CompTIA Security+ SY0-701 Practice Test 48
Question 470
A nonprofit's database virtual machine shares a physical cloud host with other customers. Performance measurements show high latency every time a neighboring tenant runs intensive reporting jobs; the provider finds no evidence of unauthorized memory access or isolation failure. Which architecture concern most closely matches the symptom?
- Resource contention from another tenant sharing finite host capacity, sometimes called a noisy-neighbor effect
- Confirmed hypervisor escape allowing the neighbor to read secret database memory
- A host-capacity problem caused solely by the database application’s own unoptimized queries, without shared resource interference
- Guaranteed host-level resource isolation because each customer has a separate cloud virtual-machine identifier
Explanation
Correct Answer
A:Even with correct guest isolation, virtual machines can contend for shared CPU, storage I/O, caches or network capacity on a physical host. The timing correlation with another tenant's reporting workload supports a noisy-neighbor performance problem. The operator can measure host and guest resource limits, request quality-of-service controls or provision dedicated capacity if necessary. This explains service degradation without asserting a confidentiality breach that telemetry does not show. Isolation, availability and performance are related but distinct attributes of shared-tenancy architecture; evidence should guide which one is actually affected. Measure host scheduling and consider isolation or capacity reservations before treating performance degradation as a security breach.
Incorrect Answers
B:Hypervisor escape is a security boundary violation that lets code cross from a guest toward the host or another guest. The case explicitly lacks evidence of such access. Shared resource contention can occur even when virtualization isolation is functioning correctly, so claiming secret-memory theft would overstate the facts. The evidence supports contention for shared resources, not a proven breach of guest isolation.
C:An inefficient query or insufficient instance size can cause similar database latency, and performance testing should consider those alternatives. However, the supplied measurements show slowdowns repeatedly coinciding with a neighboring tenant’s reporting workload on the same physical host. That correlation and the absence of memory-isolation evidence support shared-capacity contention rather than a proven defect limited to the database’s internal query plans.
D:Separate VM identifiers and guest filesystems help distinguish tenants logically, but they do not necessarily reserve independent physical CPU, memory bandwidth or storage throughput. A noisy-neighbor problem can arise when isolation prevents unauthorized memory access but scheduling and I/O resources remain shared. The case reports contention during another tenant’s heavy jobs, so a VM ID alone cannot establish performance isolation.
Question 471
A financial modelling vendor moves a proprietary licensed engine to public cloud. The software license is tied to the number and identity of physical hosts, and auditors also want evidence that selected workloads are not sharing host hardware with other tenants. Which compute allocation should be evaluated first?
- Standard shared virtual machines tagged with the software vendor's license name
- A dedicated physical host offering with documented host allocation and bring-your-own-license rules
- A single-tenant virtual network containing ordinary shared-host VMs, with the license recorded as a resource tag
- Network segmentation rules alone, without any mapping to physical hosts required by the license
Explanation
Correct Answer
B:Dedicated-host offerings allocate identified physical server capacity to a customer and can support certain host-based software licensing and co-tenancy constraints. The firm should verify the cloud provider's licensing eligibility, hardware specifications, host affinity, failover behavior and reporting evidence before committing. Dedicated hosts do not automatically encrypt every workload or replace guest hardening; their value here is the physical placement and accountability required by the license and isolation policy. A dedicated instance may differ contractually from an entire dedicated host, so the exact service terms matter. This option addresses both physical host tracking and reduced shared tenancy directly.
Incorrect Answers
A:Tags help organize billing and inventory, but do not guarantee which physical machine hosts a VM or whether unrelated tenants share it. A host-bound license may require information unavailable from ordinary shared instances. The vendor should not confuse a logical label with physical allocation evidence. Auditors need host identity and co-tenancy evidence that ordinary tags cannot supply.
C:A virtual private network or cloud VPC can isolate routing and access policy for the vendor’s systems. It does not establish that the underlying physical host is reserved for one customer or that a license tied to host identity complies with its terms. Resource tags can improve accounting but cannot substitute for provider-issued evidence of dedicated physical allocation and supported bring-your-own-license rules.
D:Network policies isolate traffic but do not prove which physical hardware executes a VM. Multiple tenants may still reside on the same host behind separate virtual networks. A licensing requirement based on host identity needs a compute placement and reporting mechanism, not merely an IP security-group rule. Both contractual licensing and physical co-tenancy criteria must be verified through the host allocation service.
Question 472
A wastewater plant must transmit real-time temperature readings from aging industrial controllers to a corporate analytics environment. The controllers cannot accept arbitrary commands from corporate networks, and there is no business need for the analytics system to initiate connections into the control zone. Which integration design gives the safest default boundary?
- Use ordinary bidirectional IP routing between the networks and rely on detection logs to identify bad commands later
- Share the same domain-administrator credentials across all controllers and analytics machines
- Permit corporate-origin VPN and MFA access to the controllers while retaining general command functions
- Use a brokered or one-direction telemetry path with explicit industrial-zone separation and no general inbound control access
Explanation
Correct Answer
D:A one-direction or tightly brokered telemetry design lets controllers export read-only measurements without granting the corporate analytics environment a general route for commands. Industrial firewalls, dedicated brokers or hardware data diodes may implement the directionality based on the required reliability and protocols. Define approved senders, destinations, formats and exception processes, and validate that alarm traffic continues during maintenance. This respects the OT safety boundary and reduces pivot opportunities from office malware. Engineers must assess whether acknowledgments or device management require separate carefully authorized channels; that exception should not erase the default separation. The design matches the stated need for outward monitoring, not remote administration.
Incorrect Answers
A:A flat or bidirectional route leaves a command path from general enterprise systems into fragile controllers. Logging can help detect misuse but may not prevent a dangerous command before it changes plant state. The business requirement is outward status reporting, so unnecessary inbound connectivity should be removed at the boundary.
B:Shared administrator credentials dramatically increase the impact of compromise and cross-domain movement. Aging controllers may have weak identity support, making shared secrets even riskier. Identity isolation and limited permissions are needed; common high-privilege accounts across IT and OT do not solve telemetry transport. A compromised office administrator could otherwise gain unnecessary authority over safety-critical plant devices.
C:VPN and MFA can authenticate remote access but do not by themselves constrain what an authenticated enterprise user or compromised endpoint may command. If analytics requires only data, retaining command functions violates least privilege. Strong login can complement a maintenance pathway, but it should not justify open general control interfaces.
Question 473
A hosting provider places application servers and the privileged interfaces of its core switches and hypervisors on the same unrestricted VLAN. A vulnerability in a public web service allows an attacker to execute code on one application host. Which new risk is created by the network design?
- Using SSH instead of Telnet on management interfaces while retaining unrestricted shared reachability
- An exploited application host may directly scan or attack privileged infrastructure-management interfaces
- Requiring 802.1X at the access port while leaving the same unrestricted post-authentication segment
- Using separate hostnames for management devices without changing their IP routing or VLAN exposure
Explanation
Correct Answer
B:Combining management and production traffic gives any compromised app server a network path to high-privilege switch, hypervisor and storage interfaces. The attacker may attempt credential theft, password guessing or management-plane exploits without first crossing a dedicated security boundary. Separate management networks or microsegmentation rules, tightly constrained jump hosts and strong administrator authentication reduce that lateral movement path. An unrestricted VLAN is not made safe simply because management services use encryption. The relevant failure mode is exposure of powerful control interfaces to a lower-trust application tier, which increases blast radius after an application compromise.
Incorrect Answers
A:Encrypted management protocols protect credentials and sessions on the wire, but a compromised application host may still attempt login abuse or exploit the management service directly. SSH does not remove unnecessary reachability. The segment boundary should restrict which dedicated management principals can connect to infrastructure endpoints. An attacker who reaches the management IP can still probe authentication and pre-login service flaws.
C:Port authentication helps determine which device may join a network, but the exploited web server is already an admitted member. If admitted application systems share unrestricted reachability with management interfaces, 802.1X does not stop lateral connections from a compromised legitimate host. Network zoning remains necessary after admission. Separate management segmentation limits what even a fully authenticated but compromised host can reach.
D:DNS names are labels, not access-control boundaries. Management services can still be reached by their IP addresses in a shared VLAN, even if their names use a different suffix. Isolation requires routing, firewall or VLAN policy that actually restricts packets between production and privileged administration planes. Attackers can bypass friendly DNS names and connect to the same management IP addresses directly.
Question 474
A cloud function runs for only three seconds each time an image arrives, yet its execution role grants permissions to delete every storage bucket in the organization. A developer argues that short-lived runtime eliminates the need for IAM scope restrictions. What is the key security issue?
- Set the function timeout to three seconds and allow the existing organization-wide delete permissions to remain.
- Share the overprivileged execution role among all image-processing functions so its permissions are managed centrally.
- Ephemeral compute does not remove the risk of an excessively privileged execution identity
- Enable storage-object versioning and recovery procedures, but retain the execution role’s permission to delete every bucket.
Explanation
Correct Answer
C:Serverless functions authenticate to cloud services using an execution identity whose permissions remain relevant even when each invocation is brief. If input handling or a dependency is compromised, a broad role can delete resources far outside the function's legitimate image-processing task. Scope permissions to specific buckets and actions, separate functions by trust boundary, monitor API calls and avoid unnecessary credential exposure. Ephemeral compute changes lifecycle and operations but is not an IAM security boundary. The architecture should assume that one invocation may be malicious and limit what it can do. This reasoning applies to containers and virtual machines as well: privilege follows identity, not runtime duration.
Incorrect Answers
A:A timeout limits how long one invocation may run, which can help contain resource consumption and runaway processing. A single authorized cloud API request can still delete important resources in far less than three seconds, so short duration is not a substitute for an IAM resource or action boundary. The risk is what the function identity can request whenever it executes, not simply how long its process survives.
B:Central role management can simplify maintenance and audit policy changes across related functions. Reusing one broadly privileged role also extends the destructive permission to more workloads, magnifying the impact of a compromised handler or dependency. Each workload should have the narrow actions and resource scope needed for its business purpose. The short execution lifetime does not reduce the authority that the shared role grants to live invocations.
D:Versioning can help recover overwritten or deleted object versions when storage settings and retention rules permit it. It is a resilience layer, not a guarantee that an excessively privileged function cannot destroy a bucket or supporting configuration. Some deletion operations can undermine recovery paths. The immediate security defect is the execution identity’s organization-wide authority, which should be scoped down even if backups and versioning also exist.
Question 475
A biotechnology firm already runs genomic analytics inside a confidential-computing enclave. Its vault accepts attestation evidence before releasing dataset keys. After a signed platform update, the enclave reports a measurement not yet on the firm's approved policy list, even though the host operator has a valid administrator account and the VM disk is encrypted. Which key-release decision maintains the intended trust boundary?
- Require independently verified attestation of an approved enclave measurement and platform state before releasing the dataset key
- Release keys to any enclave that presents a valid HTTPS certificate for the host operator's domain
- Release keys whenever the VM's disk is encrypted, irrespective of the current enclave measurement
- Allow the host administrator to override failed measurements by requesting the long-term dataset key directly
Explanation
Correct Answer
A:The vault should evaluate trustworthy attestation claims against its own approved measurements and platform policy, not treat the host's general administrator identity as proof that the confidential workload is trusted. A newly changed measurement may be legitimate, but it must be inspected and explicitly approved through a governed update process before decryption keys become available. Remote attestation helps bind key release to an intended protected execution environment and software state; a valid signature alone is not equivalent to acceptance of any measurement. The protected enclave was already selected in this scenario, so the decision being tested is fail-closed attestation and key-release authorization after a platform change, not simply choosing confidential computing as an architecture.
Incorrect Answers
B:HTTPS authenticates a communicating service endpoint when its certificate is properly validated; it does not automatically prove the enclave's measured code, configuration, or hardware state. A host operator could possess a valid transport certificate while running a workload that the data owner has never approved. Transport security remains necessary for the attestation exchange, but the vault must separately evaluate measured workload identity and policy before releasing sensitive genomic decryption material.
C:Encrypted virtual disks protect data at rest when the disk is not legitimately unlocked, but the workload will handle plaintext during computation. A malicious or unapproved runtime configuration could expose that plaintext despite encrypted storage. Disk encryption therefore cannot replace the separate evidence of the enclave's current measured state. The key vault should deny release until the modified measurement is assessed and included in the owner's authorized attestation policy.
D:Giving a cloud host administrator direct access to the long-term dataset key defeats the firm's stated requirement to reduce dependence on host-operator privilege. An administrator credential is not a substitute for measured trusted workload state; it may also have been compromised or overprivileged. Emergency recovery procedures can exist, but must use separately approved, auditable controls rather than bypassing attestation merely because a platform update changed an expected measurement.
Question 476
An MSP exposes one Internet-facing bastion with password login and accepts management connections from any source IP. Once connected, a technician can RDP to every production server; the gateway records only successful sign-ins, not privileged activity. The customer requests an improved design without removing emergency support. Which combined change addresses the major weaknesses?
- Add MFA to the bastion but retain Internet-wide entry and unrestricted access to every internal server
- Disable all bastion logging so administrative sessions consume less disk capacity
- Allowlist bastion ingress addresses but still permit every authenticated technician to access all production hosts
- Restrict entry points and downstream host/port scopes, require strong authentication and retain monitored privileged-session records
Explanation
Correct Answer
D:A secure bastion acts as a narrowly scoped administrative gateway rather than a general-purpose bridge into production. Restrict source entry, enforce phishing-resistant or strong MFA, authorize downstream RDP only to the specific target systems and time windows, and record privileged session activity for review. Use separate administrative identities and revoke access promptly after an emergency. Logging only sign-in success misses commands and lateral movement inside the session, while unrestricted RDP creates a large blast radius. The complete design preserves needed support but limits who can enter, where they can go and how their actions are investigated.
Incorrect Answers
A:MFA makes stolen passwords harder to reuse, but a bastion open to the entire Internet still faces unnecessary probing and exploitation. Once logged in, broad internal RDP routes allow compromise to spread across every server. MFA is valuable as one layer, not a replacement for narrowly defined access paths and audit trails.
B:Removing logs would weaken attribution, incident investigation and oversight for privileged support. Bastions are high-value control points, and monitoring should record session identity, destination and meaningful administrative actions subject to privacy and retention policies. Storage performance concerns can be handled through log management rather than eliminating evidence. Audit evidence must identify the technician, target server and consequential privileged operations.
C:Restricting inbound addresses reduces exposure, but any authorized or compromised technician would still reach all production servers. A single jump host should not confer blanket lateral movement rights. Downstream policies need per-customer, per-role and per-session limits aligned to specific maintenance tasks. Downstream target restrictions prevent one compromised bastion session from traversing the entire server estate.
Question 477
A digital identity startup runs a customer document processor beside untrusted plugin code submitted by third parties. An architect suggests ordinary Linux containers offer exactly the same kernel separation as full virtual machines, so the workloads can share a host without additional analysis. What technical correction should the security lead make?
- Containers normally have fully independent guest kernels identical to those in separate virtual machines
- Virtual machines do not rely on any hypervisor or virtualization boundary
- Container packaging eliminates the need to authenticate users and authorize document access
- Standard containers share the host kernel; separate VMs can provide stronger guest-kernel separation for differing trust levels
Explanation
Correct Answer
D:Ordinary containers typically share a host operating-system kernel, even though namespaces, seccomp and other controls constrain their processes. Full VMs run their own guest kernels separated by a hypervisor, which can provide a stronger boundary for workloads that must distrust each other. That does not make VMs invulnerable, nor does it mean containers are inherently inappropriate. The startup should evaluate the plugin threat model, host hardening, isolation technology, sensitive data handling and the possibility of sandbox escapes. Distinguishing the shared-kernel boundary from separate guest kernels is essential before colocating untrusted extensions with private customer documents.
Incorrect Answers
A:Linux containers isolate processes using mechanisms such as namespaces and control groups, but generally invoke the same underlying host kernel. A kernel flaw may therefore threaten several containers on a host. Calling them independent guest kernels misrepresents the boundary when untrusted plugins and sensitive documents have different threat models. That distinction matters most when the host runs code provided by untrusted external plugin developers.
B:VM isolation is normally provided by a hypervisor that mediates guest access to underlying hardware. There are differences among virtualization technologies, but it is incorrect to claim that full VMs have no relevant hypervisor boundary. That layer is one reason guest-kernel separation can reduce some shared-kernel risks. The hypervisor mediates execution boundaries between guests and the underlying host hardware.
C:Container deployment defines execution packaging and resource boundaries; it does not decide who may access customer files or what the processor may do with them. Application authentication, authorization, filesystem rights and secrets governance remain essential. Stronger runtime isolation cannot replace per-customer data controls. Authorization checks are still required whenever a processor opens documents belonging to different customers.
Question 478
A retail chain routes branch traffic into a cloud application through a centralized firewall. After an engineer adds a new peering connection, traceroutes show branch-to-application traffic going directly across the peer link without the expected inspection hop. The peering status shows connected and all apps still work. Which network check addresses the new security gap?
- Inspect actual route selection, propagation and policy-based steering to restore the required inspection path
- Treat successful peering status as proof that every packet still traverses the original firewall
- Attach a stricter inspection policy to the centralized firewall without changing the new peer-route preference
- Test only that the application responds, without checking the next hops used by branch traffic
Explanation
Correct Answer
A:Routing decisions follow effective route tables, prefix selection and propagation behavior rather than an intended diagram. The new peering path may take precedence over the inspection route even though reachability improves. Engineers should check forwarding and return paths, effective routes, transit limits and firewall insertion controls, then validate with representative flows. Correct the path without introducing asymmetric routing that breaks stateful inspection. A connected peer does not guarantee security-policy compliance. The underlying decision is whether traffic traverses the enforced inspection boundary, not whether two networks can communicate. Inspect asymmetric return paths too, since stateful gateways depend on coherent flow handling.
Incorrect Answers
B:Connection status confirms a relationship between networks, not the full path chosen for traffic. The observed traceroute contradicts the assumption that centralized inspection remains in place. Relying on peer status alone ignores how propagated or more-specific routes can bypass an intended security gateway. Connected peering can change effective next hops without preserving a mandatory security inspection service.
C:Improving a firewall rule set is helpful only for packets that actually traverse that firewall. After peering was added, traceroutes show branch application traffic taking a direct bypass path. An inspection policy that no longer sees the traffic cannot enforce the intended control. The architect must correct route selection or steering first, then verify that the inspected flow really crosses the required boundary.
D:Application reachability is a functionality test and may succeed when traffic has bypassed required security controls. The chain specifically needs proof of path enforcement, including next-hop and return routes. A simple successful HTTP response cannot establish that firewall inspection still occurs as intended. Inspect both forward and return traffic to establish that required controls remain in the path.
Question 479
A remote solar-powered pumping station records safety telemetry every second. The WAN connection to headquarters can disappear for up to six hours; local operators must continue to see alarms, and central analytics must receive the missing measurements once the link returns. Which design satisfies both resilience and data-history integrity?
- Run local alarm processing with durable timestamped queues and store-and-forward synchronization after connectivity returns
- Block local alarms until each measurement is acknowledged by the distant analytics service
- Erase sensor messages after the first failed upload so queues never fill during an outage
- Forward the missing measurements when the WAN reconnects, but omit their original timestamps and any deduplication identifier
Explanation
Correct Answer
A:Edge processing keeps safety decisions available during a WAN outage, while durable local storage retains timestamped events for later replay. After reconnection, the station can forward buffered records with sequence identifiers, deduplication and integrity checks so central analytics reconstructs the missing period without double counting. Capacity planning should account for the maximum expected outage and physical storage failure modes. The architecture must also keep alarms locally actionable rather than dependent on round-trip cloud response. This balances operational continuity and historical completeness, with authenticated encryption protecting transfers when the link is restored.
Incorrect Answers
B:Waiting for a remote acknowledgment creates a single point of failure in the WAN. During a lengthy outage, sensors might stop reporting and local alarms could be delayed when operators most need them. Remote analytics is useful, but real-time safety actions should not depend on a temporarily unavailable central service.
C:Deleting unsent events makes the station appear available while permanently losing the history needed for safety review and central analysis. Buffer management should instead be designed for expected outage duration, with controlled retention and alerts on storage exhaustion. Silent discard undermines both resilience and integrity of the reporting record. Safety auditors may need the missing interval even if local alarms kept running during the WAN failure.
D:Delayed backfill helps central analytics catch up after a disconnected period, so this proposal addresses one requirement superficially. Yet replaying readings without original timestamps or sequence identifiers makes it difficult to reconstruct real event order or distinguish a retransmission from a new measurement. The station needs durable ordered telemetry and local alarms during the outage, followed by a reliable store-and-forward synchronization process.
Question 480
A hospital's prescription API calls an internal inventory microservice over an apparently private subnet. The security team expects an attacker to compromise some internal hosts eventually, so it requires workload identity and encrypted requests instead of granting trust from network location alone. Which design is most suitable?
- Permit calls based solely on source subnet address and assume internal IP placement identifies the calling service
- Use HTTP without authentication because traffic remains inside the data-center perimeter
- Use mutual TLS with distinct service certificates, validate both peers and enforce narrow API-level authorization
- Store one permanent shared password in both services and skip service-specific identity checks
Explanation
Correct Answer
C:Mutual TLS allows both communicating services to present and validate distinct cryptographic identities, while encrypting traffic in transit. Service identity should be bound to short-lived or rotated certificates and checked against an explicit authorization policy for the inventory API. This reduces reliance on the assumption that every host inside the subnet is trustworthy. Network allowlists can still limit exposure, and the application should validate request permissions even after mTLS succeeds; identifying a service does not authorize every business operation. A compromised certificate requires revocation or rotation controls, so identity lifecycle management remains part of the design.
Incorrect Answers
A:Subnets identify routing context and can constrain connectivity, but a compromised device or misconfigured workload may communicate from an approved range. IP allowlists alone do not authenticate a specific microservice or prevent eavesdropping on internal traffic. They can supplement workload identity but cannot replace it. Service certificates distinguish workloads even when two hosts occupy the same allowed address range.
B:Private networks are useful segmentation controls, yet insiders and compromised systems may observe or alter unencrypted data in some environments. The prescription workflow handles sensitive details and requires identity validation independently of physical network placement. Plain HTTP supplies neither transport confidentiality nor strong peer authentication. An internal adversary could intercept prescription information on the private segment if it remains plaintext.
D:A common long-lived secret may authenticate a request if protected and applied correctly, but reuse obscures which service principal acted and magnifies compromise when one configuration leaks. It also does not inherently encrypt the transport. For separate workloads with strict identity requirements, independently managed mTLS identities and authorization provide stronger separation.