Skip to content
Theme

Azure

Azure infrastructure diagrams using HashiCorp Flight and Carbon icon packs for VNet networking, compute, and managed services.

VNet with App Gateway

Section titled “VNet with App Gateway”

Azure VNet with gateway, application, and data subnets. Application Gateway distributes traffic to VM Scale Sets.

architecture-beta
  group vnet(carbon:virtual-private-cloud)[Azure VNet]
  group gwsub(carbon:ibm-cloud-subnets)[Gateway Subnet] in vnet
  group appsub(carbon:ibm-cloud-subnets)[App Subnet] in vnet
  group datasub(carbon:ibm-cloud-subnets)[Data Subnet] in vnet

  service appgw(carbon:gateway-security)[App Gateway] in gwsub
  service vm1(hashicorp-flight:azure-vms-color)[VM Scale Set] in appsub
  service vm2(hashicorp-flight:azure-vms-color)[VM Scale Set] in appsub
  service sqldb(carbon:data-base)[Azure SQL] in datasub

  appgw:R --> L:vm1
  appgw:B --> T:vm2
  vm1:R --> L:sqldb
  vm2:R --> L:sqldb

AKS with F5 XC Multi-Cloud Connect

Section titled “AKS with F5 XC Multi-Cloud Connect”

Azure Kubernetes Service fronted by F5 Distributed Cloud for multi-cloud application connectivity and security.

architecture-beta
  group xc(lucide:cloud)[F5 XC]
  group aks(hashicorp-flight:azure-aks-color)[AKS Cluster]

  service mcn(f5xc:multi-cloud-network-connect)[Network Connect] in xc
  service waap(f5xc:web-app-and-api-protection)[WAAP] in xc
  service ingress(carbon:gateway)[Ingress] in aks
  service app(hashicorp-flight:docker-color)[App Pods] in aks
  service cache(carbon:datastore)[Redis Cache] in aks
  service blob(hashicorp-flight:azure-blob-storage-color)[Blob Storage]

  mcn:R --> L:waap
  waap:R --> L:ingress
  ingress:R --> L:app
  app:B --> T:cache
  app:R --> L:blob

Hub-Spoke Network Topology

Section titled “Hub-Spoke Network Topology”

Azure Hub-Spoke architecture with centralized security and shared services connecting multiple spoke VNets.

flowchart TD
  hub@{ icon: 'carbon:virtual-private-cloud', label: 'Hub VNet' }
  fw@{ icon: 'carbon:firewall', label: 'Azure Firewall' }
  vpn@{ icon: 'carbon:gateway-vpn', label: 'VPN Gateway' }
  spoke1@{ icon: 'carbon:ibm-cloud-subnets', label: 'Spoke 1 - Web' }
  spoke2@{ icon: 'carbon:ibm-cloud-subnets', label: 'Spoke 2 - App' }
  spoke3@{ icon: 'carbon:ibm-cloud-subnets', label: 'Spoke 3 - Data' }
  onprem@{ icon: 'lucide:server', label: 'On-Premises DC' }

  onprem --> vpn
  vpn --> hub
  hub --> fw
  fw --> spoke1
  fw --> spoke2
  fw --> spoke3

NVA HA with Load Balancer — Internet Traffic

Section titled “NVA HA with Load Balancer — Internet Traffic”

Inbound internet traffic hits a public load balancer, which distributes to NVA instances in the hub. The NVA forwards inspected traffic to spoke workloads. Return traffic from spokes routes through an internal load balancer back to the NVA for egress. Numbered steps show the inbound path (1-3) and return path (4-6).

flowchart TD
  subgraph internet[Internet]
    cloud@{ icon: 'lucide:globe', label: 'Internet' }
  end
  subgraph hub[Hub VNet 10.0.0.0/24]
    subgraph gwsub[Gateway Subnet 10.0.0.0/27]
      gw@{ icon: 'azure:virtual-network-gateways', label: 'VPN/ER GW' }
    end
    subgraph nvasub[NVA Subnet 10.0.0.32/27]
      intlb@{ icon: 'azure:load-balancers', label: 'Internal LB 10.0.0.36' }
      nva@{ icon: 'azure:firewalls', label: 'NVA' }
    end
    publb@{ icon: 'azure:load-balancers', label: 'Public LB' }
  end
  subgraph spoke1[Spoke1 10.1.1.0/24]
    app1@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  subgraph spoke2[Spoke2 10.1.2.0/24]
    app2@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  onprem@{ icon: 'lucide:building', label: 'On-Premises 192.168.0.0/16' }

  cloud -->|1| publb
  publb -->|2| nva
  nva -->|3| app2
  app2 -->|4| intlb
  intlb -->|5| nva
  nva -->|6| cloud
  onprem --> gw
  gw --> intlb

NVA HA with Load Balancer — On-Premises Traffic

Section titled “NVA HA with Load Balancer — On-Premises Traffic”

On-premises traffic enters through a VPN or ExpressRoute gateway and is directed to an internal load balancer fronting multiple NVA instances. The NVA inspects and forwards traffic to spoke workloads. Return traffic traverses the same internal load balancer to ensure flow symmetry, preventing asymmetric routing issues.

flowchart TD
  subgraph hub[Hub VNet 10.0.0.0/24]
    subgraph gwsub[Gateway Subnet 10.0.0.0/27]
      gw@{ icon: 'azure:virtual-network-gateways', label: 'VPN/ER GW' }
    end
    subgraph nvasub[NVA Subnet 10.0.0.32/27]
      intlb@{ icon: 'azure:load-balancers', label: 'Internal LB 10.0.0.36' }
      nva1@{ icon: 'azure:firewalls', label: 'NVA' }
      nva2@{ icon: 'azure:firewalls', label: 'NVA' }
    end
  end
  subgraph spoke1[Spoke1 10.1.1.0/24]
    app1@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  subgraph spoke2[Spoke2 10.1.2.0/24]
    app2@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  onprem@{ icon: 'lucide:building', label: 'On-Premises 192.168.0.0/16' }

  onprem -->|1| gw
  gw -->|2| intlb
  intlb -->|3| nva1
  nva1 -->|4| app2
  app2 -->|5| intlb
  intlb -->|6| nva2
  nva2 -->|7| gw
  gw -->|8| onprem

NVA HA with PIP/UDR — Active/Standby

Section titled “NVA HA with PIP/UDR — Active/Standby”

Active/standby NVA pair where the active instance (NVA1) holds the public IP address. On failure, the standby NVA2 calls the Azure API to reassign the public IP and update user-defined routes to point to itself. This approach avoids load balancers but requires API-level failover orchestration.

flowchart TD
  subgraph internet[Internet]
    cloud@{ icon: 'lucide:globe', label: 'Internet' }
  end
  subgraph hub[Hub VNet 10.0.0.0/24]
    pip@{ icon: 'azure:public-ip-addresses', label: 'Public IP' }
    subgraph gwsub[Gateway Subnet 10.0.0.0/27]
      gw@{ icon: 'azure:virtual-network-gateways', label: 'VPN/ER GW' }
    end
    subgraph nvasub[NVA Subnet 10.0.0.32/27]
      nva1@{ icon: 'azure:firewalls', label: 'NVA1 Active 10.0.0.37' }
      nva2@{ icon: 'azure:firewalls', label: 'NVA2 Standby 10.0.0.38' }
    end
  end
  subgraph spoke1[Spoke1 10.1.1.0/24]
    app1@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  subgraph spoke2[Spoke2 10.1.2.0/24]
    app2@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  onprem@{ icon: 'lucide:building', label: 'On-Premises 192.168.0.0/16' }

  cloud -->|1| pip
  pip -->|2| nva1
  nva1 -->|3| app2
  app2 -->|4| nva1
  nva1 -->|5| cloud
  onprem --> gw

NVA HA with Azure Route Server

Section titled “NVA HA with Azure Route Server”

BGP-based high availability using Azure Route Server. The Route Server establishes eBGP adjacencies with both NVA instances and dynamically programs spoke effective routes. ECMP load balances across NVAs without user-defined routes. Route Server injects next-hop entries for both NVA IPs into all peered VNets.

flowchart TD
  subgraph internet[Internet]
    cloud@{ icon: 'lucide:globe', label: 'Internet' }
  end
  subgraph hub[Hub VNet 10.0.0.0/24]
    publb@{ icon: 'azure:load-balancers', label: 'Public LB' }
    subgraph gwsub[Gateway Subnet 10.0.0.0/27]
      gw@{ icon: 'azure:virtual-network-gateways', label: 'VPN/ER GW' }
    end
    subgraph nvasub[NVA Subnet 10.0.0.32/27]
      nva1@{ icon: 'azure:firewalls', label: 'NVA1 10.0.0.37' }
      nva2@{ icon: 'azure:firewalls', label: 'NVA2 10.0.0.38' }
    end
    subgraph rssub[Route Server Subnet 10.0.0.64/27]
      rs@{ icon: 'azure:virtual-router', label: 'Route Server' }
    end
  end
  subgraph spoke1[Spoke1 10.1.1.0/24]
    app1@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end
  subgraph spoke2[Spoke2 10.1.2.0/24]
    app2@{ icon: 'azure:virtual-machine', label: 'App Server' }
  end

  cloud -->|1| publb
  publb -->|2| nva1
  nva1 -->|3| app2
  app2 -->|4| nva1
  nva1 -->|5| cloud
  rs <-.->|eBGP| nva1
  rs <-.->|eBGP| nva2
  gw --> rs

NVA HA with Gateway Load Balancer

Section titled “NVA HA with Gateway Load Balancer”

Transparent NVA insertion using Azure Gateway Load Balancer. Traffic destined for the application is transparently diverted from the public standard load balancer to the Gateway LB in a separate NVA VNet. NVAs inspect traffic and return it to the Gateway LB, which forwards it back to the application. No VNet peering or UDRs are required between the NVA and application VNets.

flowchart TD
  subgraph internet[Internet]
    cloud@{ icon: 'lucide:globe', label: 'Internet' }
  end
  subgraph nvavnet[NVA VNet]
    gwlb@{ icon: 'azure:load-balancers', label: 'Gateway LB' }
    nva1@{ icon: 'azure:firewalls', label: 'NVA' }
    nva2@{ icon: 'azure:firewalls', label: 'NVA' }
  end
  subgraph appvnet[App VNet]
    publb@{ icon: 'azure:load-balancers', label: 'Public Std LB' }
    web@{ icon: 'azure:virtual-machine', label: 'Web Server' }
  end

  cloud -->|1| publb
  publb -->|2| gwlb
  gwlb -->|3| nva1
  nva1 -->|4| gwlb
  gwlb -->|5| publb
  publb -->|6| web
  gwlb --> nva2