Skip to content

Professional Cloud Network Engineer · Domain 4: Configuring and implementing hybrid and multicloud network interconnectivity · 16% of the exam

Task 4.2: Configuring a site-to-site IPSec VPN.

HA VPN toward on-premises gateways and toward other VPC networks, and Classic VPN with route-based and policy-based tunnels.

Study it

  • Cloud Interconnect and site-to-site IPsec: Dedicated, Partner and Cross-Cloud connections, VLAN attachments, SLA topologies, MACsec, HA VPN (also over Cloud Interconnect) and Classic VPN

    Partly covered by: Hybrid Networking & DNS

Sample questions

Try each one before opening the answer. Every option is explained, with the Google Cloud documentation page that proves it.

Question 1 · choose 1

A data center has two physical VPN appliances, each with one public IP address, so that either can be taken offline for upgrades. They will connect to one HA VPN gateway, and the connection must qualify for the 99.99% availability SLA on the Google Cloud side. How should the external VPN gateway resource and the tunnels be configured?

  1. AAn external VPN gateway with SINGLE_IP_INTERNALLY_REDUNDANT for the first appliance, and both tunnels to that appliance
  2. BAn external VPN gateway with TWO_IPS_REDUNDANCY, and two tunnels from HA VPN interface 0, one to each appliance
  3. CTwo external VPN gateways with SINGLE_IP_INTERNALLY_REDUNDANT, one per appliance, each with a tunnel from HA VPN interface 0
  4. DAn external VPN gateway with TWO_IPS_REDUNDANCY, and tunnels from HA VPN interfaces 0 and 1 to peer interfaces 0 and 1
Show the answer and why
  • AAn external VPN gateway with SINGLE_IP_INTERNALLY_REDUNDANT for the first appliance, and both tunnels to that appliance

    Incorrect

    That redundancy type describes one peer gateway with one IP address; the second appliance would carry no tunnel at all.

  • BAn external VPN gateway with TWO_IPS_REDUNDANCY, and two tunnels from HA VPN interface 0, one to each appliance

    Incorrect

    Interface 1 of the HA VPN gateway would have no tunnel, and every HA VPN interface needs one for the 99.99% SLA.

  • CTwo external VPN gateways with SINGLE_IP_INTERNALLY_REDUNDANT, one per appliance, each with a tunnel from HA VPN interface 0

    Incorrect

    Both tunnels still start on interface 0; two peer gateways that one HA VPN gateway connects to are described with TWO_IPS_REDUNDANCY.

  • DAn external VPN gateway with TWO_IPS_REDUNDANCY, and tunnels from HA VPN interfaces 0 and 1 to peer interfaces 0 and 1

    Correct

    One HA VPN gateway to two peer gateways with one IP each uses TWO_IPS_REDUNDANCY, and matching tunnels on both interfaces meet the SLA.

For HA VPN to two peer VPN gateways, each with one interface and one external IP address, the external VPN gateway takes the redundancy type TWO_IPS_REDUNDANCY and the HA VPN gateway uses one tunnel to each peer. The 99.99% SLA on the Google Cloud side needs a tunnel on each HA VPN interface, matched interface 0 to peer interface 0 and interface 1 to peer interface 1; a full mesh isn't required.

Google Cloud documentation

Question 2 · choose 1

A partner's legacy VPN device supports only static, policy-based IPsec and can't run BGP. A Classic VPN tunnel to it was created as route-based in the Google Cloud console, and IKE negotiation fails because the device expects the local traffic selector 10.10.0.0/16 and the remote selector 172.20.0.0/16. What should the network engineer do?

  1. ADelete the tunnel and re-create it as policy-based, with local 10.10.0.0/16 and remote 172.20.0.0/16
  2. BEdit the existing tunnel and change its local traffic selector from 0.0.0.0/0 to 10.10.0.0/16
  3. CAdd a custom static route for 172.20.0.0/16 whose next hop is the existing Classic VPN tunnel
  4. DReplace the tunnel with an HA VPN gateway and two tunnels that use static routes to 172.20.0.0/16
Show the answer and why
  • ADelete the tunnel and re-create it as policy-based, with local 10.10.0.0/16 and remote 172.20.0.0/16

    Correct

    A policy-based tunnel sets the local and remote selectors to the ranges you give, and selectors can only be set when the tunnel is created.

  • BEdit the existing tunnel and change its local traffic selector from 0.0.0.0/0 to 10.10.0.0/16

    Incorrect

    Traffic selectors can't be changed after a tunnel has been created; the tunnel must be deleted and re-created.

  • CAdd a custom static route for 172.20.0.0/16 whose next hop is the existing Classic VPN tunnel

    Incorrect

    Routes decide which packets enter the tunnel; they don't change the traffic selectors the tunnel offers during IKE negotiation.

  • DReplace the tunnel with an HA VPN gateway and two tunnels that use static routes to 172.20.0.0/16

    Incorrect

    HA VPN supports only dynamic routing with BGP, which this device can't run.

A route-based Classic VPN tunnel created in the console sets both traffic selectors to 0.0.0.0/0 and specifies only remote ranges, used for routes. A peer that needs specific local and remote selectors requires a policy-based tunnel, and because selectors can't be edited, the tunnel is re-created. HA VPN would be the documented choice for a peer that can run BGP.

Question 3 · choose 1

Two HA VPN tunnels to the data center run as an active-active configuration and together carry close to their combined throughput. One tunnel goes down. What should the team expect?

  1. ATraffic continues on one tunnel, with the effective throughput cut in half
  2. BAll traffic stops until the failed tunnel is re-created by hand
  3. CThe remaining tunnel doubles its own capacity to absorb the full load
  4. DTraffic moves to Classic VPN tunnels that HA VPN keeps as a warm standby for failures
Show the answer and why
  • ATraffic continues on one tunnel, with the effective throughput cut in half

    Correct

    In active-active, the aggregate throughput is the sum of both tunnels, so losing one halves it.

  • BAll traffic stops until the failed tunnel is re-created by hand

    Incorrect

    A Cloud VPN tunnel that goes down restarts automatically, and the other tunnel keeps carrying traffic.

  • CThe remaining tunnel doubles its own capacity to absorb the full load

    Incorrect

    A tunnel's capacity doesn't grow when its partner fails.

  • DTraffic moves to Classic VPN tunnels that HA VPN keeps as a warm standby for failures

    Incorrect

    HA VPN keeps no Classic VPN tunnels; both HA VPN tunnels stay active in either routing configuration.

In an active-active HA VPN configuration, both tunnels carry traffic and the effective aggregate throughput is their combined throughput; when one fails, it is cut in half. Size the tunnels so that one can carry the required load if that matters. A failed tunnel restarts automatically.

Google Cloud documentation

Question 4 · choose 1

A team created an IPv4-only HA VPN gateway to connect two VPC networks and now needs the gateway to carry IPv6 as well. What must it do?

  1. AUpdate the gateway's stack type to IPv4 and IPv6 in place
  2. BDelete the gateway and re-create it with the IPv4 and IPv6 stack type
  3. CAdd a second Cloud Router with IPv6 BGP sessions to the existing gateway
  4. DConvert the gateway to a Classic VPN gateway, which supports IPv6
Show the answer and why
  • AUpdate the gateway's stack type to IPv4 and IPv6 in place

    Incorrect

    An HA VPN gateway's stack type can't be modified after creation.

  • BDelete the gateway and re-create it with the IPv4 and IPv6 stack type

    Correct

    To get a different stack type, the HA VPN gateway must be deleted and re-created.

  • CAdd a second Cloud Router with IPv6 BGP sessions to the existing gateway

    Incorrect

    The gateway's own stack type stays IPv4-only, so it can't carry IPv6 tunnels.

  • DConvert the gateway to a Classic VPN gateway, which supports IPv6

    Incorrect

    Classic VPN gateways don't support IPv6.

The stack type of an HA VPN gateway is chosen at creation and can't be changed afterwards; a gateway with another stack type means deleting and re-creating it. Classic VPN gateways don't support IPv6 at all.

Question 5 · choose 1

A company moves a site from Classic VPN to HA VPN. Its partner's firewall allow-lists the current Classic VPN gateway's external IP address, and the partner's device supports BGP. What should the company plan for?

  1. AConverting the existing Classic VPN tunnels into HA VPN tunnels in place
  2. BKeeping the Classic VPN gateway's external IP address on the new HA VPN gateway
  3. CConnecting the new HA VPN gateway to the Classic VPN gateway during the cutover
  4. DNew tunnels on an HA VPN gateway that gets two new external IP addresses, which the partner must allow-list
Show the answer and why
  • AConverting the existing Classic VPN tunnels into HA VPN tunnels in place

    Incorrect

    Existing Classic VPN tunnels can't be migrated to an HA VPN gateway.

  • BKeeping the Classic VPN gateway's external IP address on the new HA VPN gateway

    Incorrect

    The Classic VPN gateway's external IP address can't be preserved.

  • CConnecting the new HA VPN gateway to the Classic VPN gateway during the cutover

    Incorrect

    Classic VPN gateways can't connect to HA VPN gateways.

  • DNew tunnels on an HA VPN gateway that gets two new external IP addresses, which the partner must allow-list

    Correct

    HA VPN creates two new external addresses from two pools, and new tunnels replace the old ones.

Migrating from Classic VPN to HA VPN means creating new HA VPN tunnels and deleting the old ones. The Classic gateway's external IP address can't be kept: an HA VPN gateway gets two new external IP addresses from different pools, and the peer must support BGP.

Practice domain 4 →Practice all domains →