DRAG DROP (Drag and Drop is not supported)
Drag and drop the descriptions from the left onto the related OAuth-defined roles on the right.
Select and Place:

See Explanation section for answer.

Answer(s): A

Explanation:

Show Answer Next Question

QUESTION: 66

Refer to the exhibit. This cURL POST request creates an OAuth access token for authentication with FDM API requests.

What is the purpose of the file “@token_data” that cURL is handling?

This file is given as input to store the access token received from FDM.
This file is used to send authentication-related headers.
This file contains raw data that is needed for token authentication.
This file is a container to log possible error responses in the request.

Answer(s): C

Show Answer Next Question

QUESTION: 67

What is the result of a successful OAuth2 authorization grant flow?

The third-party service is provided with a token that allows actions to be performed.
The user has the application rights that correspond to the user's role within the application's database.
The application is provided with a token that allows actions on services on the user's behalf.
The user has administrative rights to the application's backend services.

Answer(s): C

Show Answer Next Question

QUESTION: 68

In the three-legged OAuth2 process, after the authorization server presents a form to the resource owner to grant access, what is the next step?

The resource owner authenticates and optionally authorizes with the authorization server.
The user who owns the resource initiates a request to the OAuth client.
If the resource owner allows access, the authorization server sends the OAuth client a redirection.
A form to allow or restrict access is submitted by the owner of the resource.

Answer(s): D

Show Answer Next Question

QUESTION: 69

Which OAuth mechanism enables clients to continue to have an active access token without further interaction from the user?

refresh grant
JWT
password grant
preshared key

Answer(s): A

Show Answer Next Question

Download PDF Next Page

Viewing page 30 of 78
Viewing questions 146 - 150 out of 507 questions

Post your Comments and Discuss Cisco 350-901 exam prep with other Community members:

Comments:

Name:

350-901 Exam Discussions & Posts
350-901 AI Verifications

Alex on: March 02, 2025
Great material
UNITED KINGDOM

careerchangerc on: May 31, 2026
This exam was harder than I expected but still manageable with the help of this website's resources.
Latvia

netgurusteve on: July 14, 2026
This exam was harder than I thought but still okay thanks to the practice questions on this website.
France

chiaradevsec on: May 01, 2026
The exam was harder than I thought with topics beyond what I prepped for. Still the practice questions were spot on and helped a lot.
Finland

scriptsandcerts on: June 03, 2026
I wasn't sure about this exam at the start but practice questions helped a lot. Now I feel like I've learned quite a bit through this website.
Latvia

AnsibleAndy on: July 05, 2026
The AI Assistant showed me the real exam questions and braindumps were surprisingly useful. Passing this exam felt very hard but I'm glad it's over.
India

CloseEnoughPass on: July 14, 2026
This exam was very hard but the exam dumps helped me focus on the right topics. Real exam questions in the dumps made all the difference in understanding what was being asked.
Egypt

xCertx on: July 02, 2026
Took two attempts but the AI Assistant and braindumps helped tackle this challenging exam. Wasn't sure of passing until the real exam questions matched what I studied.
United States

NetwrkKing on: June 01, 2026
It felt like I was making slow progress during my prep but the review time was super important. Passed in the end so happy with that result.
United Kingdom

lunchbreaklabs on: June 01, 2026
I took my time with this exam because progress was slow at first. Review time helped a lot and I passed in the end.
Sweden

omaritpro on: July 01, 2026
Very challenging exam underestimated it at first. Needed brain dumps and the AI Assistant to finally pass.
Argentina

sophiedevops on: May 10, 2026
I thought the exam would be easier but with the practice questions from this site I managed to pass.
France

linuxlover99 on: May 10, 2026
Crammed through this cert using the website’s AI assistant to save time and money but made it through eventually.
Israel

StudyGrind99 on: August 04, 2026
Between work and kids finding time to study was tough. Got through short sessions with practice questions and finally feel the relief of having this exam done.
Saudi Arabia

fatimanetsec on: May 14, 2026
Phew this exam was rough. The practice questions here really did their job though.
China

deskjobdreamer on: June 25, 2026
At first I wasn't so confident but the practice questions on this website really helped me get comfortable with this cert.
New Zealand

careerchangerc on: June 29, 2026
I was stressed about this exam but the prep material from this website was on point. Feel a bit more confident now.
Canada

sidehustlesysadmin on: June 17, 2026
Fitting in some exam prep between projects was tough, but I managed to push through slowly. Passed the exam thanks to the AI assistant on this site guiding me with what to focus on.
United States

liamsecops on: May 02, 2026
Spent nights with braindumps and the AI Assistant for this very hard exam. Barely scraped through but it feels like a weight's off my shoulders.
Netherlands

chiaradevsec on: July 01, 2026
This exam was intense. The practice questions here made it more manageable.
Malaysia

PaperCCNA on: July 24, 2026
Work's been crazy and I could only sneak in quick study sessions here and there. The AI assistant on this site really got me through to pass the exam finally some relief.
UAE

anyacloudarch on: June 03, 2026
This exam was tough but the prep questions on this site really got me through.
Australia

jasonhelpdesk on: July 29, 2026
The exam was tough but the prep on this site kept everything organized for me.
South Korea

bootcampgradb on: May 15, 2026
At first I wasn't confident at all but the practice questions on this site made me learn step by step.
Indonesia

HybridCloudH on: May 05, 2026
This exam was really stressful. Glad I had the prep material from this site to keep me on track.
Vietnam

harutodevops on: July 29, 2026
This exam was no joke. Thankful the practice questions here made it easier to handle.
United States

CoffeeAndCerts on: July 02, 2026
Honestly the exam had way more tricky parts than I prepared for. The AI assistant on this site helped fill in some gaps but it was still a tough ride.
Vietnam

katesecplus on: June 13, 2026
This exam was really stressful. The practice questions on this site made a difference though.
Hong Kong

SplunkSam on: June 10, 2026
I didn't feel confident about the exam at first, but working through the practice questions on this website made a big difference. The prep helped me see where I needed to focus.
Kenya

overtimeomar on: June 21, 2026
Man this exam was super stressful. The prep material here got me through.
Germany

GCPengineerT on: May 09, 2026
Trying to squeeze study sessions between work was a challenge but the practice questions made it easier. It felt so good to finally complete this exam and not have to worry about it anymore.
Latvia

TheCertMachine on: July 06, 2026
Spent a lot of time on braindumps but real exam questions were a lot tougher than expected and caught me off guard.
Romania

jamestechuk on: July 15, 2026
Barely scraped through this exam by relying heavily on brain dumps and real exam questions. The challenging exam seemed endless and left me completely drained.
Ukraine

PassedIt2025 on: July 07, 2026
I moved slow at first with the practice questions. Spent a lot of time just reviewing everything over and over. Managed to pass so it worked out in the end.
Pakistan

vlanjockey on: June 16, 2026
This was a very hard exam and the brain dumps helped a lot to fill gaps in my study. Spent a lot of late nights reviewing the dumps to pull through and I am relieved it's done.
Bangladesh

WindowsWizard on: July 14, 2026
After long days at work I squeezed in short study sessions for this exam. Passing felt like a big relief and these questions were pretty spot on so that helped.
China

WrongAnswerRight on: June 18, 2026
This exam was very hard and the braindumps were my last resort after numerous attempts. The AI Assistant helped understand real exam questions better.
Hong Kong

Shadreck Banda on: August 13, 2026
This is great
Bear, United States

ahmedcertkings on: July 22, 2026
The practice questions on this website saved me lots of time and money while I crammed slowly through the exam.
Mexico

laidoffleveled on: June 21, 2026
I wasn't very confident initially but the practice questions on this website really helped me understand the concepts better.
United States

syslogsam on: July 07, 2026
Real exam questions were tricky and it was a challenging exam but the brain dumps helped me eventually get throgh it.
Argentina

ServerlessSid on: July 23, 2026
The exam was tougher than I expected but the practice questions on this site helped make it manageable.
Ukraine

AzureNinja on: June 07, 2026
Juggling work and study isn't easy, but the AI assistant on this site was helpful when I squeezed in prep during lunch breaks. Happy to say I passed the exam after a slow grind.
Spain

katesecplus on: July 08, 2026
I felt the exam was tougher than I anticipated and covered more ground than expected. But having those practice questions definitely helped me understand the concepts better.
Kenya

CiscoFanJ on: May 11, 2026
The exam was harder than I thought but the practice questions on this website made it manageable.
Egypt

ExamWarrior on: June 24, 2026
It took a while to see progress with this exam prep. I spent extra time reviewing each topic thoroughly. Finally passed after all the effort.
Taiwan

homelabhero on: June 30, 2026
This exam was hard but the prep questions here really helped me pass.
Pakistan

weisyseng on: July 10, 2026
At first I was making slow progress but the practice questions helped a lot. Took my time to review everything thoroughly. Passed in the end so it all paid off.
Thailand

miguelcloudops on: July 01, 2026
The progress was slow at first but steady with the practice questions. I took my time reviewing and it paid off in the end. This website helped me pass this exam eventually.
Taiwan

OracleCertV on: May 11, 2026
Juggling work and family life made studying tough but I squeezed in some time with practice questions here and there. Managed to pass the exam so it worked out fine in the end.
Belgium

AI Explanation - verified by User2367 on: July 22, 2026
Question 4:
Q4 asks: Which permission is required to create, delete, refresh or activate a Sandbox?

The correct permission is Manage Sandboxes (the official Salesforce name; the option in the list is written as “Manage Sandbox,” which is effectively the same idea but the canonical name uses the plural).

Why: Sandboxes are a lifecycle feature. Only users with the explicit Manage Sandboxes permission can perform create, delete, refresh, or activate actions.

Why the other options aren’t correct:

- Sandbox Editor, Environment Manager are not the standard permissions to control sandbox lifecycle. - System Administrator is a broad profile level capability; you still rely on the explicit Manage Sandboxes permission to perform sandbox tasks.
Tip: If you see a singular vs. plural naming in the choices, rely on the Salesforce permission name in the docs: “Manage Sandboxes.” You can verify this in the Salesforce Help article sf.data_sandbox_manage.htm.
Mckinney, United States

AI Explanation - verified by User6405 on: July 29, 2026
Question 51:
Here’s a focused explanation of question 51.

Context: You’ve enabled auditing in Dataverse and need to trim audit data to free storage. The question asks which deletion options you can use.

What the options typically mean:

- by record: delete audit logs for specific records (per-entity instance). - between two specified dates: delete audit logs that fall within a date range (a window you specify). - by table older than a specified date: delete all audit logs for a given table that are older than a date (per-table purge). - by column: not a standard deletion criterion for auditing. Auditing is managed at the table/record level, not by a column.

About the provided answer (A, C, D): The list includes “by column,” which isn’t a typical or supported way to purge audit data. The more plausible trio is A (by record), B (date range), and C (by table older than date). If the exam expects A, C, D, that seems incorrect or mis-stated in the source.

Practical takeaway: To reduce audit storage, use per-record deletions, per-table purge by date, and/or date-range deletions as supported by the environment’s audit management UI or admin docs.

If you’d like, I can pull the official Microsoft docs to confirm the exact supported deletion criteria for audit data.
Frankfurt Am Main, Germany

AI Explanation - verified by User3976 on: July 24, 2026
Question 15:
Question 15 is about creating a single dataset that contains only metadata for 100 CSV files in a folder, without loading the actual CSV data. The three actions are performed in this order:

Step 1: From Power BI Desktop, select Get Data, choose Folder, enter the folder path, and click OK, then choose Transform data to open Power Query Editor.

- Why: The Folder connector lists every file in the folder and exposes per-file metadata.

Step 2: In Power Query Editor, expand the Attributes column.

- Why: Expanding Attributes reveals per-file metadata (e.g., name, extension, date modified, size) that you can use to build a metadata-only view.

Step 3: From the Power Query Editor, select the Content column and use Home > Combine Files (aka “Combine binaries”) to consolidate the files.

- Why: This action creates a single query that represents all files. If your goal is metadata-only, you load only this final query and disable loading for intermediate steps so the actual CSV data isn’t stored.
Key concept: using the Folder connector plus expanding Attributes and combining binaries lets you derive a single, metadata-focused dataset from many files without pulling in their contents.
Penge, United Kingdom

AI Explanation - verified by User5259 on: August 04, 2026
Question 1:

Answer: False. Agents and managers CAN create knowledge articles from Community questions.

Why: In ServiceNow, a Community question can be converted into a Knowledge article using the Create Knowledge (or Convert to Knowledge) action. This allows what customers asked to become a reusable article for others.

How it works (high level):

- Open the Community question. - Use the Create Knowledge option to generate a new article in a knowledge base. - Add or adjust metadata (category, keywords) and publish after review/approval.

Permissions: This ability depends on having the right roles/permissions (e.g., agents or knowledge managers). If you don’t see the option, check your role assignments or the knowledge-management configuration.

Quick take: If a question in Community is valuable for others, a trained agent can turn it into a Knowledge article, making it available for self-service.

Hyderabad, India

AI Explanation - verified by bill.dailey on: July 26, 2026
what does RPO ARO and MTTR stand for

RPO (Recovery Point Objective) — maximum acceptable data loss, measured in time. It specifies how far back in time you can tolerate data loss. Example: RPO of 2 hours means backups must occur at least every 2 hours.

ARO (Annualized Rate of Occurrence) — expected frequency of a threat occurring per year. It estimates how often a risk is likely to happen on average (used in risk calculations).

MTTR (Mean Time To Recover) — average time to recover/restore service after an incident. It includes detection, repair, and restoration time, and helps gauge downtime and alignment with RTO.

Virginia Beach, United States

AI Explanation - verified by User4614 on: July 24, 2026
Question 27:
Question 27 asks which solution lets you host an ML model and serve predictions in production without managing infrastructure.

Answer: A. Use Amazon SageMaker Serverless Inference to deploy the model.

Why: SageMaker Serverless Inference hosts the model and serves real-time predictions without you provisioning or managing servers. It scales automatically and handles deployment details.

Why the others aren’t suitable:

- CloudFront: a content delivery network; not for hosting or serving ML model predictions. - API Gateway: exposes APIs but requires underlying compute (e.g., Lambda or SageMaker) to actually run the model; not a true serverless model-hosting solution by itself. - AWS Batch: designed for batch processing, not real-time inference.

Quick example flow: web app ? SageMaker Serverless Inference endpoint ? model ? response.

Tokyo, Japan

AI Explanation - verified by User9126 on: July 22, 2026
Question 9:

Correct answer: B — A reservation for the server based on the MAC address.

Why this is the best choice:

- A DHCP reservation ties a specific IP address to the server’s MAC address. The server will always receive the same IP while still using DHCP, which is easy to manage and avoids manual static configurations on the server. - This helps prevent IP conflicts and allows DNS/DHCP integration to stay consistent (e.g., dynamic DNS updates you might rely on with DHCP).

Why the other options are less suitable:

- A) A separate scope for the file server using a /32 subnet — A /32 subnet is not practical for IPv4 subnets in a typical DHCP scope; separate scopes aren’t needed just for a single server. - C) A static IP address within the DHCP IP range — This can lead to IP conflicts if another device is assigned the same IP; better to keep a reservation inside the DHCP scope or use an address outside the DHCP pool. - D) SLAAC for the server — SLAAC is IPv6 auto-configuration and doesn’t align with a DHCP-managed IPv4 environment (and servers usually require stable addresses regardless of IPv6 specifics).

Quick takeaway: Use a DHCP reservation to give the file server a fixed IP while still benefiting from centralized DHCP management.

Frankfurt Am Main, Germany

AI Explanation - verified by User2505 on: August 10, 2026
Question 35:
Correct answer: A — Generative pre-trained transformers (GPT)
GPT models are designed to understand and generate human language. They can translate an employee’s natural-language request—such as “Show total sales by region for last quarter”—into a structured SQL query.
Why the others are incorrect:

Residual neural network (ResNet): Primarily used for computer vision tasks such as image classification.

Support vector machine (SVM): Commonly used for classification and regression, not natural-language query generation.

WaveNet: Designed mainly for generating or modeling audio and speech.

The key requirement is natural language to SQL generation, which is a generative language-model task. In a production system, the generated SQL should also be validated and access-controlled before execution to prevent unsafe or unauthorized database queries.
London, United Kingdom

AI Explanation - verified by User5630 on: July 29, 2026
Question 16:
Question 16 asks how to make the web app send telemetry at a fixed sampling rate.

Correct answer: B — From the code repository, modify Startup.cs (in ASP.NET Core you configure the Application Insights SDK in code, not with ApplicationInsights.config).

Why the other options aren’t correct:

- Modify ApplicationInsights.config: used in older ASP.NET apps, not typical for ASP.NET Core. - From AppInsights1, modify Usage/cost settings: this doesn’t change sampling behavior. - From AppInsights1, configure Continuous export: export settings affect data export, not sampling.

How to implement fixed-rate sampling (ASP.NET Core):

- In your Startup/Program code, configure the telemetry pipeline to use fixed-rate sampling and disable adaptive sampling. - Example (ASP.NET Core 6+ style after WebApplication.CreateBuilder):
var builder = WebApplication.CreateBuilder(args); builder.Services.Configure<TelemetryConfiguration>(telemetryConfiguration => { var builder = telemetryConfiguration.DefaultTelemetrySink.TelemetryProcessorChainBuilder; double fixedSamplingPercentage = 10; // 10% sampling builder.UseSampling(fixedSamplingPercentage); }); builder.Services.AddApplicationInsightsTelemetry(new ApplicationInsightsServiceOptions { EnableAdaptiveSampling = false }); var app = builder.Build();
- This sets a fixed sampling rate (e.g., 10%) for telemetry sent to AppInsights1.

Key concep
Mumbai, India

AI Explanation - verified by User7866 on: July 20, 2026
Question 52:

Correct answer: D.

Why: In a service offering, a service action is the activity performed by the service provider to deliver or support the service. A service desk agent provides support to a user is the actual action.

Why the others aren’t service actions:

- A: A mobile phone enabling remote work is a product/asset, not an action. - B: A password allowing a user to connect to WiFi is an access credential. - C: A license allows installation of software and is an entitlement, not an action.
Kuching, Malaysia

AI Explanation - verified by User4159 on: July 24, 2026
Question 57:

Answer: C — Spot Instances

Why: The workload is a test environment that can be interrupted and does not need to run continuously. Spot Instances let you bid on unused EC2 capacity at a deep discount, making them the most cost-effective option for interruptible workloads.

Why not the others:

- On-Demand Instances: Flexible but more expensive over time; not the most cost-efficient for interruptible tests. - Reserved Instances: Require a long-term commitment (1–3 years) and are not suited for short-lived or interruptible testing. - Dedicated Instances: Ensure physical isolation and can be costlier; unnecessary for a test environment.

Quick tips:

- Plan for interruptions (checkpoint progress, save state). - Use orchestration to handle interruptions (fallback to On-Demand if spot capacity is unavailable, or use Spot Fleet/Auto Scaling to diversify instance types). - Be aware of the two-minute interruption notice and design the test to finish or checkpoint before termination.
Gdansk, Poland

AI Explanation - verified by User2505 on: August 10, 2026
Question 50:
Correct answer: A — Purchase Provisioned Throughput for the custom model.
A model customized or fine-tuned in Amazon Bedrock cannot generally be invoked using the standard on-demand capacity used by foundation models. The company must purchase Provisioned Throughput, which allocates dedicated inference capacity for the custom model.
Why the other options are incorrect:

SageMaker endpoint: Not required because the model is being used through Amazon Bedrock.

SageMaker Model Registry: Used for managing SageMaker models, not for making a Bedrock custom model available.

Grant access in Amazon Bedrock: IAM permissions may still be needed, but granting access alone does not provide the required inference capacity.

London, United Kingdom

AI Explanation - verified by User4875 on: July 24, 2026
Question 8:
Answer: C: Patch the physical infrastructure that hosts the EC2 instances.

In the AWS shared responsibility model, AWS manages the security of the “infrastructure,” including the physical hardware and underlying virtualization.

The customer is responsible for everything else in the cloud stack: the guest OS on EC2, the NoSQL database software running on it, data, patching the OS, and configuring access controls like security groups.

So, for a self-managed NoSQL database on EC2, AWS would handle the physical hardware, while you would patch the guest OS and manage DB availability, backups, and security.

Mexico City, Mexico

AI Explanation - verified by User4873 on: August 04, 2026
Question 16:
Here’s the reasoning for Question 16.

What the dataflow options mean:

- A/B: Use DirectQuery mode. Queries hit the on-prem SQL Server every time a visual is rendered (and through the dataflow). Even with the Enhanced Dataflows Compute Engine, this keeps online processing and query latency higher, which hurts render times. - C: Use Import mode and schedule a daily refresh. Data is loaded into the Power BI service (cached in memory), so visuals don’t rely on live queries. This minimizes online processing and typically speeds up visuals. A daily refresh can bring in data up to the previous day, meeting the “current year up to yesterday” requirement. - D: Import mode with hourly refresh via Power Automate adds frequent refreshes, increasing processing overhead and isn’t needed to meet the stated goals.

Why C fits best:

- The goal is to minimize online processing and render times. Import mode achieves that by caching data in the service. - To include data up to yesterday, schedule a daily refresh after the end of the day, ensuring the data reflects the current year up to the previous day. - The gateway is still needed to refresh from on-prem sources, but among the options, Import with daily refresh best satisfies all three requirements.
Answer: C
Kuala Lumpur, Malaysia

AI Explanation - verified by studentagain200 on: August 12, 2026
Question 235:
Correct answer: A. Wiping
When laptops are decommissioned, simply deleting files does not remove the underlying data. Deleted files may still be recoverable using forensic tools.

Wiping securely overwrites the drive so the original data cannot reasonably be recovered. It is generally less expensive than physically destroying the device.

Recycling could expose the data if the drive is not securely erased first.

Shredding physically destroys the drive and is effective, but usually more costly and prevents reuse.

Deletion only removes file references; it is not a secure sanitization method.

For the exam, interpret “wiping” as secure data sanitization, not merely deleting files or performing a quick format.
Vitória, Brazil

AI Explanation - verified by User9198 on: July 27, 2026
Question 6:
Question 6 asks which option is a perspective that includes foundational capabilities of the AWS Cloud Adoption Framework (AWS CAF). The AWS CAF defines six perspectives: Business, People, Governance, Platform, Security, and Operations. Each perspective has foundational capabilities to guide cloud adoption.
Why the others aren’t CAF perspectives:

Sustainability is not a CAF perspective (it's not one of the CAF perspectives).

Performance efficiency and Reliability are pillars of the AWS Well-Architected Framework, not CAF.

Therefore, the correct answer is C: Governance. It is a CAF perspective and covers governance-related capabilities like policy, decision rights, risk, and cost governance.
Busan, South Korea

AI Explanation - verified by User1935 on: August 02, 2026
Question 3:
Question 3 asks whether subscribing to the Premier support plan would allow you to get a formal assessment of your Azure environment’s design from Microsoft, given that Basic doesn’t support this.

Answer: No.

Why:

- The Basic plan does not include architecture assessments. - Premier provides enhanced support (e.g., faster response, TAM access, proactive guidance), but a formal “design assessment” from Microsoft is typically a paid Microsoft Consulting Services engagement rather than a standard entitlement of the Premier support plan. - So, even with Premier, you wouldn’t automatically get a formal design assessment under the plan, which is why the solution does not meet the goal.

If you need such an assessment while controlling costs, you’d generally pursue a paid MCS engagement or another option outside of the basic Premier entitlement.

Melbourne, Australia

AI Explanation - verified by User8413 on: July 18, 2026
Question 4:
Question 4 presents a scenario where R&D bypasses the VPN to access services. The correct answer is Shadow IT (Option D).
Why:

Shadow IT describes employees using unsanctioned tools or methods to access systems, often without IT oversight (e.g., bypassing the VPN).

Not A (Espionage): this implies external covert actors; here the activity is internal.

Not B (Data exfiltration): there’s no explicit data transfer shown, just bypassing the VPN.

Not C (Nation-state attack): this involves external political/strategic actors; not the internal behavior described.

Key concept: Shadow IT occurs when users adopt unapproved apps, devices, or access methods, creating risks like uncontrolled data flows and weak visibility.
Helpful follow-ups:

Implement policy and training to enforce approved tools.

Use IAM, VPN enforcement, and network access controls to ensure only sanctioned access.

Monitor for unsanctioned tools and provide secure, approved alternatives.

Giza, Egypt

AI Explanation - verified by User8155 on: July 23, 2026
ok, it is a way to define user type of datatypes?
Tak — domena (DOMAIN) to sposób na zdefiniowanie user-defined typu danych, który opiera sie na istniejacym bazowym typie i narzuca na niego ograniczenia (constraints) oraz domyslna wartosc.
Co to oznacza:

Domeny tworza restrykcjony razem z typem. Wykorzystujesz je w kolumnach, aby miec spójna walidacje danych w wielu miejscach.

To nie jest pelny niestandardowy typ z wlasnym input/output funkcjonowaniem. Aby stworzyc typ z wlasnym I/O i operacjami, uzylbys CREATE TYPE (lub rozszerzen w C). Domeny to uproszczona forma ograniczen na istniejacy typ.

Przyklad:

CREATE DOMAIN positive_int AS integer NOT NULL CHECK (VALUE > 0);

CREATE TABLE orders (id serial, qty positive_int);

Zalety domen:

latwo wymuszaja wspólne reguly dla wielu kolumn/naglówków,

zmiana regul w jednym miejscu (domena) automatycznie wplywa na wszystkie kolumny ja uzywajace.

Warsaw, Poland

AI Explanation - verified by User9198 on: July 28, 2026
Question 101:

The correct option is A: Data architecture.

Why: In the AWS CAF, the Platform perspective covers the technical foundation and architecture that enable cloud workloads to run reliably and efficiently. This includes how data is modeled, stored, and accessed across the enterprise—the core of data architecture.

Quick mapping for the other options:

- Data protection aligns more with the Security perspective (focus on encryption, access control, etc.). - Data governance aligns with the Governance perspective (policies, compliance, data ownership). - Data science isn’t primarily a Platform capability; it’s more related to analytics/business outcomes and may touch other perspectives.
Busan, South Korea

AI Explanation - verified by User4486 on: August 12, 2026
Question 7:
Correct answer: A — When they are working interactively with a small amount of data
A single-node cluster has only one node, which performs both the driver and compute work without separate worker nodes. It is suitable for:

Interactive development and exploration

Small datasets

Lightweight transformations or testing

Avoiding the overhead and cost of multiple workers

Why the other options are less appropriate:

Automated reports: Jobs may benefit from job clusters or SQL warehouses.

Working with SQL in Databricks SQL: Databricks SQL warehouses are the relevant compute option.

Automatically scaling for larger data: A multi-node cluster with autoscaling is more suitable.

Large amounts of data: Distributed worker nodes are generally needed.

The answer key’s A is correct.
Mumbai, India

AI Explanation - verified by nanrany on: July 29, 2026
Question 265:
The correct answer is C: Answering billing and account inquiries.
Reasoning:

The AWS Support Concierge is a specialized team within Enterprise Support whose primary role is handling billing and account questions and helping with account setup and getting started.

They are not the source for architecture guidance or development assistance; those services are provided by Technical Account Managers (TAMs) or Solutions Architects as part of Enterprise Support.

For technical support cases, customers typically interact with AWS Support engineers or the TAMs, not the Concierge. The Concierge may route or coordinate related requests, but they don’t provide the technical resolution.

Summary:

What Concierge does: billing and account inquiries, onboarding help, coordination with AWS support.

What Concierge does not do: provide architecture/design guidance or in-depth technical support case resolutions.

Busan, South Korea

AI Explanation - verified by User1410 on: August 10, 2026
Question 3:
The answer key shows A and E, but E is incorrect as written because it says to publish the agent before enabling Responsible AI filters. The question specifically asks what to do before making the agent available.
The better answers are:

A. Configure prompt modifications to enforce tone, disclaimers, and refusal behavior at the system level.

- This applies the required behavior consistently across topics and channels. - It avoids manually adding disclaimers to every topic.

D. Configure Power Platform DLP policies to restrict unauthorized data connectors.

- DLP policies provide environment-level data protection and help prevent unauthorized data movement through connectors. - They should be configured before deployment.
Why the other options are wrong:

B: Manual topic-by-topic disclaimer edits are specifically discouraged.

C: Channel settings do not replace agent-level Responsible AI controls.

E: Publishing before enabling safety filters violates the stated deployment sequence. Responsible AI filters should be enabled before publication, not afterward.

Lagos, Nigeria

AI Explanation - verified by User4280 on: July 22, 2026
Question 6:
Question 6 asks how TCP and UDP establish a connection between two endpoints.
Correct answer: A — TCP uses the three-way handshake, and UDP does not guarantee message delivery.
Why:

TCP is connection-oriented. It establishes a connection before data transfer using a three-way handshake (SYN, SYN-ACK, ACK), enabling reliable delivery, ordering, and flow control.

UDP is connectionless. It sends datagrams without a prior connection and provides no guarantee of delivery, order, or protection against duplicates; any reliability is up to the application.

Why the other options are incorrect:

B incorrectly states UDP uses acknowledgment packets; UDP has no built-in ACKs.

C wrongly claims UDP provides reliable transfer and that TCP is connectionless.

D misstates framing bits; neither protocol uses the listed TCP/UDP handshake bits in frames.

Takeaway: TCP = reliable, connection-oriented with a handshake; UDP = best-effort, connectionless.
San Luis Obispo, United States

AI Explanation - verified by seymourpro1 on: July 27, 2026
Question 17:
Question 17 asks how to configure a single Opportunity object to support three different teams with separate fields and stages.
Key Salesforce concepts:

Sales Processes define the set of Opportunity stages for a given workflow.

Record Types let you have different data models (different fields and picklist values) on the same object.

Page Layouts control which fields are shown/required for each record type.

Why the right approach is:

Create three Sales Processes so each team can use its own stages.

Create three Record Types so each team has its own fields and picklist values.

Create three Page Layouts so each record type shows the appropriate fields.

With three teams (sales teams + services for installation), you map each team to a unique combination of sales process, record type, and page layout.
Answer: C
Boulogne-Billancourt, France

AI Explanation - verified by User9085 on: July 21, 2026
Question 17:
Question 17 asks how to classify unlabeled customer data into tiers for an ad campaign. The correct choice is B) Unsupervised learning.

Why: The data is unlabeled, so you can’t train a model with known input-output pairs. Unsupervised learning discovers structure in the data by grouping similar customers into clusters or tiers without labels.

What it does: Clustering and segmentation (e.g., k-means, hierarchical clustering, DBSCAN, Gaussian mixtures) to form meaningful groups.

Why the other options aren’t as suitable:

- Supervised learning requires labeled data to map inputs to known outputs. - Reinforcement learning learns through trial-and-error with a environment and rewards, not for static clustering. - RLHF involves learning from human feedback to shape outputs, not for unlabeled clustering.
If later you obtain labeled tiers, you could switch to supervised learning or semi-supervised methods.
Kwai Chung, Hong Kong

AI Explanation - verified by User7882 on: July 16, 2026
Question 1:

The question asks which Azure support plans allow you to open new support requests.

Key concept:

- Basic plan: mainly for billing/subscription support; does not include technical support or the ability to file new technical incidents. - Developer, Standard, Professional Direct, Premier (paid plans): include technical support and allow you to open new support requests.

Therefore, the correct choice is the plan(s) other than Basic (i.e., Developer, Standard, Professional Direct, Premier). If the provided key lists only one option as correct (A), it may be incomplete or misleading—the official plans page shows that Basic does not meet the requirement, while the paid plans do. See: https://azure.microsoft.com/en-us/support/plans/

Noida, India

AI Explanation - verified by logupolo on: August 03, 2026
Question 18:
Here’s what Question 18 is getting at and what the difference means.

What the goal is: You want to add guest (external) users to your Azure AD tenant contoso.com, one entry per external user in a CSV.

The proposed solution uses New-AzureADUser for each external user. Why this isn’t correct: New-AzureADUser creates a regular (internal) user in your directory. It does not create a guest/B2B user invite. So you’d end up with 500 ordinary accounts, not guest accounts with external access.

The correct approach: Use New-AzureADMSInvitation for each external user. This invites the external user to your directory and creates a guest user (a B2B collaboration user) in your tenant. The invited user then accepts the invitation and gains access as a guest.

Key terms and differences:

Azure AD vs directory/tenant: Azure AD is the cloud identity service. A directory (also called a tenant) is your specific instance of that service (contoso.com in your scenario). The terms are often used interchangeably in exams.

Guest user vs internal user: A guest user is external to your organization; an internal user is created directly in your directory as a normal employee/user account.

B2B invitation: The process of inviting external users so they can collaborate in your resources; New-AzureADMSInvitation is designed for this.

If you’re scripting this, you’d typically loop your CSV and call New-AzureADMSInvitation for each email. The general idea is: “invite” (guest) vs “create a new local user” (internal).
Chennai, India

AI Explanation - verified by User9707 on: July 26, 2026
Question 18:
Answer: B — S3 Versioning
Explanation:

S3 Versioning stores all versions of an object. If an object is overwritten or deleted, you can recover the previous version using its version ID.

This protects against accidental deletion or overwriting because the original data remains accessible as an older version.

How to enable:

- Console: bucket > Properties > Versioning > Enable. - CLI: aws s3api put-bucket-versioning --bucket <BUCKET> --versioning-configuration Status=Enabled.

Practical notes:

- There is extra storage cost for preserving multiple versions. - You can manage old versions with lifecycle rules (e.g., expire or transition to cheaper storage). - Versioning alone doesn’t prevent intentional deletions by users with permissions; combine with appropriate IAM/bucket policies if needed.
Other options:

S3 Lifecycle rules: automate moving/expiring data, not specifically protection against deletion.

S3 bucket policies: control access but don’t inherently preserve previous data.

S3 server-side encryption: protects data at rest, not against deletion/overwrite.

Álvaro Obregón, Mexico

AI Explanation - verified by sophielargeaudproust on: July 26, 2026
Question 97:
The correct answer is D.
Why: After deploying an AI model, documentation should capture the model's architecture, performance metrics, and updates. This supports transparency, reproducibility, and ongoing maintenance.
Why the other options are less appropriate as the core deployment documentation:

A (training data, hyperparameters, stakeholder feedback) focuses on training provenance; these details are important but are typically part of pre-deployment/training records, not the primary post-deployment documentation.

B (encountered issues, monitoring process, risks) relates to governance and monitoring rather than the core documentation you’d publish about the deployed model itself.

C (licensing information, the team, business goals) covers governance context but not the technical traceability and performance-based record needed for ongoing maintenance.

Saint-Priest, France

AI Explanation - verified by User6887 on: July 21, 2026
Question 28:
Correct answer: D
Explanation:

CAPWAP (Control And Provisioning of Wireless Access Points) creates a tunnel between the AP and the Wireless LAN Controller over the IP network. The AP and WLC communicate via this CAPWAP tunnel, which can traverse Layer 2 or Layer 3 networks.

The AP does not have to be on the same switch as the WLC. It can connect to any switch in the network, as long as there is IP reachability between the AP and the WLC. This is why option D is correct.

The other options are incorrect:

- A: Connecting the AP to the wired network won’t inherently create a loop; loops are prevented by proper network design. - B: The AP does not need to be on the same switch as the WLC. - C: The AP does not need a direct copper link to the WLC.
Notes: CAPWAP uses UDP for control and data channels (commonly on ports 5246/5247) and is designed to allow centralized management across a distributed campus.
Thiruvananthapuram, India

AI Explanation - verified by User3880 on: July 21, 2026
Question 36:
Waterfall is a traditional, linear project management and systems development approach. Key points:

It is a sequence of distinct phases that flow one after another (requirements ? design ? implementation ? verification ? maintenance).

Requirements are defined up front and changes are costly; the project often requires formal sign-off before moving to the next phase.

It emphasizes heavy documentation and minimal iteration after a phase is completed.

In the context of question 36, the BA is using a waterfall approach to capture requirements. The next step before building the solution is to obtain approval and sign-off on the requirements from stakeholders. This ensures everyone agrees the requirements are accurate, complete, and feasible, and authorizes the build to proceed, helping prevent scope creep and miscommunication.
Buenos Aires, Argentina

AI Explanation - verified by User7958 on: August 12, 2026
Question 25:
Correct answer: D — use Amazon SQS between two Lambda functions.
The architecture becomes:

Ingestion Lambda receives requests from API Gateway and places messages in an SQS queue.

Processing Lambda is triggered by the queue and loads records into Aurora PostgreSQL.

Why this works:

Decoupling: API request handling is separated from database loading.

Buffering: SQS absorbs bursts when incoming traffic exceeds the database’s processing capacity.

Controlled concurrency: Configure the SQS-triggered Lambda’s concurrency and batch size to avoid overwhelming Aurora or exhausting database connections.

Automatic retries: Messages that fail processing can return to the queue and be retried. A dead-letter queue can capture persistent failures.

Lower operational effort: The solution remains serverless and avoids managing EC2 servers or Kubernetes clusters.

The answer key’s D is correct.

A: EC2 and Tomcat add server-management overhead.

B: DynamoDB/DAX would require changing the database model and does not solve the Aurora ingestion design.

C: SNS is useful for notifications and fanout, but SQS is better for durable buffering and backpressure in this producer-consumer workflow.

Busan, South Korea

AI Explanation - verified by User1509 on: August 07, 2026
Question 103:
Question 103 is a VPN configuration simulation. The required settings are:

Strongest encryption algorithms: Choose the strongest available encryption and integrity options, typically AES with the largest key size and a secure integrity algorithm.

Encrypt all traffic: Select tunnel mode rather than transport mode. Tunnel mode encrypts the entire original IP packet, which is appropriate for a site-to-site VPN between branch networks.

Secret-password authentication: Configure a pre-shared key (PSK) on both VPN concentrators. The key must match at both ends.

The answer key shows A, but the screenshots are unavailable here, so I cannot verify which specific controls correspond to A. The important concepts are strong algorithms + tunnel mode + matching pre-shared key. Also ensure the same compatible VPN parameters are configured on both concentrators; otherwise, the tunnel will not establish.
Miami Beach, United States

AI Explanation - verified by User1935 on: August 02, 2026
Question 7:
Question 7 asks which Azure web tier plan meets these requirements:

Custom domains

10 GB storage per web app

Each app runs on dedicated compute

Load balancing between instances is included

Costs minimized

Answer: A (Standard)

Why:

- Custom domains: All paid tiers support custom domains, but you still need features like dedicated compute and load balancing. - Storage: Free/Shared have very limited resources; 10 GB per app exceeds what those tiers realistically provide. Standard offers ample storage for these apps. - Dedicated compute: Free and Shared run on shared compute. Basic and Standard provide dedicated compute, with Standard being the lowest tier that meets the other requirements. - Load balancing: Built-in load balancing across instances is included in Standard (and higher). Free/Shared don’t meet this production-style load-balancing expectation. - Cost: Basic is cheaper than Standard but lacks one or more required capabilities (notably robust load balancing and sufficient storage). Standard is the lowest tier that satisfies all the criteria.

Note: In App Service, multiple apps in the same plan share compute resources. If you truly need isolation per app, you’d typically use separate App Service Plans, which increases cost beyond a single Standard plan.

Melbourne, Australia

AI Explanation - verified by User6589 on: August 03, 2026
Question 8:
Question 8 asks which improvement opportunity found at baseline can contribute to process maturity and strengthen the customer’s overall security posture.

Post-Incident Review (A) is the after-action step used to capture lessons learned, identify process gaps, and drive improvements across the incident response lifecycle. This directly supports process maturity and a stronger security posture.

Incident Analysis (D) focuses on investigating root causes during/after an incident, which is important but more of a diagnostic activity than a standalone baseline improvement to maturity.

Fast Eradication (B) and Incident Containment (C) are tactical actions aimed at stopping the incident, not on maturing the process.

Note: The page lists D as the answer and the reference link seems unrelated to this question, which suggests an error in the provided key. Based on typical ServiceNow security incident objectives around continuous improvement, Post-Incident Review (A) is the better fit.
Hyderabad, India

AI Explanation - verified by User1887 on: July 22, 2026
Question 20:

The correct answer is Mass Transfer Tool.

Why:

The Mass Transfer Records tool is Salesforce’s built-in way to reassign ownership of existing records in bulk. In this scenario, you need to transfer ownership of the departed rep’s accounts and opportunities to a new rep while keeping the team structure intact.

How to use it (high level):

Go to Setup > Data Management > Mass Transfer Records.

Choose the object: Accounts (and then Opportunities, if needed).

Set “Transfer From” (the leaving rep) and “Transfer To” (the new rep).

Decide whether to transfer related records as well (e.g., opportunities tied to accounts) and proceed.

Review and confirm.

Why not the other options:

Data Loader / Data Import Wizard are for bulk import/export, not bulk reassigning ownership.

Dataloader.io is a third-party tool; the built-in Mass Transfer tool is the standard Salesforce method for this task.

Milan, Italy

AI Explanation - verified by User9656 on: August 05, 2026
Question 38:
Yes.
Why: The problem describes a column with case-sensitive data in a DirectQuery model, where undefined values and errors can arise from case mismatches during joins or lookups. By normalizing casing (convert all values to a single case, e.g., all upper or all lower) you ensure consistent key values across the data model. This reduces mismatches and the resulting undefined values or errors.
How to implement:

In the source query: apply UPPER() or LOWER() to the column.

In Power Query Editor: use Text.Upper or Text.Lower on the column, then load.

Notes:

This fixes the root cause when comparisons/joins depend on case.

In DirectQuery, transformations are pushed to the source or query layer; ensure the transformation is supported by your data source/connector.

After normalization, verify relationships and any downstream visuals that rely on that column.

Lagos, Nigeria

AI Explanation - verified by sangeethasoftengg on: July 22, 2026
Question 43:
Question 43 is asking what must be built before using snippets to create region-specific unsubscribe footers.

Correct answer: A. Email templates

Why:

Snippets are reusable content blocks that you insert into email content, and they require an email template (or at least an email draft) to render.

You can store region-specific footers as snippets and include them in a template, but there would be nothing to render the snippet into without a template.

The other options aren’t prerequisites for using snippets:

- Tokens are placeholders you can use in templates/snippets, but you still need a template. - Segmentation determines who receives what, not whether you can use snippets. - Smart Lists are for grouping leads, not enabling snippet usage.
Tip: To implement per-region footers, create a snippet per region and insert them into an email template (potentially using segmentation or dynamic content to decide which snippet shows).
Coimbatore, India

AI Explanation - verified by User6732 on: August 12, 2026
Question 5:
Correct answer: B. operators
The requirement is for a user who can view nonsensitive switch configuration but should not make configuration changes. On AOS-CX, the predefined operators role provides this type of read-only access and uses privilege level 1.
Why the other choices are less suitable:

administrators — Full access to switch configuration and REST API functions; excessive for this requirement.

sysadmin — Intended for highly privileged system administration, not basic monitoring.

helpdesk — More limited and focused on troubleshooting/support functions; it does not match the stated default role and level-1 requirement as directly.

The key clues are “view nonsensitive configuration information” and “level 1.” Those identify the operators role.
Sydney, Australia

AI Explanation - verified by User6346 on: July 16, 2026
Question 22:
Question 22 asks if the given solution meets the goal of reducing admin effort for deploying a large number of VMs weekly (Windows Server 2016 and Ubuntu Linux) and then removing them in the same week.

The proposed solution: use Microsoft Managed Desktop.

Why this does not meet the goal:

- Microsoft Managed Desktop is aimed at managing end-user devices (Windows 10/11) for enterprises, not at provisioning or life-cycle management of Azure VMs. - It does not inherently automate creation/deletion of large VM fleets or support mixed OS images like Windows Server 2016 and Linux.

A better fit would be Azure DevTest Labs (or similar lab services), which is designed to:

- create and manage many VMs for development/testing, - support both Windows and Linux images, - provide templates, auto-shutdown/start, and easier lifecycle management to reduce admin effort.
Answer: No. The MS Managed Desktop approach does not align with the VM lifecycle automation needs; DevTest Labs would be a better choice for this scenario.
Bengaluru, India

AI Explanation - verified by User9085 on: July 21, 2026
Question 18:
Question 18 asks which foundation model type is best for a search app that handles text and image queries. Correct answer: A) Multi-modal embedding model.

Why: A multi-modal embedding model maps both text and images into a shared vector space, enabling cross-modal similarity and retrieval (e.g., find images matching a text query or text matching an image).

What it does: Enables efficient similarity search using a vector database or nearest-neighbor indexing.

Why the other options aren’t suitable:

- Text embedding model handles only text, not images. - Multi-modal generation model generates content, not primarily for retrieval. - Image generation model creates images, not for embedding-based search.

Quick implementation note: you’d typically pair the embeddings with a vector store (e.g., ANN index) and possibly fine-tune on domain data for better alignment across modalities.

Kwai Chung, Hong Kong

AI Explanation - verified by User3892 on: July 31, 2026
role iam veut dire quoi
En bref: un rôle IAM est un ensemble d’autorisations que peut « assumer » une entité AWS (utilisateur, service, ou même un autre compte) pour accéder à des ressources, sans utiliser de clés d’accès à long terme.
Points clés:

Le rôle n’a pas de clés propres. Des identifiants temporaires (via STS) sont émis lorsque quelqu’un ou quelque chose « assume » le rôle.

La trust policy déclare qui peut assumer le rôle (par ex. une instance EC2, un utilisateur IAM, ou un service comme Lambda).

Une fois attaché à une ressource (par exemple comme un instance profile sur une EC2), l’application peut appeler les services AWS avec les permissions du rôle.

Pourquoi c’est utile:

Respect du principe du moindre privilège: on accorde uniquement ce qui est nécessaire.

Pas de stockages de clés sur les instances; les credentials se renouvellent automatiquement.

Cas d’usage:

Rôles pour EC2, rôles pour Lambda, accès cross-compte, etc.

Comment l’utiliser rapidement: 1) Créer le rôle et attacher les policies de permissions. 2) Définir la trust policy pour l’entité qui l’assume (par ex. EC2). 3) Attacher le rôle à la ressource (ex. faire un instance profile pour EC2). 4) L’application utilise les credentials fournis par le rôle via le SDK/CLI.
Bamako, Mali

AI Explanation - verified by User9156 on: July 24, 2026
Question 29:

The correct answer is D: All of the above.

Why:

Snowflake Web Interface (UI): You can create and manage Virtual Warehouses from the UI (resize, suspend/resume settings, etc.).

SQL commands: You can create and alter warehouses with DDL like CREATE WAREHOUSE and ALTER WAREHOUSE.

Data integration tools: ETL/ELT tools can create/manage warehouses by connecting to Snowflake and issuing the appropriate commands or using their config to specify a warehouse.

If you’d like, I can show quick example commands or steps for each interface.
Chennai, India

AI Explanation - verified by sophielargeaudproust on: July 26, 2026
Question 11:

Correct answer: D (Special category data)

Why:

- The system uses behavioral biometric data (typing speed, rhythm, pressure) to create a unique user profile. - When biometric data is used to uniquely identify individuals, it is treated as a special category data under privacy regimes like GDPR due to its sensitivity and impact on privacy. - Because this data can uniquely identify a person, it triggers higher safeguards and stricter processing requirements.

Quick notes:

- If the data were purely non-identifying or anonymized, the classification could be different, but here its purpose is to identify/authenticate a specific person. - Practical implications: ensure a lawful basis for processing, implement data minimization, strong security, and consider additional protections (e.g., explicit consent where required).
Saint-Priest, France

AI Explanation - verified by User3759 on: August 05, 2026
Question 6:
Question 6 answer: B — a data pipeline
You need to copy data from:

Source: on-premises SQL Server Database1

Destination: Fabric Warehouse Warehouse1

Connectivity: on-premises data gateway

A data pipeline is designed for orchestrating data movement between sources and destinations. Its Copy Data activity can use the on-premises gateway to access SQL Server and load the data into the warehouse.
Why the other options are weaker:

Dataflow Gen1: Primarily for Power Query transformations and isn’t the best Fabric-native choice for this copy operation.

KQL queryset: Queries data; it does not orchestrate copying from SQL Server.

Notebook: Could perform custom code-based movement, but it is unnecessary and less direct than a pipeline.

Exam cue: “Copy data from one system to another,” especially with a gateway or scheduled movement ? data pipeline.
Fredericton, Canada

AI Explanation - verified by User8744 on: August 12, 2026
Question 11:
Question 11: Correct answer — C
> Set expectations and agree on the requirements that the contractor will deliver with periodic progress reports.
The project manager should delegate the work clearly while retaining appropriate oversight. This means:

Define the contractor’s deliverables, responsibilities, acceptance criteria, and deadlines.

Agree on a reporting cadence to monitor progress.

Allow the local company to use its expertise and manage its own work.

This empowers the contractor without abandoning control of the project.
Why the others are weaker:

A: “Under the supervision of the project manager” suggests excessive direct control and is less empowering.

B: Micromanaging prevents the contractor from functioning independently and is inefficient.

D: Deploying the contractor’s staff under the PM’s supervision creates a staffing/control arrangement rather than empowering the contracted company.

The exam principle is delegate outcomes and monitor performance, rather than personally directing every task.
Johor Bahru, Malaysia

AI Explanation - verified by User4053 on: August 10, 2026
Question 52:
Correct answer: C — AWS Security Token Service (AWS STS).
AWS STS provides temporary security credentials—an access key, secret key, and session token—with:

A limited lifetime

Permissions defined by an IAM role or policy

The ability to access AWS APIs without using long-term IAM user credentials

An application can use STS operations such as AssumeRole to obtain credentials for a specific, restricted task.
Why the others are incorrect:

Amazon API Gateway — Builds and manages APIs; it does not generally issue AWS credentials for applications.

IAM users — Usually use long-term credentials, which are less suitable for applications.

IAM instance profiles — Provide temporary role credentials to applications running on EC2 instances, but are specifically tied to EC2. The broader requirement described is best met by AWS STS.

Itapira, Brazil

AI Explanation - verified by User2829 on: August 08, 2026
Question 3:
??????????????????Copilot???????????????????????·????????????
?:

Copilot??????????

???????????????

??????????

?????????????

???????????????????

??????????????????????Copilot??????????????????????????????????????????????????????????????????
??????????????????????

Prompts and Suggestions: Retained for 28 days

User Engagement Data: Kept for Two Years

?????????C??2?????????????????????????????????????GitHub??????????????????????????????????????????????????????
Minato-Ku, Japan

AI Explanation - verified by User9518 on: August 10, 2026
Question 63:
For Question 63, the answer key says B: User1 only.
The key concept is SharePoint Online external sharing authentication:

A recipient who already has a Microsoft account or an organizational account can authenticate by signing in.

A recipient without a usable Microsoft identity may receive a one-time passcode by email.

The passcode is sent to the same email address that was invited and is used to verify ownership of that address.

Therefore, based on the identities and guest settings shown in the missing exhibit, User1 is the only recipient who requires email passcode authentication. User2 and User3 can authenticate through existing supported identities instead.
The images containing the exact guest settings and user table aren’t visible here, so I can’t verify every identity-specific detail. But the rule to remember is: email one-time passcodes are used for external recipients who cannot authenticate with an existing Microsoft account or organizational identity, when the tenant’s guest settings allow that method.
Hyderabad, India

AI Tutor 👋 I’m here to help!

Cisco 350-901 Exam Actual Questions Developing Applications using Cisco Core Platforms and APIs (DEVCOR) (Page 30 )

QUESTION: 65

Explanation:

QUESTION: 66

QUESTION: 67

QUESTION: 68

QUESTION: 69

Cisco 350-901 Exam Actual Questions
Developing Applications using Cisco Core Platforms and APIs (DEVCOR) (Page 30 )