FBAP_002 practice question 10 of 15
An HPC environment has the following: - 100 physical hosts, each with a single 10GbE NIC, and a single NFS mount point to the FlashBlade. - A 7x52TB FlashBlade, connected to the Top-of-Rack switches via eight 40GbE connections. What is the maximum aggregate write throughput the customer should expect from the FlashBlade?
Full set: 75 questionsDomain Implement, Migrate, and Integrate FlashArray and FlashBlade Solutions
The maximum aggregate throughput of a system is determined by its primary bottleneck. In this scenario, we must evaluate the client-side bandwidth, the network bandwidth, and the storage system's write performance.
1. Client Bandwidth: 100 hosts × 10 Gbps/host = 1000 Gbps (125 GB/s)
2. Network Bandwidth: 8 ports × 40 Gbps/port = 320 Gbps (40 GB/s)
3. Storage Write Throughput: FlashBlade performance scales linearly with the number of blades. Official specifications for a 10-blade FlashBlade//S system state a maximum write throughput of approximately 3.1 GB/s. For a 7-blade system, this scales to (3.1 GB/s / 10 blades) × 7 blades ≈ 2.17 GB/s. Comparing these values, the FlashBlade chassis's write performance (≈2.1 GB/s) is the clear bottleneck.
A. Approximately 7 GB/s: This value is more aligned with the expected read throughput for a 7-blade system, which is significantly higher than its write performance.
B. Approximately 4.2 GB/s: This level of write performance would necessitate a much larger FlashBlade configuration, such as a full 15-blade chassis, not the specified 7-blade system.
C. Approximately 4 GB/s: This value does not correspond to any of the potential system bottlenecks (client, network, or storage) in this specific configuration.
1. Pure Storage, "FlashBlade//S Data Sheet" (PURE-DS-FB-S-L-EN-R10), Page 2, "Performance and Capacity" table. This document specifies performance for a 10-blade chassis, showing up to 3.1 GB/s for writes. This figure serves as the baseline for the linear scaling calculation used to determine the 7-blade system's performance.
2. Pure Storage, "FlashBlade Architecture White Paper" (pswpflashbladearchitecture.pdf), Section: "Linearly-Scaling Performance", Page 8. This document explains the fundamental architectural principle that "performance scales linearly as blades are added to the system." This supports the methodology of calculating the 7-blade performance by scaling down from the documented 10-blade specification.
3. Pure Storage, "FlashBlade Best Practices for NFS" (pure-storage-flashblade-nfs-best-practices.pdf), Section: "Performance Considerations", Page 12. This guide discusses identifying performance bottlenecks in an NFS environment with FlashBlade, reinforcing that the storage system's capabilities, client-side limitations, or the network can be the limiting factor, requiring a holistic analysis as performed here.
Keep practicing
Take the free FBAP_002 practice test
Ten exam-style questions with answers and explanations, plus the exam facts and study guides.
More questions
Other FBAP_002 practice questions
- Question 1The customer wants to build a software QA infrastructure using modern big data tools like…
- Question 2A customer has noticed their all-flash arrays are constantly running at or near 100% load…
- Question 4A customer has hundreds of TB of data that they need to archive for 20 years. The custome…
- Question 5A FlashBlade is designed with a properly implemented single 4x40Gb/s LAG. During implemen…
- Question 6A customer wants to use and orchestration tool to manage the FlashBlade. How should the c…
- Question 7A customer wants to leverage FlashBlade to create a DataHub. Which set of applications wi…
