A Video Streaming Company Uses SLB to Distribute User Requests

Multiple Choice

A video streaming company uses SLB to distribute user requests to 30 ECS instances (the 30 ECS instances have the same configuration) . Yet, the company finds that the service traffic soars dramatically every night from 20:00 to 02:00. According to their calculation, the evening traffic is 100% higher than the traffic in other periods of time. To properly respond to user requests, which of the following methods is most preferred from the perspective of cost and implementation simplicity?

A) Configure Auto Scaling appropriately to automatically add or reduce ECS instances.
B) Manually create 30 ECS instances using the custom image, and then add them to SLB every night before 20:00 and remove them after 02:00.
C) Create 30 ECS instances based on the custom image and add them to SLB. Then, configure weights for these ECS instances to respond to user requests every night from 20:00 to 02:00, and reset those weights to zero for other periods of time.
D) Configure Auto Scaling appropriately to automatically upgrade/downgrade the ECS instances.