albertaillet/kepler-model-server

Model Server for Kepler

Kepler Power Model

This repository contains source code related to Kepler power model. The modules in this reposioty connects to core Kepler project and kepler-model-db as below.

Usage

Getting Powers from Estimator

module: estimator (src/estimate/estimator.py)

/tmp/estimator.socket

Parameters of PowerRequest

key	value	description
metrics	list of string	list of available input features (measured metrics)
output_type	either of the following values: AbsPower (for node-level power model), DynPower (for container-level power model)	the requested model type
trainer_name (optional)	string	filter model with trainer name.
filter (optional)	string	expression in the form attribute1:threshold1; attribute2:threshold2.

Getting Power Models from Model Server

module: server (src/server/model_server.py)

:8100/model
POST

Parameters of ModelRequest

key	value	description
metrics	list of string	list of available input features (measured metrics)
output_type	either of the following values: AbsPower (for node-level power model), DynPower (for container-level power model)	the requested model type
weight	boolean	return model weights in json format if true. Otherwise, return model in zip file format.
trainer_name (optional)	string	filter model with trainer name.
node_type (optional)	string	filter model with node type.
filter (optional)	string	expression in the form attribute1:threshold1; attribute2:threshold2.

Posting Model Weights [WIP]

module: server (src/server/model_server.py)

/metrics
GET

Online Trainer [WIP]

module: online trainer (src/train/online_trainer.py) running as a sidecar to server

periodically query prometheus metric server on SAMPLING INTERVAL

Profiler [WIP]

module: profiler (src/profile/profiler.py)

Offline Trainer

module: offline trainer (src/train/offline_trainer.py)

:8102/train
POST

Parameters of TrainRequest

key	value	description
name	string	pipeline/model name
energy_source	valid key in PowerSourceMap	target enery source to train for
trainer	TrainAttribute	attributes for training
prome_response	json	prom response with workload for power model training

• TrainAttribute

  key	value	description
  abs_trainers	list of available trainer class names	trainer classes in the pipeline to train for absolute power
  dyn_trainers	list of available trainer class names	trainer classes in the pipeline to train for dynamic power
  isolator	valid isolator class name	isolator class of the pipeline to isolate the target data to train for dynamic power
  isolator_args	dict	mapping between isolator-specific argument name and value

Test

Build image for testing, run

make build-test

Test case	Command
Training pipeline	make test-pipeline
Model server	make test-model-server
Estimator	make test-estimator
Offline Trainer	make test-offline-trainer

For more test information, check here.

Contributing

Please check the roadmap and guidelines to join us here.