NCP-AII Question 52
Single answerA company is designing a data center to support its AI workloads. The design needs to ensure high bandwidth, low latency, and scalability for future expansions. Which of the following network topologies is most suitable for these requirements?
- A
Fat-tree topology
- B
Bus topology
- C
Ring topology
- D
Star topology
Show answer and explanation
Correct answer: A
Explanation
The fat-tree topology is the most suitable choice for data centers supporting AI workloads because it provides a hierarchical structure that ensures high bandwidth and low latency, meeting the performance demands of AI and HPC applications. It also allows for scalability, accommodating future expansions without significant redesign.
- A. Correct.
Fat-tree topology is commonly used in modern data centers due to its high bandwidth, low latency, and scalability features. It is particularly effective for AI and HPC workloads.
- B. Incorrect.
Bus topology is outdated and lacks the scalability and performance needed for modern data center environments.
- C. Incorrect.
Ring topology introduces latency issues and is not suitable for high-performance computing or AI workloads due to its sequential data transmission.
- D. Incorrect.
Star topology, while simple, is not scalable for large data centers and could become a bottleneck for AI infrastructure.