jvm heap failure with large number of instances

[maxiaohao]

Large number of instances can failure the jvm - see class AbstractMockEc2Instance, which makes use of java.util.Timer and can be a probable cause. That means every mock EC2 instance runs in a separate jvm thread based on java.util.Timer, which consumes considerable amount of resources and large number of threads could easily fail the jvm. For now we set a large jvm heap size in gradle.properties of 4GB heap in memory to afford 30k mock EC2 instances successfully, which is fairly a large number though. However, we need to explore further options that can substitute the current java.util.Timer based model and have the performance improved.

[nottoseethesun]

This is an enhancement, as aws-mock already works fine for a large number of instances. This is only to increase that already large number.