I'm fairly new to aws but from what I've learned so far load balancer takes a request and forwards it to some node. What I'm looking for is a scenario in which user sends one request (e.g. render 1000 frames of some animation) and load balancer then sends to individual nodes chunks of computation to perform (node A renders first 10 frames, node B renders next 10 frames and so on, and if node B fails then some other node will still render frames 10-20 and nothing gets lost). Is this even a job for load balancer? How do you go about implementing such a system? If there is some tutorial/book/course/etc I'd be thankful for any recommendations.
AWS Elastic Load Balancers (ELB) are not meant for that. Instead you can try something like Render Farm Deployment Kit on AWS.