The Microsoft AZ-700 (Designing and Implementing Microsoft Azure Networking Solutions) exam is about choosing, implementing and troubleshooting the paths that let cloud workloads communicate securely. A working network diagram is not enough: effective routes, DNS responses, policy decisions and application health must agree with the intended architecture. Microsoft updated the English exam skills on July 27, 2026, so current preparation should follow that version of the blueprint.
Imagine a company with an on-premises data center and two Azure application spokes. A central hub provides hybrid connectivity and inspection. One application needs private database access; another must accept customer HTTPS traffic without exposing its back-end servers. The engineering problem is to prove reachability where permitted, deny unapproved access and maintain a recoverable operational design.
Build an IP and routing plan that survives growth
Begin with nonoverlapping address spaces and explicit subnet functions. For example, reserve separate ranges for hub gateways and security appliances, application spokes, and private endpoints, then validate them against on-premises networks and planned acquisitions. Address overlap can make route propagation and return traffic ambiguous even when a tunnel is up.
VNet peering provides connectivity between networks, but it does not by itself make arbitrary multi-hop transitive routing work. Document which prefixes each network can reach, whether forwarded traffic is allowed, and which user-defined routes, gateway settings and network virtual appliances govern the path. Evaluate both the forward path and the return path; asymmetric routing through inspection often produces intermittent failures that look like application defects.
Build a change sheet showing source subnet, destination subnet, expected next hop and allowed protocol. Use effective routes and observed traffic rather than relying only on a hand-drawn arrow. The detailed Azure NSG, route and Bastion troubleshooting discussion is useful when a subnet appears connected but security and routing rules disagree.
Choose hybrid connectivity for the service requirement
A site-to-site VPN can be appropriate for a bounded requirement or a resilient alternate path. ExpressRoute can provide private network connectivity for workloads with specific availability, bandwidth or network-governance needs. Neither should be selected solely because its name sounds more enterprise-ready. Consider recurring charges, failover, route control, provider dependencies and the time it takes to recover a broken circuit.
For dynamic routes, identify which BGP prefixes the on-premises edge advertises and which routes Azure learns. The existence of a BGP session does not prove every intended network is reachable. Investigate prefix filtering, competing more-specific routes, overlapping addresses and gateway-transit configuration when one branch reaches an Azure spoke and another does not.
Rehearse loss of the preferred connection with the owners of the application and security controls. A backup tunnel might become active yet send flows around required inspection or fail to carry DNS to the private service. The recovery test should cover reachability, security policy, application correctness and the return to the preferred path.
Make private access and name resolution agree
Private endpoints give supported Azure services a private address in the consumer network, but clients must resolve the service name to the intended address and must have an allowed network path. The failure mode is often DNS rather than the endpoint resource itself. For clients in an on-premises branch, design conditional forwarding or another supported resolver path for the private DNS zone.
Test an application server in each location. Query the service’s normal hostname, record the returned address, and confirm the response reflects the required private endpoint rather than an unexpected public path. Then test the TCP/TLS connection and the service’s own authorization. Successful DNS resolution is necessary but not sufficient to prove access.
The Microsoft Azure private endpoints and private DNS article helps distinguish endpoint and zone configuration. Be clear that service endpoints and Private Link solve related but different access problems; neither substitutes for identity permissions, and Private Link is not automatically an organization-wide firewall policy.
Deliver public applications without opening the back end
Choose the application-delivery component based on the traffic: Layer 4 distribution, Layer 7 routing and web controls, or global edge acceleration have different requirements. Azure Load Balancer, Application Gateway and Azure Front Door should be compared by protocol, health probing, TLS handling, regional or global footprint, protection needs and operating cost.
For the public HTTPS application, specify where TLS terminates, how the origin is protected and what happens when one backend becomes unhealthy. A working health probe must reflect application readiness, not merely an open port. Test a deployment that returns a valid HTTP response from the wrong service version; network success alone can disguise an application failure.
Review whether a Web Application Firewall policy is needed, which rules have a demonstrated purpose and how blocked requests are investigated. Monitor false positives through an approval process rather than disabling broad protection after one incident.
Distinguish routing failures from denied traffic
Suppose an application spoke cannot connect to a private database. Use a reproducible query, source identity, destination name, resolved address, port and timestamp. Check effective routes first, then network-security rules, firewall or appliance logs, and the database’s own access control. One failed test does not prove every upstream service is broken.
Virtual network flow logs, Network Watcher tools and connection monitors can provide complementary evidence. The Azure DNS and Network Watcher diagnostics article offers a more focused diagnostic workflow. Do not confuse a blocked flow with a closed application port or incorrect credentials; the corrective change differs.
Write an acceptance matrix: app-spoke A can reach the designated private service over its permitted port; app-spoke B cannot; on-premises administrators can reach approved management endpoints through the governed path; the external customer can reach only the public application entry point. Confirm each allow and deny result after a routing or firewall change.
Align the experiment with the current AZ-700 scope
The current Microsoft AZ-700 study guide covers core network infrastructure, connectivity services, application delivery, private access and network security. Those subject areas form an integrated scenario, not five unrelated lists of product names.
Keep the lab isolated and budgeted. Record the original design, exact configuration change, before-and-after observations and rollback step. If a test improves reachability but breaks a security denial or bypasses required monitoring, it has failed. Practitioners preparing for AZ-700 should learn to explain those trade-offs with evidence rather than treat the credential as a guarantee of professional outcomes.