Professional Cloud Network Engineer · Domain 6: Configuring, implementing and managing a cloud network security solution · 13% of the exam
Task 6.4: Configuring self-managed network virtual appliance and Packet Mirroring.
Multi-NIC appliance VMs between VPC networks, an internal load balancer or policy-based routes for highly available appliances, out-of-band Network Security Integration and Packet Mirroring to self-managed collectors.
Study it
Network virtual appliances, Network Security Integration and Packet Mirroring
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
Firewall appliance VMs in us-central1 sit behind an internal passthrough Network Load Balancer, and a static route 0.0.0.0/0 with that load balancer as next hop sends the VPC network's egress through them. VMs in us-central1 work, but packets from VMs in us-east1 that use the route are dropped. The appliances pass their health checks. What should the network engineer do?
AAdd appliance VMs in us-east1 to the same load balancer's backend service
BEnable global access on the load balancer's forwarding rule
CRe-create the forwarding rule and the route so that symmetric hashing is turned on
DCreate a second static route 0.0.0.0/0 for us-east1 with a network tag and the same next hop
Show the answer and why
AAdd appliance VMs in us-east1 to the same load balancer's backend service
Incorrect
An internal passthrough Network Load Balancer's backends must be in its own region.
BEnable global access on the load balancer's forwarding rule
Correct
With global access disabled, the next hop is reachable only from its own region, and packets from other regions using the route are dropped.
CRe-create the forwarding rule and the route so that symmetric hashing is turned on
Incorrect
Symmetric hashing keeps both directions of a flow on one appliance; it doesn't make the next hop reachable from another region.
DCreate a second static route 0.0.0.0/0 for us-east1 with a network tag and the same next hop
Incorrect
The new route would still point to a next hop that rejects traffic from other regions while global access is off.
A route whose next hop is an internal passthrough Network Load Balancer applies across the VPC network, but without global access the load balancer accepts next hop traffic only from its own region; packets sent from another region are dropped. Enabling global access on the forwarding rule makes the appliances usable from every region.
An appliance VM has nic0 and nic1, both attached to the same VPC network. The team wants a load balancer in that network to send traffic to the appliance on nic1. What should the network engineer point out?
Anic1 can receive load-balanced traffic once it gets its own external IP address
BWhen interfaces share a VPC network with nic0, only nic0 can receive load-balanced traffic
CTwo interfaces of one VM can never be attached to the same VPC network
Dnic1 receives load-balanced traffic only through a policy-based route
Show the answer and why
Anic1 can receive load-balanced traffic once it gets its own external IP address
Incorrect
An external address doesn't change which interface may receive load-balanced traffic.
BWhen interfaces share a VPC network with nic0, only nic0 can receive load-balanced traffic
Correct
The multiple network interfaces page states that only nic0 receives load-balanced traffic in this case.
CTwo interfaces of one VM can never be attached to the same VPC network
Incorrect
Regular VPC networks allow several interfaces in the same network, following the documented rules.
Dnic1 receives load-balanced traffic only through a policy-based route
Incorrect
A policy-based route doesn't override which interface may receive load-balanced traffic.
In regular VPC networks, a VM can have several network interfaces in the same network, but Cloud Load Balancing and static routes have extra rules: when other interfaces share a VPC network with nic0, only nic0 can receive load-balanced traffic.
A security team uses Network Security Integration out-of-band integration to send copies of selected traffic to a partner's inspection appliances. Where does the team define which traffic is mirrored?
AIn a policy-based route whose next hop is the partner's appliances
BIn the VPC network's subnet settings, as a mirroring flag per subnet
CIn mirroring rules of a global network firewall policy
DIn a Cloud Armor security policy attached to the backend services
Show the answer and why
AIn a policy-based route whose next hop is the partner's appliances
Incorrect
Routes steer the original packets; out-of-band integration clones traffic without changing its path.
BIn the VPC network's subnet settings, as a mirroring flag per subnet
Incorrect
Subnets have no mirroring flag; selection is done with mirroring rules.
CIn mirroring rules of a global network firewall policy
Correct
Out-of-band integration uses mirroring rules in the firewall policy to select the traffic that is mirrored.
DIn a Cloud Armor security policy attached to the backend services
Incorrect
Cloud Armor filters requests to load balancers; it doesn't mirror VM traffic.
With out-of-band integration, Network Security Integration Packet Mirroring clones traffic to the producer's appliances according to mirroring rules that you specify in a global network firewall policy, including negative rules to exclude traffic. The original traffic keeps its path.
A team uses VPC Packet Mirroring to send traffic to its own collectors but only needs inbound TCP traffic from 203.0.113.0/24 to the mirrored VMs, to keep the extra bandwidth low. How should the mirroring policy be set up?
AMirror all traffic and drop the unneeded packets on the collectors
BNarrow the policy with filters on protocol, CIDR range and direction
CAdd a firewall rule that denies other traffic so it isn't mirrored
DMirror only egress TCP traffic toward 203.0.113.0/24, because ingress can't be filtered
Show the answer and why
AMirror all traffic and drop the unneeded packets on the collectors
Incorrect
Mirroring everything copies far more traffic than the team needs.
BNarrow the policy with filters on protocol, CIDR range and direction
Correct
Packet Mirroring filters can select by protocol, CIDR ranges, IP version and direction, which narrows the mirrored traffic.
CAdd a firewall rule that denies other traffic so it isn't mirrored
Incorrect
Denying traffic changes what the VMs receive; filters select what is mirrored.
DMirror only egress TCP traffic toward 203.0.113.0/24, because ingress can't be filtered
Incorrect
Mirroring can be set for ingress only, egress only or both.
Instead of collecting all traffic, a Packet Mirroring policy can use filters on protocol, CIDR ranges, IP version and direction (ingress-only, egress-only or both) to limit what is copied.