GooseFSx Types
GooseFSx C50 Type
C50 has two specifications: C50.T3 and C50.T12. Except for capacity, the two specifications are the same in performance, price, features, and robustness.
Features
High performance, performance scales linearly with capacity, 100MB/s - 120MB/s read bandwidth per TiB, 100MB/s - 120MB/s write bandwidth per TiB, 10,000 read IOPS per TiB, 10,000 write IOPS per TiB.
Starting capacity, step length and usage recommendations as shown in the table below:
|
C50.T12 | 36TiB | 12TiB | Data volume exceeds 200TiB or grows fast. |
C50.T3 | 9TiB | 3TiB | Data volume within 200TiB |
If you need to further enhance performance, or if purchase failed due to resource insufficient in AZ, or have another requirement, you can contact Storage Architect or consult Online Customer Service to ask for help. Applicable Scenarios
AI training and inference scenarios: The data accelerator GooseFSx C50 accelerates AI training and reasoning performance. It enables seamless data flow with the data lake COS, providing immediate access to COS model data for AI computing power. This avoids expensive computing resources waiting for data and allows training results to be stored in COS as needed, reducing overall storage costs.
Gene analysis scenario: The data accelerator GooseFSx C50 accelerates high-performance storage requirements for gene sequencing and local alignment stages. It enables free data flow with the data lake COS, implements immediate access to COS samples on the Omics platform, achieves automatic archiving of gene analysis results to COS, and delivers them to end users via COS Internet distribution capability.
Autonomous driving scenario: The data accelerator GooseFSx C50, combined with data lakes COS and data migration, forms an end-to-end autonomous driving solution. It provides an integrated service for the entire process and full cycle, including data collection (road and vehicle data migration to the cloud and uploading to data lakes), computation (immediate training), and storage (long-term persistent storage).
High-performance computing scenario: The data accelerator GooseFSx C50 offers high-performance, low-latency, high-throughput parallel file service, fully meeting the high-throughput, low-latency needs of HPC computation. Integrated with the data lakes COS, it provides ultra-high performance, hyperscale, low-cost storage service.
GooseFSx C60 Type
C60 has two specifications: C60.T2 and C60.T12. Except for capacity, the two specifications are the same in performance, price, features, and robustness.
Features
High performance, performance scales linearly with capacity, 200MB/s read bandwidth per TiB, 200MB/s write bandwidth per TiB, 20,000 read IOPS per TiB, 20,000 write IOPS per TiB.
Starting capacity, step length and usage recommendations as shown in the table below:
|
C60.T12 | 36TiB | 12TiB | Data volume exceeds 100TiB or grows fast. |
C60.T2 | 4.5TiB | 1.5TiB | Data volume within 100TiB |
If you need to further enhance performance, or if purchase failed due to resource insufficient in AZ, or have another requirement, you can contact Storage Architect or consult Online Customer Service to ask for help. Applicable Scenarios
AI training and inference scenarios: Data accelerator GooseFSx C60 accelerates AI training and inference performance. It enables data to flow freely with data lakes like COS, providing immediate access to model data in COS for AI computing power. This avoids expensive compute resources waiting for data and allows training results to be stored in COS as needed, reducing overall storage costs.
Genetic Genetic analysis scenario: The data accelerator GooseFSx C60 speeds up high-performance storage needs for gene sequencing and local alignment stages. It enables seamless data flow with the data lake COS, allowing the Omics platform to immediately access COS samples, automatically archive genetic analysis results to COS, and deliver them to end users via COS's Internet distribution capability.
Autonomous driving scenario: The data accelerator GooseFSx C60, combined with data lakes COS, data migration, and other products, forms an end-to-end autonomous driving solution. It offers an integrated service covering the entire process and full cycle, including data collection (road and vehicle data uploading to the cloud and migration to data lakes), computation (immediate training), and storage (long-term persistent storage).
High-performance computing scenario: The data accelerator GooseFSx C60 offers high-performance, low-latency, high-throughput parallel file service, fully meeting the high-throughput, low-latency needs of HPC computation. Integrated with the data lakes COS, it provides ultra-high-performance, hyperscale, low-cost storage service.
CAE/CAD scenario: GooseFSx C60 accelerates small object read/write for CAE/CAD applications. Integrated with COS data lakes, it provides ultra-high-performance, hyperscale, low-cost storage service.
Film and television rendering scene: Data accelerator GooseFSx C60 integrates with data lakes COS to provide integrated storage service. Rendering material is archived in COS at low cost. Data accelerator GooseFSx C60 is pulled up as needed for rendering, and final outputs are stored in COS for long-term maintenance. They are delivered to end users via COS Internet distribution capability.
GooseFSx C70 Type
Features
High performance, performance scales linearly with capacity, 600MB/s read bandwidth per TiB, 200MB/s write bandwidth per TiB, 30,000 read IOPS per TiB, 20,000 write IOPS per TiB.
The initial capacity is 13.9453125TiB (~14TiB), with a step length of 4.6484375TiB (~4.5TiB). Capacity specifications include 13.9453125TiB (~14TiB), 18.59375TiB (~18.5TiB), 23.2421875TiB (~23TiB), 27.890625TiB (~28TiB), 32.53906250TiB (~32.5TiB).
If you have another requirement, or if purchase failed due to resource insufficient in AZ, you can consult Online Customer Service to ask for help. Applicable Scenarios
AI training and inference scenarios: The data accelerator GooseFSx C70 accelerates AI training and inference performance. It enables seamless data flow with the data lakes COS, providing immediate access to model data for AI computing power. This avoids expensive computing resources waiting for data and allows training results to be stored in COS as needed, reducing overall storage costs.
Autonomous driving scenario: The data accelerator GooseFSx C70, combined with data lakes COS, data migration, and other products, forms an end-to-end autonomous driving solution. It offers an integrated service covering the entire process and full cycle, including collection (road and vehicle data uploading to the cloud and entering the lake), computation (immediate training), and storage (long-term persistent storage).
High-performance computing scenario: The data accelerator GooseFSx C70 offers high-performance, low-latency, high-throughput parallel file service, fully meeting the high-throughput, low-latency needs of HPC computation. Integrated with data lakes COS, it provides ultra-high performance, hyperscale, low-cost storage service.
GooseFSx Type Comparison
Note:
Involved in resource reservation. If you have further needs, you can consult online customer service to ask for help. Due to network reasons, performance may fluctuate by around 10%.
Main performance specifications comparison as follows:
|
Initial capacity (TiB) | 13.95 (approx. 14) | C60.T2 ： 4.5 C60.T12 ： 36 | C50.T3 ： 9 C50.T12 ：36 |
Step length capacity (TiB) | 4.65 (approx. 4.5) | C60.T2 ：1.5 C60.T12 ： 12 | C50.T3 ： 3 C50.T12 ：12 |
Read bandwidth MB/s per TiB 1 | 600 | 200 | 100 - 120 |
Write bandwidth MB/s per TiB 1 | 200 | 200 | 100 - 120 |
Read IOPS per TiB | 30,000 | 20,000 | 10,000 |
Write IOPS per TiB | 20,000 | 20,000 | 10,000 |
Latency ms | 0.1~1 | 0.1~1 | 0.1~1 |
Note 1: If you need to further enhance performance, you can contact Storage Architect or consult Online Customer Service to ask for help.