set nexmark benchmark for temporal join with high cache miss rate

[st1page]

No description provided.

[TennyZhuang]

This may be difficult because NEXMark has good locality since the data is generated, which is a business characteristic. Unless we use some tricks like limiting memory to a very small amount, I don't think this makes much sense. Perhaps we can write some queries or use cases with lower locality.

[st1page]

There are some NEXMark query not having good locality such as q20, I guess we can jsut change them to temporal join

[lmatz]

There are some NEXMark query not having good locality such as q20, I guess we can jsut change them to temporal join

Great

I am also adding a new Kube-bench pipeline that uses RW's datagen to ingest data into Kafka first and then run customized queries reading from Kafka, I will try to finish the basics soon
so everyone can add more stuff into it.