. Cost (not just monthly price)

    • Monthly fees
    • Cost to store data
    • Cost to transfer data over the internet
    • Cost to increase processing capacity
    • Cost to maintain infrastructure over several years
    • Example idea: a provider that seems cheap may become expensive as usage grows (e.g., from 10,000 to 1,000,000 customers).

  2. Safety / Security

    • Protect sensitive data (e.g., banking, hospital, universities) from:
      • Unauthorized access
      • Cyberattacks
      • Accidental loss
    • Mechanisms mentioned:
      • Encryption
      • Multi-factor authentication (e.g., passwords + face/voice)
      • Permanent monitoring
      • Compliance through international security certifications

  3. Availability

    • Services remain operational most of the time
    • Example: an online store failing during peak hours can cause major financial losses
    • Use SLA (Service Level Agreement) with high availability targets (e.g., 99.9% or 99.99%)

  4. Scalability

    • Ability to scale up/down based on demand
    • Example: traffic spikes during holidays/sales
    • Benefit: temporarily increase compute capacity and pay only for peak resources

  5. Geographic coverage

    • Data centers distributed across many countries/continents
    • Goals:
      • Bring services closer to users
      • Reduce response times and improve user experience
    • Reliability concept: if one region fails, providers can replicate to another region to maintain continuity

  6. Specialized services

    • Providers may excel in specific areas, such as:
      • Advanced AI tools
      • Database specialization (e.g., enterprise databases)
      • Strengths in specific financial applications

  7. Compatibility and technical support

    • Compatibility: if a company already uses a technology stack (e.g., Microsoft ecosystem), it should integrate easily
    • Technical support: often critical because cloud services must frequently be 100% available; immediate help is needed when problems occur

Provider profiles (representative examples and highlighted strengths)

1) AWS (Amazon Web Services)

  • Why it matters:
    • One of the first to offer cloud at massive commercial scale
  • Brief origin story:
    • Around ~2000, Amazon engineers faced slow infrastructure provisioning
    • This led to the on-demand infrastructure approach
    • AWS emerged in 2006
  • Current strengths mentioned:
    • Data centers across many regions
    • 200+ solutions/services
  • Representative services:
    • Amazon EC2 (Elastic Compute Cloud)
      • Create virtual machines in minutes (e.g., Linux, specified CPUs/RAM)
    • Amazon S3 (Simple Storage Service)
      • Store documents, photos, videos, backups, app data, static websites
      • Advantage emphasized: durability via multiple data copies
    • Amazon RDS (Relational Database Service)
      • Managed relational databases (provider handles maintenance, updates, backups)
    • AWS Lambda
      • Serverless concept: upload code; it executes automatically on events (developer doesn’t manage servers directly)

2) Microsoft Azure

  • Launch year: 2010
  • Growth reason:
    • Strong presence in private/government organizations via Windows, Office, SQL Server, Windows Server, authentication services
  • Strength described:
    • Organizations can extend familiar Microsoft technologies instead of starting from scratch
  • Representative services:
    • Azure Virtual Machines
    • Storage
      • Secure, scalable backup for documents/media
    • Microsoft Entra (identity management)
      • Authorization/authentication and user access controls
    • AI integrations:
      • Machine learning, image analysis, natural language processing, intelligent assistants

3) Google Cloud

  • Release year: 2008
  • Why it developed infrastructure:
    • Began as an internet search engine; the scale of searches/emails/videos drove infrastructure growth
  • Strengths mentioned:
    • Large-scale data analysis
      • BigQuery: analyze massive data quickly (example: supermarket sales patterns by season/region/branches)
    • AI
      • Research impact in voice recognition, translation, intelligent search
      • AI access via Gemini Enterprise (train/deploy models without building huge data centers)
  • Other notable product mentioned:
    • Google Cloud Storage
    • Kubernetes: deploy container-based applications across platforms

Other providers briefly profiled (additional strengths)

  • IBM Cloud
    • Focus: complex enterprise solutions
    • AI via Watson
    • Business AI: assistants, document analysis, task automation, decision support
  • Oracle Cloud
    • Strong background in database management systems
    • Easier migration of existing Oracle databases
    • Also offers storage, networking, AI, and data analytics
    • Differentiator: technology integration
  • Alibaba Cloud
    • Major in Asia-Pacific (e.g., China, Indonesia, Malaysia)
    • Origin: e-commerce infrastructure
  • DigitalOcean
    • Developer/startup-focused simplicity
    • Less overwhelming than feature-heavy platforms
  • UBH Cloud / UBH (as spoken)
    • European focus: distributed data centers and data sovereignty
    • Emphasis on compliance with European privacy regulations

Key conclusion from Topic 1

  • There is no universal “best” provider:
    • Choice depends on organizational needs, investments, current situation, and constraints.
  • The session hints at future trends, including selecting multiple providers later (multi-cloud).

Topic 2: Trends in cloud computing

What “trend” means (definition given)

  • A trend is not a passing fad.
  • It is a change driven by real needs and likely to evolve and persist over coming years.

Trend 1: AI as a Service (cloud + AI)

  • Motivation:
    • Training AI requires:
      • Enormous data
      • Significant processing power (e.g., thousands of GPUs over weeks)
    • Few organizations can build that infrastructure themselves
  • Core idea:
    • Providers offer AI as a service so organizations can rent infrastructure instead of building data centers
  • Example use cases:
    • Classifying documents (hospital: medical images)
    • Detecting fraudulent transactions (bank)
    • Building virtual assistants (university)
  • Top platforms named:
    • Microsoft Azure AI Foundry
    • Amazon Bedrock
    • Google Gemini Enterprise
  • Benefit emphasized:
    • Simpler interfaces help developers use AI without being deep ML experts
    • Supports democratization of AI, enabling smaller organizations and even individuals

Trend 2: Edge computing (processing closer to the user)

  • Problem it addresses:
    • Devices generate information continuously; sending everything to distant data centers can be impractical
  • Core idea:
    • Bring computation closer to where data is produced using edge servers
  • Why it matters:
    • Response time/latency can be critical
  • Examples:
    • Self-driving cars (fractions of a second for detected pedestrians)
    • Robot-assisted remote surgeries
    • Automated factories / robotics
    • Traffic systems / train-related systems
  • Edge tasks:
    • Analyze locally near the event/user
    • Send only necessary results to the cloud for storage, training, or heavier computation
  • Clarification:
    • Edge computing complements cloud computing
    • Cloud remains important for:
      • Storing large datasets
      • Training AI models
      • Running compute-heavy applications
  • Where edge servers may be deployed:
    • Inside devices
    • On antennas
    • At traffic lights
    • Within facilities/factories

Trend 3: Serverless computing (developer does not manage servers)

  • Motivation:
    • Even if the server exists “in the cloud,” someone must still:
      • Provision it
      • Patch OS/security updates
      • Monitor resources (energy, disk space)
      • Configure networking rules
    • This creates ongoing operational/admin overhead
  • Core idea:
    • Serverless means the provider manages infrastructure tasks:
      • Installing, configuring, updating, securing, scaling
    • Developers only:
      • Write code
      • Upload it
      • Let the platform run it when events occur
  • Clarification:
    • Servers still exist; what disappears is the developer’s need to manage them directly
  • Example scenario:
    • University ID photo upload service
    • Traditional: server runs 24/7, consuming resources even when idle
    • Serverless: code executes only when a photo is uploaded; the function stops after completion
  • Pay model emphasized:
    • Pay more efficiently for the time code runs, not idle uptime
  • Providers/platforms named for serverless:
    • AWS Lambda
    • Azure Functions
    • Google Cloud Functions
  • Limitation stressed:
    • Serverless is not always best for constant 24/7 workloads; virtual servers may be better
    • Best suited for smaller tasks, automation, and event processing

Trend 4: Multicloud and hybrid cloud (architectural strategy)

Multicloud

  • Why it exists:
    • Earlier strategy: choose one provider and move most services there
    • But relying entirely on one provider introduces risk:
      • Provider disruption/major outage
      • Vendor lock-in / dependency (migration becomes costly/complex)
  • Core idea:
    • Use multiple providers to leverage strengths
    • Example composition:
      • AWS for certain compute features (Lambda mentioned)
      • Google for AI model development
      • Azure for authentication/authorization
  • Benefits:
    • More flexibility
    • Reduced dependency on one platform
  • Challenges:
    • Need staff trained in multiple clouds
    • Integration complexity may increase despite standards/protocols

Hybrid cloud

  • Core idea:
    • Combine local infrastructure with public cloud services
  • Example given:
    • Hospital/bank/university keeps sensitive legal/regulated data (e.g., clinical info) on premises
    • Uses cloud for:
      • Administrative apps
      • Backups
      • Training AI models
  • Summary phrasing:
    • Today’s cloud is not just migration—it’s building intelligent architectures
    • Must maintain cloud benefits: availability, security, flexibility

Trend 5: Green cloud computing / digital sustainability

  • Issue addressed:
    • Cloud usage consumes electricity and generates heat in data centers
    • Cooling systems also use significant energy
  • Core goal:
    • Increase processing capacity without uncontrolled energy growth
    • Achieve better energy/environmental efficiency
  • Strategies mentioned:
    • Renewable energy
      • Use solar, wind, hydro to reduce emissions
    • More efficient hardware
      • Modern servers deliver more capacity with less energy
    • Improved cooling systems
      • Use outside air when conditions allow
      • Advanced liquid cooling
    • Virtualization to improve utilization
      • Previously, physical servers might run at ~20% utilization but still consume similar energy
      • With virtualization, multiple VMs share one physical server
      • Higher utilization can reduce the number of active servers
    • Automatic resource optimization
      • Providers monitor idle resources and redistribute workloads to turn off unused capacity
  • Business/strategic emphasis:
    • Sustainability affects purchasing decisions
    • Many organizations include carbon footprint reduction targets
    • Green cloud becomes competitive/social responsibility, not only environmental concern

Q&A / additional conceptual clarifications (from comments/questions)

Q1: Will physical storage devices disappear?

  • Unlikely to disappear soon.
  • Storage tech changes over time (analogy: RAM chips, floppy disks, CDs/DVDs, USB, external drives).
  • Cloud reduces risk of user data loss, but:
    • Cloud servers still require physical storage
    • The difference is who owns/manages the hardware (the provider)
  • Physical devices remain important for backups/jobs and local-storage scenarios.

Q2: What houses cloud internet infrastructure?

  • Data center
  • Includes more than buildings:
    • Servers, storage, networking equipment
    • Security systems
    • Backup power (generators, battery banks)
    • Advanced cooling
    • Fire protection
    • Biometric/physical access controls
    • Fiber connectivity and global distribution

Q3: Best cloud platform to learn without a student email address?

  • Major providers offer free options, so student status isn’t required.
  • By learning goal:
    • AWS: broad ecosystem; free tier; labs/experiments
      • Limitation: may require a bank card for verification
    • Azure: free credit and labs; easier learning for those already using Microsoft tools
    • Google Cloud: free account with access to key services; interactive labs and prepared environments

Q4: Best practices to reduce risk of losing confidential data and mitigating hacker attacks

  • Key concept: shared responsibility model
    • Provider secures physical infrastructure (data centers, servers, network, power, cooling, availability)
    • Customer secures their information
  • Practical best practices:
    • Encrypt information
    • Use multi-factor authentication
    • Apply least privilege permissions
    • Maintain backups
      • Prefer backups in other regions

Q5: Why are virtual machines required if organizations already have physical computers?

  • Virtualization improves utilization and reduces wasted hardware capacity.
  • Example scenario:
    • A powerful server (e.g., 32 cores, 128GB RAM) might use only ~10–15% capacity while running one service
  • With virtualization, multiple independent VMs can run on the same hardware:
    • Web server VM
    • Database VM
    • Email VM
    • Human resources VM
    • Developer VM
  • Benefit:
    • Avoid buying separate physical servers for each function
    • Lower costs and improve scaling/efficiency

Closing summary (speaker’s wrap-up)

  • Cloud computing is an evolving ecosystem, not only rented servers.
  • Connections covered:
    • Providers (AWS/Azure/IBM/Oracle/Alibaba, etc.)
    • Cloud evolution: AI integration, edge computing, serverless, multicloud/hybrid architectures, green/sustainable cloud
  • Emphasized recommendation:
    • Understand relationships between concepts rather than memorizing names
  • Final reflection:
    • Why providers grow: innovation and competition for smarter/faster/safer/easier-to-use services.

Speakers / sources featured

  • Single unnamed speaker/instructor (host/teacher in the course; examples are referenced as “as I tell my students,” with no personal name given).
  • Cloud provider companies mentioned as sources/examples:
    • Amazon Web Services (AWS)
    • Microsoft (Azure)
    • Google Cloud
    • IBM
    • Oracle
    • Alibaba
    • DigitalOcean
    • UBH Cloud (as spoken)
  • No other specific named authors/sources are cited.

Original video