Posts

Google Cloud Application Load Balancer

  Google Cloud Application Load Balancer (Layer 7) Let's understand this in a simple, beginner-friendly way. 1. What is an Application Load Balancer? An Application Load Balancer (ALB) works at Layer 7 – Application Layer of the OSI model. It distributes HTTP/HTTPS traffic based on application-level information, such as: URL path ( /users , /payments ) Hostname ( api.example.com ) HTTP headers Example: A user visits www.example.com . /products → Product service /payments → Payment service /images → Cloud Storage The load balancer checks the URL and sends the request to the correct backend. 2. What can be used as a backend? Google Cloud Application Load Balancer can route traffic to: Compute Engine VMs Google Kubernetes Engine (GKE) Cloud Run Cloud Storage (backend buckets, supported by global external ALB) External backends over the internet or hybrid connectivity 3. Types of Application Load Balancers Type Meaning Global external Accepts internet traffic and supports global, mu...

Managed Instance Group (MIG)

  Think of a Managed Instance Group (MIG) as a Kubernetes ReplicaSet, but for VMs . As a backend developer, imagine you need 10 identical application servers running your Spring Boot application. Instead of manually creating and managing 10 VMs, you: Define a blueprint ( Instance Template ) Tell Google Cloud to maintain 10 copies Google automatically creates, heals, scales, and updates them Core Components 1. Instance Template = Blueprint Contains VM configuration: Machine type OS image Startup scripts Metadata Network settings Example: Template: - Ubuntu - 4 vCPU - 16 GB RAM - Install Java - Deploy Spring Boot app This template acts like a Docker image for VMs. 2. Managed Instance Group Suppose: Desired instances = 3 Google creates: VM-1 VM-2 VM-3 All from the same template. MIG  ├── VM-1  ├── VM-2  └── VM-3 Self-Healing Feature If one VM crashes: VM-2 dies MIG notices: Current = 2 Desired = 3 Automatically creates: VM-2 (new) using the same template. This is call...

Shared VPC and VPC Network Peering

  Shared VPC and VPC Network Peering Peering means creating a direct private connection between two networks so they can communicate with each other. Simple Real-World Analogy Imagine two separate office buildings: Building A has its own internal network. Building B has its own internal network. Normally, employees would communicate through the public internet. With peering , you build a private bridge between the two buildings. Building A  <==== Private Bridge ====>  Building B Now traffic: Doesn't go over the public internet Is faster Is more secure Uses private addresses In GCP VPC Peering Two VPC networks are connected directly: VPC-A  <==== Peering ====>  VPC-B After peering: VM in VPC-A can talk to VM in VPC-B Uses private RFC1918 IPs (10.x.x.x, 172.16.x.x, 192.168.x.x) No VPN required No public IP required VPC-A VM = 10.1.1.10 VPC-B VM = 10.2.1.20 After peering: 10.1.1.10 --- > 10.2.1.20 communication works privately. VPC Peering is a n...

Cloud Peering

  Cloud Peering Used to connect your network to Google public services such as: Google Cloud Gmail YouTube Google APIs Google Workspace Two options: Direct Peering Carrier Peering 1. Direct Peering Definition: A direct connection between your network and Google's network. How it works: Uses BGP route exchange. Connects at a Google Edge PoP (Point of Presence) . Google Edge PoP (Point of Presence) is a physical location where Google's global network connects to other networks and the Internet . Think of it like this: 🏢 Your Company Network → 🌐 Google Edge PoP → ⚡ Google Global Backbone → ☁️ Google Services Simple Analogy Imagine Google owns a huge highway network. Google Edge PoP = Highway Entry Gate Once your traffic enters through that gate, it travels on Google's fast private backbone network. Features: Access to all Google services. 10 Gbps per link. Lower latency than normal internet routing. No SLA. Requirements: Must be present at a Google Edge PoP. Must meet...

GCP Interconnect Services

  GCP Interconnect Services Problem Statement A company has an on-premises datacenter and wants to communicate with resources running inside Google Cloud VPC. Possible options: Cloud VPN Dedicated Interconnect Partner Interconnect Cross-Cloud Interconnect The choice depends on bandwidth, cost, location, and business requirements. Dedicated Interconnect What is it? Dedicated Interconnect provides a private physical connection between an on-premises network and Google's network. Company Datacenter          |     Company Router          |      Fiber Link          |     Google Router          |       Cloud Router          |          VPC The company and Google meet at a common colocation faci...

GCP Connectivity Services and Cloud Interconnect and Peering

Image
 This topic explains how your company connects its office or data center to Google Cloud and Google's network. Imagine your company has its own data center and wants to connect to GCP. Google provides different ways to establish this connection. 1. Four types of connections Service Simple meaning Dedicated Interconnect Direct physical connection to Google's network Partner Interconnect Connection through a Google-approved partner Direct Peering Direct connection to Google's public network Carrier Peering Connection to Google's public network through a service provider Dedicated vs Shared Dedicated connection Your company connects directly to Google's network using a dedicated physical link. Example: Your company has a data center and needs a high-bandwidth private connection to GCP. Shared connection Your company connects to Google's network through a network service provider. Example: Your company uses a telecom provider to connect to GCP. 2. Layer 2 vs Layer 3...