Skip to content

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

Task 4.4: Configuring Network Connectivity Center.

Hybrid spokes on VPN tunnels and VLAN attachments, site-to-site data transfer, router appliances and solving transitivity problems.

Study it

  • Network Connectivity Center for hybrid networks: hybrid spokes, site-to-site data transfer, router appliances and transitivity

    Lesson coming

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

In a Network Connectivity Center hub, the workload networks app1 and app2 are VPC spokes and use global dynamic routing. The routing network transit, which isn't a VPC spoke and uses regional dynamic routing and legacy best path selection, holds VLAN attachments in us-east4 that are attached to the hub as hybrid spokes. The data center learns the app1 and app2 subnets. VMs of app1 in us-east4 reach the data center, but VMs of app1 in us-west1 have no route to it. Which change fixes this?

  1. AAdd transit to the hub as a VPC spoke, so that the global dynamic routing of the VPC spokes applies to the attachments' routes
  2. BSet includeImportRanges to ALL_IPV4_RANGES on the hybrid spokes of the VLAN attachments
  3. CSwitch transit to standard best path selection so that its routes are shared with every region
  4. DSwitch transit's dynamic routing mode from regional to global
Show the answer and why
  • AAdd transit to the hub as a VPC spoke, so that the global dynamic routing of the VPC spokes applies to the attachments' routes

    Incorrect

    A routing network doesn't have to be a VPC spoke for hybrid and VPC spokes to connect, and the regions follow transit's own routing mode.

  • BSet includeImportRanges to ALL_IPV4_RANGES on the hybrid spokes of the VLAN attachments

    Incorrect

    That setting makes the data center learn the VPC spokes' subnets, which it already does; it doesn't decide where routes toward the data center are programmed.

  • CSwitch transit to standard best path selection so that its routes are shared with every region

    Incorrect

    NCC hybrid spokes support only the legacy best path selection mode of Cloud Router.

  • DSwitch transit's dynamic routing mode from regional to global

    Correct

    With transit in regional mode, the NCC dynamic routes from its hybrid spokes are programmed only in us-east4; in global mode they are programmed in all regions.

When VPC spokes and hybrid spokes share a hub, the NCC dynamic routes from the hybrid spokes are programmed in the VPC spokes according to the dynamic routing mode of the routing VPC network that holds the hybrid spokes: regional mode keeps them in the region of each hybrid spoke, global mode programs them in all regions. The routing network doesn't need to be a VPC spoke, hybrid spokes need legacy best path selection, and includeImportRanges governs the other direction, the advertisement of VPC spoke subnets toward the data center.

Question 2 · choose 1

A network team wants to attach a branch office's VPN connection to a Network Connectivity Center hub as a hybrid spoke. The branch currently connects through a Classic VPN tunnel. What must the team use for the spoke?

  1. AThe existing Classic VPN tunnel, added to the hub as a VPN spoke
  2. BA Router appliance spoke that points to the Classic VPN gateway
  3. CHA VPN tunnels, added to the hub as a Cloud VPN spoke
  4. DA VPC spoke for the network that holds the Classic VPN gateway
Show the answer and why
  • AThe existing Classic VPN tunnel, added to the hub as a VPN spoke

    Incorrect

    Classic VPN tunnels aren't supported as NCC spoke resources.

  • BA Router appliance spoke that points to the Classic VPN gateway

    Incorrect

    Router appliance spokes are VMs running a third-party appliance, not Classic VPN gateways.

  • CHA VPN tunnels, added to the hub as a Cloud VPN spoke

    Correct

    NCC VPN spokes use HA VPN tunnels; Classic VPN tunnels aren't supported.

  • DA VPC spoke for the network that holds the Classic VPN gateway

    Incorrect

    A VPC spoke exchanges subnet routes; it doesn't turn the Classic VPN tunnel into a hybrid spoke.

NCC hybrid spokes are built from Cloud Interconnect VLAN attachments, HA VPN tunnels or Router appliance VMs. Classic VPN tunnels aren't supported, so the branch needs HA VPN tunnels before it can join the hub as a VPN spoke.

Google Cloud documentation

Question 3 · choose 1

A team deploys SD-WAN router appliance VMs as Network Connectivity Center Router appliance spokes. Following Google's guidance for these spokes, how should the VMs and their BGP sessions be laid out?

  1. ATwo or more VMs in different zones, each peering with two redundant Cloud Router interfaces
  2. BTwo VMs placed together in the same zone, each peering with two redundant Cloud Router interfaces
  3. CTwo VMs in different zones, both peering with one shared Cloud Router interface
  4. DOne VM with two network interfaces, each running its own BGP session
Show the answer and why
  • ATwo or more VMs in different zones, each peering with two redundant Cloud Router interfaces

    Correct

    Google recommends at least two router appliance VMs in different zones, each peering with a pair of redundant Cloud Router interfaces.

  • BTwo VMs placed together in the same zone, each peering with two redundant Cloud Router interfaces

    Incorrect

    VMs in one zone share that zone's failures; the guidance calls for VMs in different zones.

  • CTwo VMs in different zones, both peering with one shared Cloud Router interface

    Incorrect

    Each VM must peer with a pair of redundant Cloud Router interfaces, not with one interface.

  • DOne VM with two network interfaces, each running its own BGP session

    Incorrect

    One VM is a single point of failure; the 99.99% Compute Engine SLA applies to VMs in multiple zones.

Router appliance instances are Compute Engine VMs, so the Compute Engine SLA applies: 99.5% for a single VM and 99.99% for VMs in multiple zones. Google recommends running at least two VMs in different zones, each peering with a pair of redundant Cloud Router interfaces.

Question 4 · choose 1

VPC networks blue and green are VPC spokes of the same Network Connectivity Center hub, which uses the mesh topology. To reach a network that is peered only with green, an engineer proposes also connecting blue and green directly with VPC Network Peering. What should the network engineer say?

  1. AThe peering works and makes green's peered network reachable from blue
  2. BThe peering works, but only after blue is removed from the hub's export filters
  3. CThe peering works, provided both spokes switch to legacy best path selection
  4. DVPC Network Peering isn't supported between two VPC spokes of one hub
Show the answer and why
  • AThe peering works and makes green's peered network reachable from blue

    Incorrect

    VPC Network Peering isn't supported between two VPC spokes that are also connected through the hub.

  • BThe peering works, but only after blue is removed from the hub's export filters

    Incorrect

    Export filters don't make peering between two spokes of the same hub possible.

  • CThe peering works, provided both spokes switch to legacy best path selection

    Incorrect

    Best path selection has no bearing on whether two spokes may peer.

  • DVPC Network Peering isn't supported between two VPC spokes of one hub

    Correct

    The VPC spokes overview lists this as a limitation of VPC spokes.

VPC spokes on one hub already exchange subnet routes through the hub, and NCC doesn't allow VPC Network Peering between two VPC spokes that are also connected using the hub; only two spokes in the edge group of a star topology hub may peer. Reaching a network that is only peered with one spoke needs another design.

Practice domain 4 →Practice all domains →