Question #1765
A financial institution is designing a high-frequency trading system that requires ultra-low latency communication between servers. The architecture must ensure minimal network latency between Amazon EC2 instances. Which two configurations should be implemented together? (Choose two.)
Enable enhanced networking with SR-IOV on the EC2 instances.
Distribute the EC2 instances across multiple Availability Zones.
Launch all EC2 instances within a single cluster placement group.
Configure each EC2 instance with multiple elastic network interfaces.
Use EC2 instances that are optimized for Amazon EBS.
Explanation
The correct answers are A and C.
Why A (Enhanced Networking with SR-IOV) is correct:
SR-IOV bypasses the hypervisor for direct network interface access, reducing latency and jitter—critical for high-frequency trading.
Why C (Cluster Placement Group) is correct:
Cluster placement groups place EC2 instances in the same rack, minimizing network hops and ensuring ultra-low latency communication.
Why other options are incorrect:
- B: Distributing across AZs increases latency due to inter-AZ communication.
- D: Multiple ENIs don't reduce latency; they increase bandwidth or segmentation.
- E: EBS optimization targets storage I/O, not network latency.
Key Points:
1. Use SR-IOV (enhanced networking) for direct hardware access.
2. Cluster placement groups ensure physical proximity for low-latency communication.
3. Avoid multi-AZ deployments for latency-sensitive systems.
Answer
The correct answer is: AC