Video Summary - Up in the Cloud – Your first flight to the cloud

Summary of the session: “Your first flight to the cloud” (Google Cloud intro)

Event / structure (context)

The “Balad Wad” initiative (Egypt Edition) is organized by Google for Developers in collaboration with ITI (Information Technology Institute).
Sessions run every Friday ( 4 PM Egypt / 5 PM Dubai ).
Each session includes:
- A motivational kickoff by Google Developer Relations
- Hands-on training led by a Cloud trainer

Core technological concepts taught

1) What “the Cloud” is (basic definition)

The cloud is described as a way to:

Run programs and store data on online servers accessible from anywhere.

Key benefits emphasized:

Less operational burden (the provider manages much of the infrastructure)
Faster deployment and quicker readiness to market
The cloud is framed as the “computer/power” used to run solutions and models

2) Service models (layers of responsibility)

The trainer breaks cloud responsibility into three layers:

IaaS (Infrastructure as a Service)
- You manage/configure more (e.g., OS configuration); the cloud provides underlying infrastructure.
- Example mentioned: Compute Engine (discussed later as Virtual Machines)
PaaS (Platform as a Service)
- You focus on deploying applications; the provider handles runtime/middleware/OS.
- Examples: App Engine, Cloud Run
SaaS (Software as a Service)
- You simply log in and use software; most setup is handled for you.
- Examples referenced: Google Workspace and Gemini (SaaS-like usage)

Google Cloud architecture & key product features

3) Google Cloud advantages (why Google Cloud specifically)

Strong functionality + security across the stack
Fast innovation through continuous service releases
A large global footprint with availability across many countries/regions

4) Regions, zones, and latency/regulation

Region: a geographic location chosen based on latency and compliance
Zone: an isolated data center within a region

Numbers mentioned:

43+ regions
130 zones

5) Network Edge Locations and CDN concept

Edge Locations are positioned as the foundation for CDNs (Content Delivery Networks).
CDNs cache content closer to users to reduce:
- Latency
- Bandwidth load

Additional mentions:

200+ edge locations
Security tied to Cloud Armor (referred to as protection against “DDoS / disk attacks” in subtitles)

Services overview (what they’ll use during training)

6) Google Cloud is more than storage/compute

Training positions Google Cloud as enabling:

Data
Machine Learning
AI
Other enterprise capabilities

7) Vertex AI / renamed concept: “Agent Platform”

The trainer states that Vertex AI was renamed, and the available offering is now “Enterprise Agent Platform” / “Agent Platform.”
This is connected to Generative AI agent workflows

Detailed breakdown of key compute/storage components

8) Infrastructure-level compute (IaaS)

Compute Engine (VMs)
- You select CPU/RAM/storage type
- You install/configure software yourself

Machine type categories mentioned:

Standard
High-memory
High-CPU
GPU-supported configurations

GPU is mentioned due to ML/GenAI relevance.

9) Container/cluster and managed compute

Kubernetes Engine
- Presented as managed clustering
- Built around Docker / Kubernetes concepts
Kubernetes is emphasized as handling container orchestration

10) PaaS: Cloud Run / Cloud Functions (event-driven)

Cloud Run
- Runs containers/services
- Scales with demand
- Pay-for-usage emphasized
Cloud Run Functions / Cloud Functions
- Event-driven execution
- Runs when specific events occur (e.g., a trigger like a click)
- Pay only when invoked

Database & storage services (with use cases)

11) Object storage / “cloud storage”

Stores large unstructured data (e.g., images/videos/media)
Compared to “Google Drive with a link”-style access

12) Relational databases

Cloud SQL
- Managed MySQL/PostgreSQL-like environments
- Reduces admin overhead by letting you choose:
  - Instance type/version
  - Region

13) Other database/storage options and why choose them

Trainer listed multiple options aligned to different workload patterns:

Firestore (non-relational)
- For mobile/web apps with document-style data
Spanner
- Highlighted as scalable and high availability
Bigtable
- Referred to with a subtitle artifact (“Pick Table”)
- Suitable for write-heavy / real-time workloads
Cloud data warehouse
- Described as a warehouse for analytical queries on very large datasets
- Mentions terabytes-scale datasets

Scaling, load balancing, and networking components

14) Auto-scaling types

Auto-scaling for compute resources, scaling based on factors like:
- CPU or load

Mentioned:

Prescriptive auto-scaling using forecasting (ML-based prediction of future capacity needs)

15) Load balancers

Purpose:

Distribute traffic across servers for scale and performance

Two types:

Application Load Balancer
- HTTP/HTTPS
- Routing using URL path/host
- Mentions routing rules and SSL-related concepts (including SSL certificates and Cloud Armor)
Network Load Balancer
- TCP/UDP
- Routing based on IP/ports
- Positioned for databases, gaming, mail servers (as stated)

Hands-on labs (tutorial execution)

16) Getting Cloud credits and creating a project

High-level steps for free credit/testing:

Sign in with a Google account
Fill a credit form (name/email/coupon code)
Create a new Google Cloud project (trainer mentions using a “trainer” account)
Use Cloud Console / Cloud Shell for CLI access

17) Lab 1: Create a VM (Compute Engine) + web server setup

Using the Console:

Create an instance and select:
- Region/zone
- Machine family (examples like f1-micro / low-cost)
- OS
- Disk options
- Networking (HTTP/HTTPS public settings via firewall)
Review estimated monthly cost

Using CLI:

Alternative path using gcloud

Server access:

SSH into the VM

Setup:

Install/configure Nginx (Engine X) using package manager steps (subtitle text indicated an “install engine x” style process)

Startup Script concept:

Commands that run automatically when the instance starts to configure services

18) Lab cleanup

Delete firewall rules
Delete VM instances (via Console and via CLI)

19) Lab 2: Cloud SQL with Service Account + Cloud SQL Proxy

High-level flow (partially described due to time):

Create a Service Account
Grant permissions (IAM roles) so the VM/server can access Cloud SQL privately
Configure VPC connectivity (including mention of VPC peering)
Create a Cloud SQL instance (example mentioned: PostgreSQL)
Use Cloud SQL Proxy from Compute Engine to connect securely
Install an SQL client inside the VM and connect via localhost through the proxy

AI / Agent Platform overview (mentioned, then deferred)

After transitioning from lab to product overview, the trainer mentioned:

Agent Platform components:
- Model Garden (choose models and tune)
- An agent layer that interacts with tools/data

Mentions related to enterprise agent workflows:

Building RAG / enterprise search
Vector Search and document retrieval grounding
“Grounding / ranking” through filtering/searching within documents
Future topics referenced:
- “RAG Engine”
- enterprise agent workflows

Main speakers / sources

Nour — Supervises GDG/GDE programs (Google Developer Groups / Google Developer Expert) for North Africa & Middle East
Ramesh Chandra — Developer Relations Manager for Google (North Africa, Middle East, Turkey, Central Asia)
Fadi Nabil — Cloud Solutions Architect and Google Developer Expert; trainer for the session (Google Cloud / GT Giza Coding)

Up in the Cloud – Your first flight to the cloud

Key takeaways