Technical Lab: Choose between regional and cross-region load balancers
Questions
Question 1 — Multiple Choice
A company has critical applications deployed in three Azure regions: East US, West Europe, and Southeast Asia. The requirement is that, in case of complete failure of an entire region, traffic should be automatically redirected to healthy regions without manual intervention.
Which Azure Load Balancer tier meets this requirement?
A) Regional Azure Load Balancer with Standard SKU, configured with load balancing rules between VMs in different availability zones within the same region
B) Azure Load Balancer with Basic SKU in internal mode, manually replicated between regions via automation scripts
C) Azure Cross-Region Load Balancer, which operates at the global tier and routes traffic between regional backends distributed across multiple regions
D) Azure Application Gateway with WAF enabled, configured with backends in multiple regions
Question 2 — Technical Scenario
An architect is designing the load balancing solution for an internal corporate application. Users are all located in a single Azure region, and the requirements include support for high availability ports (HA Ports) to balance multiple protocols simultaneously on network virtual appliances (NVAs).
The architect proposes the following configuration:
Type: Azure Load Balancer
SKU: Standard
Mode: Internal
Rule: HA Ports enabled
Scope: Regional (single region)
After reviewing the proposal, a colleague states that it is incorrect because the Cross-Region Load Balancer also supports HA Ports and should be preferred for offering greater resilience. What is the correct assessment?
A) The colleague is right; the Cross-Region Load Balancer supports HA Ports and is always preferable to regional for offering greater availability
B) The colleague is wrong; the Cross-Region Load Balancer does not support HA Ports, and the proposed configuration with the Standard internal regional Load Balancer is adequate for the scenario
C) The colleague is partially right; HA Ports are supported in the Cross-Region Load Balancer, but only in external mode
D) The proposed configuration is incorrect because HA Ports require Basic SKU, not Standard
Question 3 — True or False
The Azure Cross-Region Load Balancer can have individual virtual machines located in different regions as direct backends, the same way a regional Load Balancer references VM NICs.
True or False?
Question 4 — Technical Scenario
An organization runs a high-volume public application with backends in two regions: Brazil South and East US. It currently uses a regional Azure Load Balancer Standard in each region, with separate public IPs. External clients need a single entry point with automatic failover between regions.
The team decides to add a Cross-Region Load Balancer in front of the two regional ones. During planning, the following question arises: do the existing regional Load Balancers need to be replaced or can they be reused as backends of the Cross-Region Load Balancer?
A) The regional Load Balancers need to be replaced by Application Gateway instances, as the Cross-Region Load Balancer only accepts Application Gateway as a backend
B) The existing regional Load Balancers Standard can be reused directly as backends of the Cross-Region Load Balancer, without need for replacement
C) The regional Load Balancers need to be downgraded to Basic SKU before being added as backends of the Cross-Region Load Balancer
D) The regional Load Balancers need to be converted to internal mode before being accepted as backends of the Cross-Region Load Balancer
Question 5 — Multiple Choice
When comparing the regional Azure Load Balancer Standard with the Azure Cross-Region Load Balancer, which of the statements below correctly describes a behavioral difference between the two?
A) The regional Load Balancer operates at OSI model layer 7, while the Cross-Region Load Balancer operates at layer 4
B) The Cross-Region Load Balancer distributes traffic based on latency between regions and preserves the client's source IP to the regional backend, while the regional Load Balancer distributes traffic within a single region using five-tuple hash
C) The Cross-Region Load Balancer supports health probes based on HTTP/HTTPS, while the regional Load Balancer supports only TCP
D) The regional Load Balancer Standard does not support availability zones, while the Cross-Region Load Balancer offers zone redundancy natively
Answer Key and Explanations
Answer Key — Question 1
Answer: C
The Azure Cross-Region Load Balancer is the only Microsoft Layer 4 load balancing service designed to operate globally across regions. It maintains a single global public IP address and, when an entire regional backend fails, automatically redirects traffic to healthy backends in other regions, without need for manual intervention or DNS changes.
Alternative A describes the regional Load Balancer Standard with zone redundancy, which protects against datacenter failure within a region, but not against entire region failure. Alternative D describes Application Gateway, which operates at layer 7 and is not a native multi-region Load Balancer for generic TCP/UDP traffic. Choosing A instead of C would result in absence of automatic inter-regional failover, exposing the application to total unavailability during a regional failure.
Answer Key — Question 2
Answer: B
The Cross-Region Load Balancer does not support HA Ports. This functionality is exclusive to the regional Azure Load Balancer Standard in internal mode. HA Ports allow simultaneously balancing all protocols (TCP, UDP) on all ports with a single rule, an essential scenario for NVAs that need to inspect arbitrary traffic.
Alternative A contains a factually incorrect statement about HA Ports support in the Cross-Region Load Balancer. Alternative C proposes a mode distinction that doesn't exist for this functionality. Alternative D reverses the correct relationship: HA Ports are exclusive to Standard SKU, not Basic. The consequence of adopting the Cross-Region Load Balancer in this scenario would be the inability to configure the necessary rules for the NVAs, making the architecture unfeasible.
Answer Key — Question 3
Answer: False
The Cross-Region Load Balancer does not accept individual VMs as direct backends. Its backend pool is composed exclusively of frontend IP addresses of regional Load Balancers Standard. This layered architecture is intentional: the Cross-Region Load Balancer acts as a global entry point and delegates local distribution to regional balancers, which in turn manage the VMs.
This behavior is a critical distinction from the regional Load Balancer, which directly references VM NICs in the backend pool. Confusing the two models leads to unfeasible architecture designs, where one tries to add VMs directly to the Cross-Region Load Balancer backend and the configuration is rejected by Azure.
Answer Key — Question 4
Answer: B
The existing regional Load Balancers Standard with public IP can be reused as backends of the Cross-Region Load Balancer. The Cross-Region Load Balancer accepts as backend pool members the frontend IPs of regional Load Balancers Standard, regardless of whether they are already in production. This allows the organization to add the global failover layer without redesigning the existing regional infrastructure.
Alternative A describes a non-existent restriction; Application Gateway is not the accepted backend type. Alternative C reverses the compatibility logic: Basic SKU is what is not supported, not Standard. Alternative D confuses the operation mode with the compatibility requirement; the Cross-Region Load Balancer works with public frontends, not internal ones.
Answer Key — Question 5
Answer: B
The Cross-Region Load Balancer uses anycast to direct traffic to the closest point of presence to the client and routes to the appropriate regional backend based on latency and availability. It preserves the client's source IP to the regional backend, allowing lower layers to apply source-based policies. The regional Load Balancer Standard uses a five-tuple hash algorithm (source IP, source port, destination IP, destination port, protocol) to distribute connections among VMs within a region.
Alternative A reverses the layers: both operate at OSI model layer 4. Alternative C is incorrect because both support HTTP/HTTPS and TCP probes. Alternative D is false: the regional Load Balancer Standard has full support for availability zones, this being one of its central characteristics compared to Basic SKU.