Add complex benchmarks to vHive

[ustiugov]

We plan to add more reprepresentative benchmarks to vHive framework, particularly, multi-function apps, apps with various serverful components (e.g., Apache Kafka), data-intensive apps, apps with complex control and data flow.

The confirmed list of functions with committed contributors:

Function name	Status	Contributor	Type	Knative infra	Link
serving ubench	done	core team	example	serving	link
eventing ubench	done	core team	example	eventing	link
Video Analytics	done	core team	data-intensive	serving	link
MapReduce (Corral)	done	core team	data-intensive	serving	link
gg laptop-to-Lambda	wip	Stanford	complex dataflow, parallel	serving	#370
Anomaly detection	scheduled	individual, AWS Hero	parallel	serving or eventing
Stacking ensemble training	wip	individual, AWS Hero	parallel	serving	#365
Iterative RandomForest training	wip	core team	parallel	serving
ML ETL pipelines	scheduled	ETH		serving
Latency-critical	wip	core team	auth, encrypt, etc.	serving

Next steps: genome sequencing, other ETL pipelines.

[ustiugov]

We welcome the contributions from the open community here! Please get in touch if you would like to add your benchmarks or request particular benchmarks to be added.

[ustiugov]

the benchmarks (beta mode) are public in the vSwarm repo.