Examkin

Soru

Zorluk: ZorHigh-Performing and Scalable Storage Solutions

A financial firm is deploying a distributed quantitative modeling simulation across 5050 Amazon EC2 instances. The simulation requires a shared file system that can scale throughput to 20 GiB/s20\text{ GiB/s} and handle up to 1,000,0001,000,000 IOPS with sub-millisecond latencies. The simulation periodically reads raw historical data stored in an Amazon S3 bucket, processes it, and writes the results back to S3. The file system must remain highly available to ensure uninterrupted running of long-term modeling jobs. Which two actions should a solutions architect recommend to meet these requirements? (Select TWO.)

  1. Deploy an Amazon FSx for Lustre Persistent SSD file system and link it to the Amazon S3 bucket using a Data Repository Association (DRA).Cevap
  2. Mount the file system using the Lustre client on the EC2 instances, and use data repository tasks to synchronize modified data back to Amazon S3.Cevap
  3. C
    Deploy an Amazon EFS file system configured with Max I/O performance mode and Provisioned Throughput set to 20 GiB/s20\text{ GiB/s}.
  4. D
    Deploy an Amazon FSx for Lustre Scratch SSD file system and configure AWS DataSync to schedule hourly replication of output data to the Amazon S3 bucket.
  5. E
    Provision a single Amazon EBS io2 Block Express volume with Multi-Attach enabled, configured with 1,000,0001,000,000 IOPS, and attach it to all 5050 EC2 instances.

Cevap

Deploy an Amazon FSx for Lustre Persistent SSD file system linked to the Amazon S3 bucket using a Data Repository Association (DRA), mount it using the Lustre client on the EC2 instances, and use data repository tasks to synchronize modified data back to Amazon S3.
The correct options recommend deploying an Amazon FSx for Lustre Persistent SSD file system linked to Amazon S3 using a Data Repository Association (DRA), mounting it using the Lustre client on the EC2 instances, and using data repository tasks to synchronize modified data back to Amazon S3. FSx for Lustre Persistent SSD deployment provides the required sub-millisecond latencies, up to hundreds of GB/s of throughput, and millions of IOPS, while offering high availability (data is replicated within the Availability Zone, and failed components are automatically replaced). The Data Repository Association natively integrates the file system with Amazon S3, allowing raw data to be loaded automatically and results to be synchronized back via data repository tasks.

Adım Adım Çözüm

1
Analyze workload scaling and shared access requirements
Identified that the file system needs to support 5050 concurrent EC2 instances, deliver 20 GiB/s20\text{ GiB/s} throughput, 1,000,0001,000,000 IOPS, sub-millisecond latencies, high availability, and direct S3 integration.
This establishes that a high-performance shared file system (HPC optimized) with native S3 integration is required.
2
Evaluate and select the storage service and deployment type
Selected Amazon FSx for Lustre Persistent SSD because Persistent deployment offers high availability, SSD storage meets the latency and throughput goals, and it natively links with Amazon S3. EFS and EBS Multi-Attach are ruled out due to throughput limits and maximum instance attachment limits.
Ensures the performance, scaling, and high-availability criteria are met.
3
Determine the data synchronization mechanism
Configured a Data Repository Association (DRA) and Lustre client on the EC2 instances, using data repository tasks to sync data back to Amazon S3.
Provides seamless, high-performance data import/export between Lustre and S3.

Anahtar Kavram

Selecting high-performance shared file systems with native S3 integration for HPC workloads
Tahmini Süre:3m 0s
Bu soruyu puanla

Konular

High-Performing and Scalable Storage Solutions

Bu konudan daha fazla soru