[poolifier.git] / docs / worker-choice-strategies.md

# Worker choice strategies

## Table of contents

- [Strategies](#strategies)
  - [Fair share](#fair-share)
  - [Weighted round robin](#weighted-round-robin)
  - [Interleaved weighted round robin](#interleaved-weighted-round-robin)
- [Statistics](#statistics)
  - [Median](#median)

## Strategies

All duration or timestamp are expressed in milliseconds.

### Fair share

Its goal is to distribute the load evenly across all workers. To achieve this, the strategy keeps track of the average task execution time for each worker and assigns the next task to the worker with the lowest task end prediction time: `task_end_prediction = max(current_time, task_end_prediction) + average_task_execution_time`.  
By default, the strategy uses the average task execution time for each worker but it can be configured to use the average task event loop utilization (ELU) active time instead.

### Weighted round robin

The worker weights are maximum tasks execution time, once the worker has reached its maximum tasks execution time, the next task is assigned to the next worker. The worker default weights are the same for all workers and is computed given the CPU cores speed and theirs numbers.

### Interleaved weighted round robin

The worker weights are maximum tasks execution time. The rounds are the deduplicated worker weights.  
During a round, if worker weight is inferior to the current round, the next task is assigned to the next worker. Once all workers have been assigned a task, the next round starts.  
The worker default weights are the same for all workers and is computed given the CPU cores speed and theirs numbers. So the default rounds consists of a unique worker weight.

## Statistics

Worker choice strategies enable only the statistics that are needed to choose the next worker to avoid unnecessary overhead.

### Median

Strategies using the average task execution time for each worker can use the median instead. Median is more robust to outliers and can be used to avoid assigning tasks to workers that are currently overloaded. Median usage introduces a small overhead: measurement history must be kept for each worker and the median must be recomputed each time a task has finished.
Commit	Line	Data
9e2ab226 JB	1	# Worker choice strategies
9e2ab226 JB	2
4f8e444d JB	3	## Table of contents
	4
	5	- [Strategies](#strategies)
	6	- [Fair share](#fair-share)
	7	- [Weighted round robin](#weighted-round-robin)
	8	- [Interleaved weighted round robin](#interleaved-weighted-round-robin)
	9	- [Statistics](#statistics)
	10	- [Median](#median)
c63b7b38	11
9e2ab226 JB	12	## Strategies
9e2ab226 JB	13
4f8e444d JB	14	All duration or timestamp are expressed in milliseconds.
4f8e444d JB	15
9e2ab226 JB	16	### Fair share
	17
	18	Its goal is to distribute the load evenly across all workers. To achieve this, the strategy keeps track of the average task execution time for each worker and assigns the next task to the worker with the lowest task end prediction time: `task_end_prediction = max(current_time, task_end_prediction) + average_task_execution_time`.
2567b3a3	19	By default, the strategy uses the average task execution time for each worker but it can be configured to use the average task event loop utilization (ELU) active time instead.
9e2ab226 JB	20
	21	### Weighted round robin
	22
251f0d71	23	The worker weights are maximum tasks execution time, once the worker has reached its maximum tasks execution time, the next task is assigned to the next worker. The worker default weights are the same for all workers and is computed given the CPU cores speed and theirs numbers.
afe102b2	24
9e2ab226 JB	25	### Interleaved weighted round robin
9e2ab226 JB	26
251f0d71	27	The worker weights are maximum tasks execution time. The rounds are the deduplicated worker weights.
12d8ed1b	28	During a round, if worker weight is inferior to the current round, the next task is assigned to the next worker. Once all workers have been assigned a task, the next round starts.
251f0d71 JB	29	The worker default weights are the same for all workers and is computed given the CPU cores speed and theirs numbers. So the default rounds consists of a unique worker weight.
251f0d71 JB	30
9e2ab226 JB	31	## Statistics
9e2ab226 JB	32
afe102b2 JB	33	Worker choice strategies enable only the statistics that are needed to choose the next worker to avoid unnecessary overhead.
afe102b2 JB	34
9e2ab226 JB	35	### Median
9e2ab226 JB	36
c63b7b38	37	Strategies using the average task execution time for each worker can use the median instead. Median is more robust to outliers and can be used to avoid assigning tasks to workers that are currently overloaded. Median usage introduces a small overhead: measurement history must be kept for each worker and the median must be recomputed each time a task has finished.