Migrating 10TB Database to Cloud Storage via Network

PCA By islever Editorial Last updated

Designing network, storage, and compute resources.

Short answer

D Upload the data with gcloud storage cp.

Question

Your company wants to migrate their 10-TB on-premises database export into Cloud Storage. You want to minimize the time it takes to complete this activity and the overall cost. The bandwidth between the on-premises environment and Google Cloud is 1 Gbps. You want to follow Google-recommended practices. What should you do?
  1. Develop a Dataflow job to read data directly from the database and write it into Cloud Storage.
  2. Use the Data Transfer appliance to perform an offline migration.
  3. Use a commercial partner ETL solution to extract the data from the on-premises database and upload it into Cloud Storage.
  4. Upload the data with gcloud storage cp. Correct Answer

Key Takeaways

The core trap is assuming large data always requires an appliance; however, with sufficient bandwidth (1 Gbps), direct upload is faster and more cost-effective than shipping hardware.

This question addresses the optimal method for migrating a 10-TB dataset over a 1 Gbps connection, highlighting the trade-offs between online CLI tools and offline appliances.

Common mistake: Candidates often select the Data Transfer Appliance (B) due to its association with 'large-scale' migrations, failing to calculate that 10 TB fits within a week on a 1 Gbps link.

Expert Analysis

Why the Answer Is Correct

Correct answer: D — Upload the data with gcloud storage cp. The most efficient way to transfer 10 TB of data over a 1 Gbps network is using gcloud storage cp (or gsutil). A 1 Gbps link can transfer approximately 10.5 TB in 24 hours, meaning the migration would take roughly 30 hours. This is significantly faster than waiting for a physical appliance to be shipped, provisioned, and returned, which typically takes at least one week. Additionally, using the network avoids the logistical costs and delays associated with the appliance service.

Why the Other Options Are Wrong

Option A (Dataflow) introduces unnecessary complexity and cost for a simple bulk export of database files. Option C (Commercial ETL) is also unnecessarily complex and costly for a raw data migration task. Option B (Data Transfer Appliance) is designed for scenarios where online transfer would exceed one week or when bandwidth is extremely limited (e.g., <100 Mbps). Since the network bandwidth here allows for a rapid transfer, the appliance is not the recommended choice for minimizing time and cost.

Community Comment Notes

Several community members correctly identified that the appliance threshold is based on transfer duration exceeding one week. As user cchiaramelli noted, "Since 10TB would take way less than a week for that bandwidth, I would go for D." User JohnDohertyDoe reinforced this by calculating the 30-hour transfer time. While some users like mgm7 cited a "5TB object size" limit, this refers to individual object limits (not total dataset size) and is irrelevant to the transfer mechanism choice. The consensus among technical experts points to D as the correct operational strategy.

Official Reference

https://cloud.google.com/architecture/migration-to-google-cloud-transferring-your-large-datasets

This link opens the official vendor documentation in a new tab.

Exam Strategy

Always evaluate the specific constraints given in the scenario. If the bandwidth is high enough to transfer the data volume in less than a week, prefer online methods (CLI/GSUtil) over offline appliances to save time and money.

Frequently Asked Questions

Why is the Data Transfer Appliance not used for 10TB?

Appliances are recommended only if online transfer exceeds one week. On a 1 Gbps link, 10TB takes ~30 hours, making the appliance slower and more expensive.

Does the 5TB object limit affect this migration?

No. The 5TB limit applies to single objects. You can split the database dump into multiple smaller files and upload them all concurrently using gsutil/gcloud.

More PCA FAQ →

Discussion 47 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 30 of 47

  1. cchiaramelli 👍 15 Answered D
    This Transfer Appliance docs says it is suitable when "It would take more than one week to upload your data over the network" Since 10TB would take way less than a week for that bandwidth, I would go for D https://cloud.google.com/transfer-appliance/docs/4.0/overview#suitability
  2. mgm7 👍 7 Answered B
    maximum object size in GCS is 5TB
  3. Nora9 👍 6 Answered B
    option B (Use the Data Transfer appliance to perform an offline migration) seems to be the most appropriate. It addresses the need for a speedy transfer of a large amount of data and is a cost-effective solution recommended by Google for large-scale data migrations. This option circumvents potential network bandwidth limitations and provides a reliable way to transfer large datasets. Why option D is not a good choice : D. Upload the data with gcloud storage cp: This method uses the gcloud command-line tool to copy files to Cloud Storage. While simple, it might not be the most efficient for a 10-TB migration given the 1 Gbps bandwidth. The process could be slow and may require additional handling for potential interruptions and resuming uploads.
  4. JohnDohertyDoe 👍 6 Answered D
    Since we would want to do it in the shortest time possible, using gsutil cp would take only 30 hours to move 10TB. So answer is D. https://cloud.google.com/architecture/migration-to-google-cloud-transferring-your-large-datasets?hl=en#online_versus_offline_transfer
  5. madcloud32 👍 5 Answered B
    Answer is B. Max cp limit of file is 5 TB
Show 25 more comments
  1. MikeH20 👍 4
    Respectfully, I feel option D is the better choice. According to https://cloud.google.com/transfer-appliance/docs/4.0/overview#suitability, Google recommends using a transfer appliance if the data transfer would take longer than 1 week. At 1 Gbps and 10TB of data, the transfer would take 30 hours (https://cloud.google.com/architecture/migration-to-google-cloud-transferring-your-large-datasets#online_versus_offline_transfer). We also need to take into account shipping time for the appliance itself, to and from. That would take a couple weeks. The question mentions to minimize time and cost, and follow Google best practices. In this case, D checks those boxes. The question does not mention a customer concern about network interruptions. If they did, then B could be argued as the more appropriate answer.
  2. Zonci 👍 4 Answered B
    B. Use the Data Transfer Appliance to perform an offline migration. using the Data Transfer Appliance aligns with Google's recommended practice for large-scale migrations where bandwidth limitations are a concern, ensuring efficient, secure, and cost-effective transfer of your on-premises database export into Google Cloud Storage.
  3. sheucm89 👍 4
    Vote D. The bandwidth is 1 Gbps. It will only take less than 3 hours to complete the upload of 10TB data by using copy command. The Data Transfer appliance can take several days to weeks. The question ask to minimize the time and the overall cost, so D is suitable.
  4. Prudvi3266 👍 4 Answered D
    As we have 1Ggigabit network we can transfer this with CLI command quicker than Transfer appliance. which takes time like more than week. Incase of lesser band width may be we use transfer appliance
  5. shop2 👍 4
    1 Gbps and not 1GB BW, this means it will take 23 hours approx, still faster than the appliance.
  6. Roro_Brother 👍 4 Answered B
    Because there is 10 TB of data, it's B
  7. david_tay 👍 3 Answered B
    B is correct answer. I just did a through questioning in Gemini and this is a good explanation (after referring to official GCP resources too): Time: Transferring 10TB over a 1 Gbps connection will take a very long time (likely more than a day, even under ideal conditions). The gcloud storage cp command, while capable of resumable uploads for individual files, doesn't handle interruptions well for large directory transfers. Estimated 22 hours to complete transfer, provided in the most ideal situation for the network. But still a big risk if a large file is interrupted halfway and needs to start over. Although likely this 10TB transfer will be via many files, there still could be some large files in between. Hence to optimize cost and time B using transfer appliance is the answer.
  8. Murtuza 👍 3
    The Data Transfer Appliance is a physical device that can be sent to the customer to transfer large amounts of data quickly and securely. The appliance is designed to transfer large amounts of data, and it comes with built-in security features to protect the data during the transfer process. It is also easy to set up and use. The appliance can be connected to the on-premises environment, and the data can be transferred directly to Google Cloud Storage, which minimizes the database load and reduces the time it takes to complete the migration. Therefore, the best option to migrate a large on-premises database export into Cloud Storage, while minimizing cost, time, and database load, is to use the Data Transfer Appliance.
  9. a53fd2c 👍 3
    Option B. Cloud Storage object limit is 5 TB. https://cloud.google.com/storage/quotas?hl=en#objects https://cloud.google.com/storage/quotas?hl=en#objects
  10. Murtuza 👍 3
    The google math is 1TB data, 1GB bandwidth will take 3 hrs. We have 10TB of data so we are looking at 30 HRS hence the storage appliance
  11. Murtuza 👍 3
    Duplicate so review question #146 and the choice over there is offline transfer appliance.This is a tricky question
  12. mstaicu 👍 3
    So how would it work with B ? The data needs to still end up in GCS. Also, who says the export is 1 large file ?
  13. klefevre08 👍 3
    so no answer is possible according to you ?...
  14. ryaryarya 👍 2 Answered D
    No where in the question does it say the database export is one file (for anyone who is hung up on the 5TB object size limit). And even if it was, it still needs to end up in GCS, using the appliance doesn't solve that. Also it would be way more expensive and take a lot longer before your data would be available in GCS using the appliance.
  15. exam4c3 👍 2
    your link says: Transfer Appliance is a good fit for your data transfer needs if It would take more than one week to upload your data over the network, but this workload takes 1 day to complete by CLi
  16. MahAli 👍 2 Answered D
    Appliance is overkilling for 30 hours cli command, the appliance could take more than a week with shipping, how many times you run jobs which take hours to completed? Many times I would go with D
  17. VishalMoon 👍 2 Answered B
    I think the key factor here is "Google Recommended Practices". Based on this alone we have to select B. If this was not there and given the 1 GBPS speed, D would have been feasible.
  18. desertlotus1211 👍 2
    What about cost? And time to get to GCP Region/Zone? And chain of custody of the data as well as upload and available? What is the of cost of that? How ling will it take?
  19. AlainP1980 👍 2 Answered D
    Google practices: gsutil or gcloud storage commands are ok up to few TB bandwidth is 1 Gbps minimize the time and the overall cost => answer D
  20. AlainP1980 👍 2 Answered D
    Google practices: gsutil or gcloud storage commands are ok up to few GB bandwidth is 1 Gbps minimize the time and the overall cost => answer D
  21. AlainP1980 👍 2 Answered D
    Google practices: gsutil or storage are ok up to few GB bandwidth is 1 Gbps minimize the time and the overall cost => answer D
  22. bigdapper 👍 2
    Yep, I've been discussing this with Gemini also and I agree with the logic. I trust Gemini's assessment.
  23. Murtuza 👍 2
    Answer is B google practices is to use a storage transfer appliance for any data that is over 1TB
  24. rrope 👍 2 Answered B
    The best option is B. Use the Data Transfer appliance to perform an offline migration.
  25. smithloo 👍 2
    thanks for sharing it <a href="https://www.qualitybacklink.net">Link building SEO</a>

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

Log in to comment, report an error, or add a note about this question.

Multi-Regional vs Regional Cloud Storage for Global Software Downloads

S-grade

Surviving a Zone Outage With SQL Server on GCP

S-grade

Fixing BigQuery connection errors from a Compute Engine VM

S-grade

Moving 75 TB into Cloud Storage with Google-recommended practices

S-grade

Practice All PCA Questions

Access 81 questions with complete answers and detailed explanations.

View Full PCA Practice Test →

← Back to PCA Study Guide