By: Edward Folarin, Delivery Manager
Data transfer/migration is a key consideration in any organization’s decision to move into the cloud. If a sound strategy is applied, migration of on-premise data to the cloud is usually a seamless process. When an organization fails to do so, however, it risks running into challenges stemming from deficiencies in technical resources, inadequate planning, and/or incompatibility with legacy systems, to name a few.
Data transfer via AWS Snowball is no exception. If performed incorrectly or out of order, some of the seemingly insignificant tasks related to the data migration process can become substantial obstacles that adversely affects a timeline. The AWS Snowball device can be simple to use if one is familiar with other AWS data transfer services and/or follows all of the steps provided in the AWS Snowball User Guide.However, neglecting a single step can greatly encumber an otherwise ordinary data transfer process.
According to AWS on its service:
“AWS Snowball is used to transport terabytes or petabytes of data to and from AWS, or who want to access the storage and compute power of the AWS Cloud locally and cost effectively in places where connecting to the internet might not be an option.”
When preparing to migrate data from on-premises storage into AWS via a Snowball device, an organization should be aware of the importance of 11 easily overlooked tasks and considerations associated with planning for the data move. They are as follows:
1. Understanding the specifics of the data being moved to the cloud. Ensure that it is compatible and can transfer seamlessly to the cloud via AWS Snowball. Follow a cloud migration model to help layout specific details and avoid surprises during the data transfer process.
2. Verifying and validating the amount of data being transferred. Snowball is intended for large data transfers (over 10 terabytes). Using it for smaller data transfers is not a cost-effective option.
3. Verifying that the workstation meets the minimum requirement for the data transfer. It should have a 16-core processor, 16 MB of RAM, and a RJ45 or SPF+ network connection.
4. Performing a data transfer test on the workstation an organization plans to use to complete the task. This will not only equip the organization with an understanding of the amount of time needed to perform the transfer, but will provide an opportunity to try various methods of transferring data. Additionally, it will assist with estimating the time the Snowball device will need to be in the organization’s possession, as well as its associated cost.
NOTE: The Snowball Client must be downloaded and installed before this step is performed.
5. Creating a specific administrative IAM user account for the data transfer process via the management console. This account will be used to order, track, create and manage Snowball Import/Export jobs and return the device to AWS.
NOTE: It is important to avoid using personal IAM user accounts if individuals will be responsible for ordering the device and performing the data transfer.
6. Following the “Object Key Naming convention” when creating S3 buckets. It is also important to confirm that the selected S3 bucket name aligns with the expectations of the stakeholders.
7. Confirming the point of contact/s and shipping address for the Snowball device. This is especially important if the individual ordering the device is different from the one performing the data transfer.
8. Setting up SNS notifications to help track the stages of the snowball job. This will keep the stakeholders informed of the shipping status and the importing of data to the S3 bucket.
9. Being aware of how holidays could affect the progress or process of the data-transfer timeline. This is important because additional costs are accrued 10 days after the Snowball is delivered.
10. Considering the organization’s administrative processes that might hinder or delay the data transfer process. By factoring in internal processes (e.g., Change Request management, stakeholder buy-in, technical change moratoriums, etc.) into the timeframe it will take to receive the device, start the job, and ship it back to AWS can help prevent unnecessary fees.
NOTE: The Snowball device has no additional cost if it is returned within 10 days from the date it is received. Following that time, however, a daily late fee of $15 is applied until the date AWS receives it.
11. Keeping the original source data intact till the data import is confirmed. It is very important that source data remain intact until the Snowball device has been returned to AWS, the data import has been completed, and the customer has validated the data in the S3 bucket(s).
Transferring data from on-premises to an AWS Snowball can be an uneventful endeavor when thorough planning is done in advance of ordering the device. Taking these 11 planning tasks and considerations into account are essential to eliminating some of the potential headaches and stress occasionally associated with this type of activity.
Refer to AWS Snowball Documentation for additional information and specific instructions not covered in this article.
If you or your organization has more questions, reach out to us at [email protected]