Design High-Performing Architectures

A financial services company runs a reporting platform on an Amazon Aurora MySQL cluster with one writer and four reader replicas. Two of the readers are large, memory-optimized instances dedicated to heavy analytical queries, while the other two smaller readers serve the customer-facing dashboard. Currently the application uses the built-in reader endpoint, which round-robins connections across all four readers, causing the dashboard to occasionally hit the large analytics instances and analytical queries to land on the small instances, resulting in inconsistent performance. The team wants to ensure dashboard traffic only uses the two small readers and analytics traffic only uses the two large readers, without managing connection logic in application code. What should the solutions architect recommend?

A company runs an internal reporting application used only during business hours, with unpredictable spikes when analysts run ad-hoc queries. Traffic drops to near zero on nights and weekends. The team wants a fully managed relational database that automatically scales compute capacity up and down to match demand and minimizes cost during idle periods, without requiring them to manually resize instances. Which solution best meets these requirements?

A media company serves a global single-page application from Amazon S3 through Amazon CloudFront. The product team wants to insert a custom HTTP security header and modify the response object based on the viewer's country before content is returned to the browser, without changing the origin. The logic must run at CloudFront edge locations with minimal added latency and must not require managing servers. Which solution best meets these requirements?

A media company is migrating 40 TB of files from an on-premises NFS file server to Amazon EFS as part of a data center consolidation. The migration must complete over an existing 1 Gbps Direct Connect link within a weekend, preserve file metadata and permissions, and automatically verify data integrity after transfer. The team wants to minimize custom scripting and operational effort. Which solution best meets these requirements?

A gaming startup is launching a new mobile game with a companion leaderboard service backed by Amazon DynamoDB. Traffic patterns are completely unknown at launch and are expected to swing dramatically—from near-zero during off-peak hours to massive, unpredictable spikes when the game is featured or goes viral. The team wants to avoid throttling during sudden spikes and does not want to spend time forecasting throughput or managing capacity. Which DynamoDB configuration best meets these requirements?

A gaming company runs a leaderboard service on Amazon DynamoDB. Traffic follows a very predictable daily pattern: read and write volume ramps up gradually every morning, peaks in the evening, and drops overnight. The workload is steady and forecastable with no sudden unexpected spikes. The team wants the lowest possible cost while ensuring the table always has enough throughput to avoid request throttling during the daily peaks. Which approach best meets these requirements?

A media company runs a CPU-bound video transcoding service on a fleet of EC2 instances behind an Auto Scaling group. The workload is Linux-based, compiled from open-source code, and consistently drives sustained high CPU utilization with minimal memory usage. The engineering team wants to maximize price-performance and reduce compute cost without sacrificing throughput. Which change should the solutions architect recommend?

A media company runs a nightly video-rendering pipeline that processes thousands of independent frames. Each frame render is a stateless job that can be checkpointed and retried if interrupted. The batch must finish before morning, but the workload is highly parallelizable and completely fault-tolerant. The team wants to minimize compute cost while maintaining enough capacity to meet the deadline. Which compute approach best fits these requirements?

A company runs a stateless web application on an EC2 Auto Scaling group. During sudden traffic surges, newly launched instances take about 8 minutes to complete a lengthy bootstrapping process (installing packages, downloading large data files, and warming an application cache) before they can serve traffic. This delay causes request latency and dropped connections during scale-out events. The application code and dependencies change infrequently. Which solution BEST reduces the time to serve traffic when scaling out, with the least ongoing operational effort?

A media company runs a content-processing application on a fleet of EC2 instances that share files through an Amazon EFS file system. Most of the day the workload is light, but several times a week batch rendering jobs generate large, unpredictable spikes in throughput that far exceed the baseline. During these spikes, users report that file operations become slow because the file system exhausts its burst credits. The team wants a solution that automatically delivers high throughput during unpredictable spikes without provisioning a fixed capacity or managing credits. Which approach best meets these requirements?

A company is migrating a fleet of Windows-based application servers to Amazon EC2. The application requires a shared file system that all instances can mount concurrently using the SMB protocol, with full support for Windows NTFS permissions and Active Directory integration. The team needs a fully managed solution that scales automatically. Which AWS storage service should the solutions architect recommend?

A research company runs a high-performance computing (HPC) workload on a fleet of hundreds of Linux EC2 instances. The workload processes large genomics datasets that currently reside in an Amazon S3 bucket. The compute nodes require a POSIX-compliant shared file system that can deliver hundreds of gigabytes per second of throughput and sub-millisecond latencies while processing the data, and results must be written back to S3 when jobs complete. Which storage solution best meets these performance requirements?

A company operates a real-time multiplayer gaming backend across EC2 instances behind Network Load Balancers in three AWS Regions (us-east-1, eu-west-1, ap-southeast-1). Players worldwide report inconsistent connection latency and slow failover when a regional endpoint becomes unhealthy. The backend uses raw TCP and UDP protocols. The company wants players to connect to the nearest healthy Region automatically through static entry-point IP addresses, with fast rerouting during Regional failures. Which solution best meets these requirements?

A financial services company runs a stateless REST API on an Application Load Balancer backed by Auto Scaling groups in two AWS Regions (us-east-1 and eu-west-1). The API returns highly dynamic, non-cacheable responses. Global users report inconsistent latency and slow initial connections, and the company wants a static set of IP addresses that can be allow-listed by enterprise clients' firewalls, along with fast regional failover. Which solution best meets these requirements?

A data engineering team stores raw CSV files in an Amazon S3 data lake. They need to periodically clean the data, convert it to a columnar format for efficient querying, and catalog the resulting schema so it can be queried by Amazon Athena. The team wants a fully managed, serverless solution that requires no infrastructure to provision or manage. Which approach best meets these requirements?

A company runs a self-managed PostgreSQL database on a single EC2 instance. The database currently uses a 500 GB gp2 EBS volume, but the application team reports that the volume cannot sustain the required 12,000 IOPS during peak business hours without adding more capacity. The team wants to guarantee the required IOPS while minimizing storage cost, and the workload does not require sub-millisecond latency or multi-attach. Which EBS configuration should the solutions architect recommend?

A media rendering company runs a fleet of EC2 instances that perform video transcoding. Each job downloads source files, writes large amounts of intermediate scratch data during processing, and uploads the final output to Amazon S3. The intermediate data is fully disposable and does not need to survive an instance stop or termination. The team reports that the intermediate write/read operations are bottlenecking performance and wants the highest possible local disk throughput at the lowest cost. Which storage option should the solutions architect recommend for the scratch data?

A retail company ingests clickstream events into an Amazon Kinesis Data Stream. The analytics team needs to compute continuous, tumbling-window aggregations (such as page views per product every minute) and push the results to a live operational dashboard with minimal latency. They want a solution that requires no server management and can process the stream in real time using SQL or Apache Flink. Which approach best meets these requirements?

A media company runs a website that generates a high-volume clickstream. They need to ingest events in real time and have multiple independent consumer applications process the same records: one for a live analytics dashboard (sub-second latency), and another that aggregates data and writes hourly batches to a data lake. Ordering of events per user session must be preserved. Which ingestion solution best meets these requirements?

A media company ingests a continuous stream of application log events, currently reaching several thousand records per second. The analytics team needs the raw events delivered to an Amazon S3 data lake in near-real time, converted to Apache Parquet format, and partitioned by event date so that Amazon Athena can query them efficiently. The team wants to minimize the amount of custom code and infrastructure they have to manage. Which solution best meets these requirements?