An enterprise is planning to migrate its on-premises data center to AWS. The migration inv… — AWS Certified Solutions Architect - Professional

An enterprise is planning to migrate its on-premises data center to AWS. The migration involves two distinct datasets:

* A $1.2\text{ PB}$ dataset of historical logs stored on an on-premises Hadoop Distributed File System (HDFS). This data needs to be migrated to Amazon S3 Standard and then transitioned to Amazon S3 Glacier Flexible Retrieval after $30\text{ days}$ .
* A $120\text{ TB}$ dataset of active user documents stored on an on-premises SMB file share. This data must be migrated to an Amazon FSx for Windows File Server file system. The documents are actively modified, and the migration must minimize cutover downtime by replicating changes up to the final cutover window.

The enterprise has a dedicated $1\text{ Gbps}$ Direct Connect connection. A maximum of $300\text{ Mbps}$ can be dedicated to the migration to avoid disrupting production traffic. The migration must be completed within a strict $40\text{ day}$ window. Security policy mandates that all data must be encrypted at rest using customer-managed keys (CMKs) in AWS KMS, and the migration process must maintain files' metadata (such as permissions and timestamps).

Which combination of actions will meet these requirements? (Select TWO.)

For the HDFS dataset, order multiple AWS Snowball Edge Storage Optimized devices. Configure the Snowball jobs to use a customer-managed KMS key (CMK) for encryption. Copy the HDFS data to the devices, ship them to AWS to import the data into an Amazon S3 Standard bucket, and configure an S3 Lifecycle policy on the bucket to transition the objects to Amazon S3 Glacier Flexible Retrieval after $30\text{ days}$ .Answer
For the SMB dataset, order AWS Snowball Edge Storage Optimized devices to seed the data. Copy the SMB files to the devices and import them into an intermediate Amazon S3 bucket. Deploy AWS DataSync agents on-premises, and configure a DataSync task to copy the seeded data from the S3 bucket to the FSx for Windows File Server file system. Configure a second DataSync task to sync incremental changes directly from the on-premises SMB share to the FSx for Windows File Server file system before cutover.Answer
C
Deploy AWS DataSync agents on-premises, and configure a DataSync task to copy the HDFS dataset directly from the on-premises HDFS cluster to Amazon S3 Standard over the Direct Connect connection using a customer-managed KMS key (CMK). Configure a target S3 Lifecycle policy to transition the data to S3 Glacier Flexible Retrieval.
D
For the HDFS dataset, order multiple AWS Snowball Edge Storage Optimized devices. Configure the Snowball jobs to encrypt the data at rest using the default AWS-managed KMS key for Amazon S3 (aws/s3). Copy the HDFS data to the devices, ship them to AWS to import the data into an Amazon S3 Standard bucket, and configure an S3 Lifecycle policy to transition the objects to S3 Glacier Flexible Retrieval.
E
Deploy AWS DataSync agents on-premises to migrate the SMB dataset. Configure a Direct Connect Gateway connected directly to a Virtual Private Gateway (VGW) in a central hub VPC, and route replication traffic transitively through the hub VPC to FSx for Windows File Server file systems residing in multiple spoke VPCs without utilizing AWS Transit Gateway.

Answer

Ordering multiple AWS Snowball Edge Storage Optimized devices utilizing customer-managed KMS keys (CMKs) to migrate the HDFS data to Amazon S3 Standard with an S3 Lifecycle transition policy, combined with ordering AWS Snowball Edge devices to seed the SMB dataset to an intermediate Amazon S3 bucket, followed by using AWS DataSync to copy that seeded data to FSx for Windows File Server and sync subsequent incremental changes.

To migrate

1.2\text{ PB}

of data within a

40\text{ day}

window over a

300\text{ Mbps}

bandwidth limit, an offline migration using AWS Snowball Edge is required since an online transfer would take over a year. The Snowball jobs must use a Customer Managed Key (CMK) because default AWS-managed keys cannot be configured for Snowball Edge imports. An S3 Lifecycle policy will automate the transition to Glacier Flexible Retrieval. For the

120\text{ TB}

SMB dataset, an online transfer would take over

46\text{ days}

, exceeding the

40\text{ day}

limit. Seeding the initial dataset using Snowball Edge to an intermediate S3 bucket, using AWS DataSync to copy that seed to FSx for Windows File Server, and then running incremental DataSync tasks directly from the on-premises SMB share to FSx for Windows File Server enables a successful migration within the timeline.

Step-by-Step Solution

Calculate the transfer capabilities for the network path.

At a maximum allocated bandwidth of

300\text{ Mbps}

with a standard

80\%

network efficiency, the maximum transfer rate is approximately

2.59\text{ TB/day}

This calculation determines if either of the datasets can be transferred entirely online within the

40\text{ day}

window.

Evaluate the migration timeline for the

1.2\text{ PB}

HDFS dataset.

1.2\text{ PB}

dataset requires approximately

463\text{ days}

of continuous online transmission, which exceeds the

40\text{ day}

timeline. Therefore, the HDFS dataset must be migrated offline using multiple AWS Snowball Edge Storage Optimized devices.

This rules out online-only tools like AWS DataSync for the bulk HDFS migration.

Evaluate the migration timeline for the

120\text{ TB}

SMB dataset.

120\text{ TB}

dataset requires approximately

46.3\text{ days}

of online transmission, which also exceeds the

40\text{ day}

window. The initial

120\text{ TB}

must be seeded offline using Snowball Edge, after which a DataSync task can sync incremental updates directly over the network.

Seeding the bulk SMB dataset offline reduces the remaining online transfer volume to only daily deltas, which easily fit within the timeline.

Select the appropriate encryption and routing configurations.

The Snowball Edge jobs must use a Customer Managed Key (CMK) because default AWS-managed KMS keys like `aws/s3` are not supported. Transitive routing of replication traffic via a Direct Connect Gateway directly to a Virtual Private Gateway (VGW) is unsupported, meaning that AWS Transit Gateway is required for multi-VPC spoke routing.

This satisfies the security compliance (CMKs) and networking topology constraints.

Key Concept

Large-Scale Data Transfer using Snow Family, DataSync, and Transfer Family

Estimated Time:3m 0s

Question

Answer

Step-by-Step Solution

Key Concept

Common Mistakes

Topics

More questions from this topic