Apache Spark or Akka Cluster for stream processing?

Question

Scala

What is better to use for streaming data processing?
There are tasks that take small strings as input, then the tasks send many requests to various APIs (vk, twitter, yandex) and finally save the result to MongoDB. Time to complete one task 0.5 - 3 sec. Disk operations are not performed.
Which option to choose for the best performance?

Answer 1

Parallel tasks. And a special database, like this https://clickhouse.yandex/