I have some algorithms that receive market data as events and process them to send order requests with an API. The problem is scaling these algorithms to support many users and many algorithm instances. If i use K8s, I have to assign a pod to each instance that is not optimized. If i group some of users into one pod, then how can i found that what is the optimized number of instances? (process depends on market changes) Some people proposed using common data processing platforms. What is the best practice?
Asked
Active
Viewed 56 times
1 Answers
0
After some research i found AKKA and actor model that can fit my requirements. https://akka.io/
sohrab haghayegh
- 31
- 6