Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi,...
Transcript of Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi,...
![Page 1: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/1.jpg)
Cooper: Task Colocation with Cooperative Games
Qiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. LeePresented by: Qiuyun Llull
Duke University
HPCA – Feb 7, 2017
1
![Page 2: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/2.jpg)
Task Colocation in Datacenters
2
Datacenters colocate applications to increase server utilization
![Page 3: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/3.jpg)
3
Colocation Contention
Colocation interference can lead to performance degradation
![Page 4: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/4.jpg)
4
System Setting
• Alvin, Ben, and Dan are working towards HPCA papers.• They share a cluster and divide processors equally.• Ben’s applications are memory intensive.• Alvin and Dan’s applications are not memory intensive.
![Page 5: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/5.jpg)
5
System Setting
![Page 6: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/6.jpg)
6
Strategic Behavior
• Alvin, Ben, and Dan are strategic.• Can smaller, separate clusters improve performance?• Alvin and Dan share separate cluster to improve performance.
![Page 7: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/7.jpg)
7
Strategic Behavior
![Page 8: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/8.jpg)
8
Without incentives, strategic users may…• Bypass common management policy• Migrate tasks for better colocations• Procure private machines
Strategic action fragments cluster and harms efficiency
Strategic Behavior
![Page 9: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/9.jpg)
9
Pursues Performance• Predicts contention quickly and accurately• Colocates tasks for system performance• Colocates tasks with complementary demands
Neglects Incentives• Overlooks strategic behavior• Fails to encourage users to colocate
Prior Research
![Page 10: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/10.jpg)
10
Stability• No group of users break away to form separate system
Satisfied Preferences• More users colocate with preferred tasks
Fair Attribution of Costs• Users that contribute more to contention suffer higher losses
Incentivizing Colocation
![Page 11: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/11.jpg)
11
Example
PreferencesA : B > C > DB : A > C > DC : A > B > DD : C > A > B
![Page 12: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/12.jpg)
12
Example
PreferencesA : B > C > DB : A > C > DC : A > B > DD : C > A > B
![Page 13: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/13.jpg)
13
A framework that incentivizes strategic users to colocateby providing desirable system outcomes:
• Stability• Satisfied Preferences• Fair Attribution of Costs
Cooper
![Page 14: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/14.jpg)
14
• System Setting• Incentivizing Colocation• Cooper Colocation Framework• Evaluation
Agenda
![Page 15: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/15.jpg)
15
• Strategic agents are users and tasks• Utility is task performance• Colocation preferences describe preferred co-runners• If u(A,B) > u(A,C), then A prefers B over C
• Actions are -- participate or break away
Cooperative Game
![Page 16: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/16.jpg)
16
Colocations are stable when no group of users can improve their performance by changing colocation.
Game Equilibrium
![Page 17: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/17.jpg)
Cooper Framework
17
Query Interface
Preference Predictor
ActionRecommender
Users
Agents Coordinator
System Profiler
ColocationPolicies
Job Dispatcher
Machines
![Page 18: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/18.jpg)
Colocation Policies
18
Query Interface
Preference Predictor
ActionRecommender
Users
Agents Coordinator
System Profiler
ColocationPolicies
Job Dispatcher
Machines
![Page 19: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/19.jpg)
Matching people in life
19
![Page 20: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/20.jpg)
20
Stable MatchingAlgorithm partitions tasks into two sets• Tasks in one set propose.• Tasks in other set accepts, rejects.
Task updates co-runners • Accept proposal if performance improves
Algorithm terminates when all tasks matched
[1] D. Gale and L. Shapley, “College admissions and the stability of marriage,” American Mathematical Monthly, 1962. [2] R.W.Irving, “An efficient algorithm for the stable roommates problem,” Journal of Algorithms, pp. 577–595, 1985.
![Page 21: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/21.jpg)
21
A
F
E
D
B
C
D>F>E
E>F>D
E>D>F
A>B>C
A>C>B
C>B>A
Stable Matching
![Page 22: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/22.jpg)
22
Stable PoliciesStable Marriage Random (SMR)• Partition tasks randomly
Stable Marriage Partition (SMP)• Partition tasks with domain-specific knowledge• Memory-intensive tasks propose
Stable Roommate (SR)• No partition• Any task proposes to any other.
![Page 23: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/23.jpg)
23
Baseline PoliciesGreedy (GR)• Colocate tasks to minimize performance loss
Complementary (CO)• Colocate tasks with complementary resource demands
![Page 24: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/24.jpg)
Preference Predictor
24
Query Interface
Preference Predictor
ActionRecommender
Users
Agents Coordinator
System Profiler
ColocationPolicies
Job Dispatcher
Machines
![Page 25: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/25.jpg)
25
• Profile colocation performance with sparse samples
• Rate co-runners with profiles
• Predict ratings with collaborative filtering• Infer ratings based on task similarity• Suppose A: B > C and A is similar to D• Then D: B > C
• Construct preference list per task based on ratings
Preference Predictor
![Page 26: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/26.jpg)
Action Recommender
26
Query Interface
Preference Predictor
ActionRecommender
Users
Agents Coordinator
System Profiler
ColocationPolicies
Job Dispatcher
Machines
![Page 27: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/27.jpg)
27
• Assess assigned matches for each task
• Search preference list for better co-runners• Suppose X: A > B, and X matched to B• X messages A to suggest new match
• Recommend break away • Suppose A also prefers X over assigned match. • X, A should break away
Action Recommender
![Page 28: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/28.jpg)
Cooper Recap
28
Query Interface
Preference Predictor
ActionRecommender
Users
Agents Coordinator
System Profiler
ColocationPolicies
Job Dispatcher
Machines
![Page 29: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/29.jpg)
29
• System Setting• Incentivizing Colocation• Cooper Colocation Framework• Evaluation
Agenda
![Page 30: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/30.jpg)
Workloads• PARSEC for multithreaded benchmarks• Spark for task-parallel machine learning
System Measurements• 10 nodes, each with 2 processors and 24 cores• Two tasks share a processor each with half the cores
System Simulation• 500 nodes with varied task populations• Simulate colocations with system profiles
30
Experimental Methods
![Page 31: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/31.jpg)
Fair Attribution of Costs
31
Stable Marriage Random (SMR)
swap
t.bo
dytr.
dedu
pca
neal
svm
linea
rstr
eamc.
decis
iongr
adien
tna
iveco
rrelat
.0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
Thro
ughp
ut P
enal
ty
Complementary (CO)sw
apt.
body
tr.de
dup
cane
alsv
mlin
ear
strea
mc.de
cision
grad
ient
naive
corre
lat.0.00
0.05
0.10
0.15
0.20
0.25
Thro
ughp
ut P
enal
ty
Tasks that contribute more to contention suffer higher penalties
x-axis sorts applications by memory intensity
![Page 32: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/32.jpg)
32
Satisfied Preferences
More users colocate with preferred tasks.
SR/CO SMR/CO SMP/CO
020
060
010
00N
umbe
r of A
gent
s
Improved PerformanceUnchanged PerformanceDegraded Performance
![Page 33: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/33.jpg)
33
Stability
Fewer users break away to form separate system
0 1 2 3 4 5
020
040
060
080
0
CO
0 1 2 3 4 5
020
040
060
080
0
GR
0 1 2 3 4 5
020
040
060
080
0
SMR#
user
s br
eakin
g aw
ay
x-axis: gains (%) for which tasks break away
![Page 34: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/34.jpg)
Loss
Rel
ative
to S
tand
-Alo
nestable matches
0.1
0.0
0.1
0.2
0.3
0.4
0.5
34
Performance
Stable colocations preserve system performance
![Page 35: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/35.jpg)
More in the paper …Cooper Implementation• Profiler and preference predictor• Adapted matching algorithms• Action recommender and job dispatcher
Cooperative Game Theory• Shapely value for fair division• Extending beyond pairs
Experimental Results• Sensitivity to system scale and job mix• Comprehensive policy comparisons
35
![Page 36: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/36.jpg)
ConclusionCooperative Games for Shared Systems• Formalize interactions between strategic users• Incentivize user participation• Enable fair task colocation
Management Desiderata• Fair attribution of costs• Satisfied preferences• Stability
Fairness versus Performance• Stable colocations satisfy more users• Stable colocations preserve system performance
36
![Page 37: Cooper: Task Colocation with Cooperative GamesQiuyun Llull, Songchun Fan, Seyed Majid Zahedi, Benjamin C. Lee Presented by: Qiuyun Llull Duke University HPCA – Feb 7, 2017 1. Task](https://reader035.fdocuments.in/reader035/viewer/2022071414/610cea61c008371b6c75bd04/html5/thumbnails/37.jpg)
Q & A
Thank you!
35