GCP Associate Cloud Engineer (Week 4)

Interconnecting Networks

• Cloud VPN
  • Securely connects on-premises network to GCP VPC network using an IPSec VPN tunnel
    • IPSec: IPsec is commonly used to secure VPNs. While a VPN creates a private network between a user’s computer and the VPN server, IPsec protocols implement a secure network that protects VPN data from outside access
  • Useful for low-volume data connections
  • 99.9% SLA
  • Supports:
    • Site-to-site VPN
    • Static & Dynamic routes (Cloud Router)
    • IKEv1 & IKEv2 ciphers (Internet Key Exchange)
• Dynamic routes (Cloud Router)
  • BGP (Border Gateway Protocol)

• Cloud Interconnect
  • Dedicated Interconnect provides direct physical connection between your on-premises network and GCP network
    • Requirement: A cross-connect between the google network and router in a common co-location facility to establish a BGP session (The two routers themselves cannot establish BGP session?)
      
  • Partner Interconnect provides connectivity through a supported service provider
    
• Comparison
  

• Peering
  • Peering is a method that allows two networks to connect and exchange traffic directly without having to pay a third party to carry traffic across the Internet
  • Direct Peering provides a direct connection between your business network and Google’s
    • Broad-reaching edge network locations
    • Exchange BGP routes
    • Reach all of Google’s services
    • Peering requirements
    • No SLA
  • Edge Points of Presence (PoPs)
    • Where google’s network connects to the rest of the internet
  • Carrier Peering provides connectivity through a supported partner
    • Carrier Peering partner
    • Reach all of Google’s services
    • Partner requirements
    • No SLA

• Sharing VPC Networks
  • Shared VPC
    
    • Allows an organization to connect resources from multiple projects to a common VPC network
    • A host project
    • Communicate within Shared VPC using private IP, host project with external networks using public IP
  • VPC peering
    • Bi-directional peering, then communicate with private IP addresses even in different organizations
    • Decentralized or distributed approach to multi project networking

Load Balancing and Autoscaling

• Managed Instance Groups
  • Deploy and manage identical VM instances based on template
  • Instance group can be resized
  • Manager ensures all instances are running
  • Typically used with autoscaler to offer autoscaling capabilities
    • Dynamically add/remove instances
    • Autoscaling policy:
      • CPU utilization
      • Load balancing capacity
      • Monitoring metrics
      • Queue-based workload
  • Can be single zone or regional
• Health check
• HTTP(S) Load Balancing
  • Layer 7 of OSI model
  • Global load balancing
  • Single anycast IP address
    • Anycast is a network addressing and routing methodology in which a single destination IP address is shared by devices in multiple locations
  • HTTP on port 80 or 8080, HTTPs on port 443
  • IPv4 & IPv6
  • Autoscaling, prewarming-free
  • URL maps, default closest (?)
• Backend services
  
  • Health check
  • Session affinity (optional, default round robin algo ?)
  • Time out setting (30s default)
  • One or more backends
    • An instance group (managed or unmanaged)
    • A balancing mode (CPU utilization or RPS)
    • A capacity scaler

• HTTP(S) load balancing
  • Target HTTP(S) proxy
  • One signed SSL certificate installed
  • Client SSL session terminates at the load balancer
  • Support the QUIC transport layer protocol
• SSL certificates
  • Required for HTTP(S) load balancing
  • Up to 10 SSL certificates (per target proxy)
  • Create an SSL certificate resource

• SSL Proxy Load Balancing
  • Global load balancing for encrypted, non-HTTP traffic
  • Terminates SSL session at load balancing layer
  • IPv4 or IPv6 clients
  • Benefits:
    • Intelligent routing
    • Certificate management
    • Security patching
    • SSL policies
• TCP Proxy Load Balancing
  • Global load balancing for unencrypted, non-HTTP traffic
  • Terminates TCP sessions at load balancing layer
  • IPv4 or IPv6 clients
  • Benefits:
    • Intelligent routing
    • Security patching
• Network load balancing
  • Regional, non-proxied load balancer
  • Forwarding rules (IP protocol data)
  • Traffic: UDP; TCP/SSL ports
  • Backends: Instance group; Target pool
    • Target pool resource defines a group of instances that receive incoming traffic from forwarding rules
• Internal load balancing
  • Regional, private load balancing
    • VM instances in the same region
    • RFC 1918 IP addresses
  • TCP/UDP traffic
  • Reduced latency, simpler configuration
  • Software-defined, fully distributed load balancing
  • Supports 3-tier web services

Infrastructure Automation

• Deployment Manager
  • An infrastructure deployment service that automates the creation and management of GCP resources
    • Repeatable deployment process (Template-driven)
    • Declarative language
    • Parallel deployment, use reference to indicate dependence
• GCP Marketplace
  • Quickly deploy functional software packages that run on GCP
  • Single bill for GCP and third-party services
  • Manage solutions using Deployment Manager
  • Notifications when a security update is available
  • Direct access to partner support

Managed Services

• BigQuery
  • Serverless, highly scalable and cost-effective cloud data warehouse
  • Fully managed
  • Petabyte scaleSQL interface
  • Very fast
  • Free usage tier
• Cloud Dataflow
  • Serverless, fully managed data processing
  • Batch and stream processing with autoscale
  • Open source programming using Beam
  • Intelligently scale to millions of QPS
  • Extract, transform, and load (ETL) orchestration is a common mechanism for building big data pipelines
• Cloud Dataprep
  • Visually explore, clean and prepare data for analysis and machine learning
  • Serverless, works at any scale
  • Suggestion ideal data transformation
  • Focus on data analysis
  • Integrated partner service operated b Trifacta
• Cloud Dataproc
  • A service for running Apache Spark and Apache Hadoop clusters
  • Low cost (per-second, preemptible)
  • Super fast to start, scale and shut down
  • Integrated with GCP
  • Managed service
  • Simple and familiar

A leased line is a dedicated, fixed-bandwidth data connection.

Terraform is a tool for building, changing, and versioning infrastructure safely and efficiently.

Infrastructure as code is the process of managing infrastructure in a file or files rather than manually configuring resources in a user interface.

Containers in the Cloud

• Infrastructure as a Service allows you to share compute resources with other developers by virtualizing the hardware using virtual machines

• Strong Consistency & Eventual consistency
  • Strong Consistency offers up-to-date data but at the cost of high latency
  • While Eventual consistency offers low latency but may reply to read requests with stale data since all nodes of the database may not have the updated data
• cURL
  • curl is used in command lines or scripts to transfer data
  • The name stands for “Client URL”