Google Cloud Computing Foundation Course - Week 4 Lecture Notes Summary
Lecture 20: Configuring Elastic Apps with Autoscaling
• Autoscaler Overview:
• Managed instance groups use autoscaling to add and remove VM instances based
on predefined policies.
• Policies define the minimum and maximum number of replicas in the group.
• Autoscaler adjusts resources by provisioning VMs from Compute Engine
templates.
• Instance Capacity:
• Additional VMs offer diminishing returns as the group grows larger (e.g., 10th
VM adds 10% capacity).
• Autoscaler behaves conservatively to avoid running out of capacity by adding
extra VMs when needed.
• VM Removal Strategy:
• Autoscaler prefers underutilization rather than resource exhaustion.
• It typically removes one VM at a time to stay close to target utilization.
Lecture 21: Exploring PaaS with App Engine
• App Engine Overview:
• Fully managed, serverless platform for running highly scalable applications.
• Ideal for rapid development and deployment without the need for managing
infrastructure.
• Reduces operational overhead (no server management or deployment
configuration).
• Key Features:
• Supports a range of programming languages: Java, PHP, Python, [Link], C#,
Ruby, etc.
• Command-line management, production debugging, and use of tools like Cloud
SDK, IntelliJ IDEA, Visual Studio.
• Automatically scales based on application traffic.
• App Engine Environments:
• Standard Environment: Fully managed, scales down to zero, but with runtime
and configuration limitations.
• Flexible Environment: Allows custom runtimes and infrastructure options, but
always running (even during low usage).
• Storage Options:
• Cloud Storage, Cloud SQL, and Bigtable for different data storage needs.
• Memcache for caching, and support for large object storage (up to 5 TB).
• Scaling and Load Balancing:
• Automatic scaling to meet demand and load balancing across multiple regions for
high availability.
• Can host different versions of apps for development, testing, staging, and
production.
Lecture 22: Event-Driven Programs with Cloud Functions
• Cloud Functions Overview:
• Cloud Functions are serverless and run based on specific events in GCP (event-
driven architecture).
• Ideal for small, independent units of functionality.
• Use Cases:
• Connecting and extending GCP services using event-driven automation.
• Examples include responding to file uploads, database changes, or Pub/Sub
messages.
• Key Advantages:
• Automatically managed infrastructure with seamless scaling from a few
invocations to millions.
• Easily integrates with other GCP services such as Cloud Storage, Pub/Sub, Cloud
Spanner, and Translation API.
Lecture 23: Containerizing and Orchestrating Apps with GKE
• Google Kubernetes Engine (GKE) Overview:
• GKE provides managed container orchestration using Kubernetes.
• A hybrid between Infrastructure-as-a-Service (IaaS) and Platform-as-a-Service
(PaaS).
• Ideal for teams deploying or maintaining containerized workloads without
infrastructure management.
• Containerization Benefits:
• Provides an abstraction layer over the operating system and hardware.
• Containers are lightweight, start quickly, and enable scalability and resource
isolation.
• Containers vs VMs:
• Containers scale workloads independently and more efficiently than VMs.
• They offer the flexibility of IaaS but with the performance benefits of PaaS.
• Kubernetes Capabilities:
• Kubernetes allows container orchestration at scale, deploying containers across
clusters.
• Features include automatic scaling, rolling updates, and management of container
health and storage.
• GKE Features:
• Fully managed environment for running Docker containers.
• Provides access to open-source Kubernetes with support for hybrid cloud
environments.
Lecture 24: Summary
• Cloud Compute Options:
• GCP provides four main computing options: Compute Engine, App Engine,
Cloud Functions, and GKE.
• Compute Engine delivers VMs with IaaS.
• App Engine simplifies app development with a serverless PaaS model.
• Cloud Functions allows event-driven automation without infrastructure
management.
• GKE is used for container orchestration with the benefits of both IaaS and PaaS.
QUESTIONS & ANSWERS
Lecture 20: Configuring Elastic Apps with Autoscaling
1. What does autoscaling do in GCP?
• Autoscaling automatically adjusts the number of virtual machines (VMs) in
response to changes in load, based on policies.
2. What are the policies involved in autoscaling?
• Policies include the minimum and maximum number of instance replicas.
3. How does the percentage utilization of a VM change in larger groups?
• The more VMs in a group, the less percentage each new VM adds to the overall
capacity.
4. What happens if autoscaler needs to remove VMs to meet a target?
• It conservatively removes one VM rather than overshooting the target and
removing too many.
5. Why does the autoscaler prefer underutilization over running out of resources?
• To avoid service disruption when there is a sudden need for more capacity.
Lecture 21: Exploring PaaS with App Engine
6. What is App Engine in GCP?
• App Engine is a fully managed serverless platform for building and running
highly scalable applications.
7. What are the main benefits of using App Engine?
• It eliminates the need for managing infrastructure, allowing developers to focus
on writing code.
8. Which programming languages are supported by App Engine?
• Languages like Java, PHP, [Link], Python, C#, .NET, Ruby, Go, and custom
runtimes.
9. What is the difference between the standard and flexible environments in App
Engine?
• The standard environment is fully managed and scales to zero when idle, while
the flexible environment runs on Docker containers and does not scale down to
zero.
[Link] are some storage options in App Engine?
• Options include App Engine Memcache, Cloud Storage, Cloud SQL, and Cloud
Bigtable.
Lecture 22: Event-Driven Programs with Cloud Functions
[Link] is a Cloud Function in GCP?
• Cloud Functions are serverless pieces of code that execute in response to events
in the cloud.
[Link] programming languages can be used for Cloud Functions?
• You can use [Link], Python, and Go for Cloud Functions.
[Link] kind of events trigger Cloud Functions?
• Events like file uploads to Cloud Storage, changes in Cloud Datastore, and
messages from Cloud Pub/Sub.
[Link] do Cloud Functions scale?
• They scale automatically from a few invocations per day to millions without
manual intervention.
[Link] is a trigger in Cloud Functions?
• A trigger binds a function to a specific event or set of events that the function
responds to.
Lecture 23: Containerizing and Orchestrating Apps with GKE
[Link] is Google Kubernetes Engine (GKE)?
• GKE is a managed environment for deploying containerized applications using
Kubernetes.
[Link] is the primary benefit of using containers in GKE?
• Containers allow scalability and independence of workloads while abstracting the
underlying infrastructure.
[Link] does containerization work in GKE?
• Containers virtualize the operating system, allowing lightweight, independent
scalability and flexibility like infrastructure-as-a-service.
[Link] is the role of Kubernetes in GKE?
• Kubernetes orchestrates containers, managing clusters, scaling, rollouts, and
rollbacks of applications.
[Link] makes GKE ideal for organizations?
• It allows easy management and scaling of containerized workloads without
handling the underlying server infrastructure.
[Link] is the difference between virtual machines (VMs) and containers in GKE?
• VMs virtualize hardware while containers virtualize the OS, making containers
faster and more lightweight.
General GCP Topics
[Link] are the four main computing options in GCP?
• Compute Engine, App Engine, Cloud Functions, and Google Kubernetes Engine
(GKE).
[Link] is Compute Engine in GCP?
• It provides infrastructure-as-a-service (IaaS) through virtual machines running in
Google’s data centers.
[Link] does autoscaling work with Managed Instance Groups?
• It adjusts the number of instances in a group automatically based on the current
demand.
[Link] are the differences between infrastructure-as-a-service (IaaS) and platform-
as-a-service (PaaS)?
• IaaS gives full control over virtual machines and infrastructure, while PaaS
abstracts this to focus on application development.
[Link] does App Engine handle scalability?
• App Engine automatically scales based on application traffic and resource needs.
[Link] is the purpose of load balancing in App Engine?
• It distributes traffic to ensure consistent performance and scalability of the front-
end services.
[Link] is Cloud SQL in GCP?
• Cloud SQL is a fully-managed relational database service for MySQL,
PostgreSQL, and SQL Server databases.
[Link] is the purpose of Cloud Bigtable in GCP?
• Cloud Bigtable is a NoSQL database designed for heavy read/write throughput
and big data workloads.
[Link] are the advantages of using Google Cloud Functions?
• Cloud Functions remove infrastructure management, automatically scale, and are
event-driven.
[Link] are Cloud Functions connected to other GCP services?
• Cloud Functions can be seamlessly integrated with GCP services such as Cloud
Datastore, Cloud Spanner, and Cloud Vision API.
[Link] role does event-driven architecture play in GCP?
• It allows functions to run in response to specific events, promoting a decoupled
and reactive system architecture.
[Link] is Pub/Sub in GCP?
• Pub/Sub is a messaging service that decouples event producers and consumers,
facilitating asynchronous communication.
[Link] are the key features of Kubernetes?
• Kubernetes manages containerized workloads and services, including scaling,
rolling updates, and managing container health.
[Link] are containers considered more efficient than VMs?
• Containers start faster, use fewer resources, and provide a higher level of
abstraction than VMs.
[Link] are microservices, and how do containers support them?
• Microservices are small, independently deployable units of application logic, and
containers help by isolating and scaling them individually.
[Link] are the benefits of using Docker containers?
• Docker containers provide portability, consistency, and a lightweight alternative
to traditional VMs.
[Link] are some use cases for Google Kubernetes Engine (GKE)?
• GKE is ideal for managing microservices architectures, scaling containerized
apps, and handling production-grade workloads.
[Link] are the common cloud storage options for GCP applications?
• Cloud Storage, Cloud SQL, Cloud Datastore, and Cloud Bigtable.
[Link] does GKE integrate with Kubernetes?
• GKE uses Kubernetes to orchestrate and manage container clusters on Google
Cloud.
Security and Monitoring
[Link] security tools are available for applications in GCP?
• Tools like Cloud Security Scanner and IAM (Identity and Access Management)
are available.
[Link] does GCP handle redundancy and high availability?
• GCP uses load balancing, autoscaling, and regional replication to ensure
redundancy and availability.
[Link] can you monitor the health of GCP applications?
• Tools like Stackdriver provide diagnostics, monitoring, and debugging for apps in
production.
[Link] is Stackdriver in GCP?
• Stackdriver provides monitoring, logging, and diagnostics to manage applications
in GCP.
[Link] are the advantages of using Cloud Load Balancing?
• It helps distribute traffic, ensuring application performance and high availability
across regions.
Cost and Optimization
[Link] does GCP manage costs for scaling applications?
• GCP scales resources based on usage, reducing costs by stopping unused
instances or reducing resource allocation.
[Link] does App Engine help minimize operational costs?
• It only consumes resources when code is running and eliminates the need for
infrastructure management.
[Link] is the pricing model for GCP services like Cloud Functions?
• Cloud Functions are billed based on the number of invocations, duration, and
resources used.
[Link] can autoscaling reduce costs in GCP?
• By dynamically adjusting resources based on demand, autoscaling ensures
efficient use of resources.
[Link] is the role of a service mesh in GCP?
• A service mesh manages communication between microservices, ensuring
reliability and security.
