october 23, 2015 1 grid computing: from old traces to new applications fribourg, switzerland...

28
June 23, 2022 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup , Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed, Catalin Dumitrescu, Mathieu Jan, Dick Epema Parallel and Distributed Systems Group, TU Delft Big thanks to our collaborators: U Wisc./Madison, U Chicago, U Dortmund, U Innsbruck, LRI/INRIA Paris, INRIA Grenoble, U Leiden, Politehnica University of Bucharest, Technion, … DGSim The Failure Trace Archive

Upload: kerrie-beasley

Post on 03-Jan-2016

218 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20231

Grid Computing:From Old Traces to New Applications

Fribourg, Switzerland

Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed, Catalin Dumitrescu, Mathieu Jan, Dick Epema

Parallel and Distributed Systems Group, TU Delft

Big thanks to our collaborators: U Wisc./Madison, U Chicago, U Dortmund, U Innsbruck, LRI/INRIA Paris, INRIA Grenoble, U Leiden, Politehnica University of Bucharest, Technion, …

DGSimThe FailureTraceArchive

Page 2: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

About the Speaker

• Systems• The Koala grid scheduler• The Tribler BitTorrent-compatible P2P file-sharing• The POGGI and CAMEO gaming platforms

• Performance• The Grid Workloads Archive (Nov 2006)• The Failure Trace Archive (Nov 2009)• The Peer-to-Peer Trace Archive (Apr 2010)• Tools: DGSim trace-based grid simulator, GrenchMark

workload-based grid benchmarking

• Team of 15+ active collaborators in NL, AT, RO, US• Happy to be in Berkeley until September

April 20, 20232

Page 3: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

The Grid

An ubiquitous, always-on computational and data storage platform on which users can seamlessly run their (large-scale) applications

April 20, 20233

Shared capacity & costs, economies of scale

Page 4: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20234

The Dutch Grid: DAS System and Extensions

VU (85 nodes)

TU Delft (68) Leiden (32)

SURFnet6

10 Gb/s lambdas

UvA/MultimediaN (46)

UvA/VL-e (41)

• 272 AMD Opteron nodes 792 cores, 1TB memory• Heterogeneous: 2.2-2.6 GHz single/dual core nodes• Myrinet-10G (excl. Delft)• Gigabit Ethernet

DAS-4 (upcoming)• Multi-cores: general purpose, GPU, Cell, …

DAS-3: a 5-cluster grid

Clouds• Amazon EC2+S3, Mosso, …

Page 5: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20235

Many Grids Built

DAS, Grid’5000, OSG, NGS, CERN, …

Why grids and not The Grid?Why grids and not The Grid?

Page 6: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20236

Agenda

1. Introduction2. Was it the System?3. Was it the Workload?4. Was it the System Designer?5. New Application Types6. Suggestions for Collaboration7. Conclusion

Page 7: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20237

The Failure Trace ArchiveFailure and Recovery Events

20+ traces online

http://fta.inria.frhttp://fta.inria.fr

D. Kondo, B. Javadi, A. Iosup, D. Epema, The Failure Trace Archive: Enabling Comparative Analysis of Failures in Diverse Distributed Systems, CCGrid 2010 (Best Paper Award)

Page 8: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Was it the System?

• No• System can grow fast• Good data and models to support system

designers• Yes• Grid middleware unscalable [CCGrid06,HPDC09]• Grid middleware failure-prone [CCGrid07,Grid07]• Grid resources unavailable [CCGrid10]• Inability to load balance well [SC|07]• Poor online information about resource availability

April 20, 20238

Page 9: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 20239

Agenda

1. Introduction2. Was it the System?3. Was it the Workload?4. Was it the System Designer?5. New Application Types6. Suggestions for Collaboration7. Conclusion

Page 10: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202310

The Grid Workloads ArchivePer-Job Arrival, Start, Stop, Structure, etc.

6 traces online

http://gwa.ewi.tudelft.nlhttp://gwa.ewi.tudelft.nl

1.5 yrs >750K >250

A. Iosup, H. Li, M. Jan, S. Anoep, C. Dumitrescu, L. Wolters, D. Epema, The Grid Workloads Archive, FGCS 24, 672—686, 2008.

Page 11: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202311

Grid Systems

How Are Real Grids Used?

Data Analysis and Modeling• Grids vs. parallel production environments such

as clusters and (small) supercomputers• Bags of single-processor tasks vs. single parallel jobs• Bigger bursts of job arrivals• More jobs

Parallel production environments

Page 12: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202312

Bags-of-Tasks (BoTs)

Grid Workloads

Analysis: Grid Workload Components

Time [units]

Workflows (WFs)

• BoT size = 2-70 tasks, most 5-20• Task runtime highly variable,

from minutes to tens of hours

• WF size = 2-1k tasks, most 30-40

• Task runtime of minutes

Page 13: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Was it the Workload?

• No• Similar workload characteristics across grids• High utilization possible due to single-node jobs• High load imbalance• Good data and models to support system designers

[Grid06,EuroPar08,HPDC08-10,FGCS08]• Yes• Too many tasks (system limitation)• Poor online information about job characteristics +

High variability of job resource requirements • How to schedule BoTs, WFs, mixtures in grids?

April 20, 202313

Page 14: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202314

Agenda

1. Introduction2. Was it the System?3. Was it the Workload?4. Was it the System Designer?5. New Application Types6. Suggestions for Collaboration7. Conclusion

Page 15: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202315

Problems in Grid Scheduling and Resource Management

The System1. Grid schedulers do not own resources

themselves• They have to negotiate with autonomous local schedulers

• Authentication/multi-organizational issues

2. Grid schedulers interface to local schedulers• Some may have support for reservations, others are queuing-based

3. Grid resources are heterogeneous and dynamic• Hardware (processor architecture, disk space, network)

• Basic software (OS, libraries)

• Grid software (middleware)

• Resources may fail

• Lack of complete and accurate resource information

Page 16: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202316

Problems in Grid Scheduling and Resource Management

The Workloads

4. Workloads are heterogeneous and dynamic• Grid schedulers may not have control over the full

workload (multiple submission points)

• Jobs may have performance requirements

• Lack of complete and accurate job information

5. Application structure is heterogeneous• Single sequential job

• Bags of Tasks; parameter sweeps (Monte Carlo), pilot jobs

• Workflows, pipelines, chains-of-tasks

• Parallel jobs (MPI); malleable, coallocated

Page 17: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202317

The Koala Grid Scheduler

• Developed in the DAS system• Has been deployed on the DAS-2 in September 2005• Ported to DAS-3 in April 2007• Independent from grid middlewares such as Globus• Runs on top of local schedulers

• Objectives:• Data and processor co-allocation in grids• Supporting different application types• Specialized application-oriented scheduling policies

Koala homepage: http://www.st.ewi.tudelft.nl/koala/

Page 18: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202318

Koala in a Nutshell

• Parallel Applications• MPI, Ibis,…• Co-Allocation• Malleability

• Parameter Sweep Applications• Cycle Scavenging• Run as low-priority jobs

• Workflows

A bridge between theory and practice

Page 19: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Euro-Par 2008, Las Palmas, 27 August 200819

Inter-Operating Grids Through Delegated MatchMaking

Inter-Operation Architectures

Hybrid hierarchical/ decentralize

d

Decentralized

Hierarchical

Independent

Centralized

Delegated MatchMakin

g

Page 20: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202320

Inter-Operating Grids Through Delegated MatchMaking

The Delegated MatchMaking Mechanism

1. Deal with local load locally (if possible)2. When local load is too high, temporarily bind resources from

remote sites to the local environment. • May build delegation chains. • Delegate resource usage rights, do not migrate jobs.

3. Deal with delegations each delegation cycle (delegated matchmaking)

Delegate

Local load too high

Resource request

Resource usage rights

Bind remote resource

The Delegated MatchMaking Mechanism=The Delegated MatchMaking Mechanism=Delegate Resource Usage Rights, Delegate Resource Usage Rights,

Do Not Delegate JobsDo Not Delegate Jobs

Page 21: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202321

• DMM• High goodput• Low wait time• Finishes all jobs

• Even better for load imbalance between grids

• Reasonable overhead• [see thesis]

What is the Potential Gain of Grid Inter-Operation?

Delegated MatchMaking vs. Alternatives

Independent

Centralized

Decentralized

DMM

(Higher is better)

Grid Inter-Operation (through DMM)Grid Inter-Operation (through DMM)delivers good performancedelivers good performance

Page 22: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202322

4.2. Studies on Grid Scheduling [5/5]

Scheduling under Cycle Stealing

Scheduler

CS-Runner

Node

submits PSA(s)

JDF

grow/shrink

messagesregisters

Clusters

Launcher

Launcher

Head Node

KCMKCM

submitslaunchers

deploys, monitors,

and preempts

tasks

monitors/informs

idle/demanded resources

CS Policies:• Equi-All: grid-wide basis • Equi-PerSite: per cluster

CS Policies:• Equi-All: grid-wide basis • Equi-PerSite: per cluster

Application Level Scheduling:• Pull-based approach• Shrinkage policy

Application Level Scheduling:• Pull-based approach• Shrinkage policy

Launcher

Launcher

O. Sonmez, B. Grundeken, H. Mohamed, A. Iosup, D. Epema: Scheduling Strategies for Cycle Scavenging in Multicluster Grid Systems. CCGRID 2009: 12-19

Requirements1.Unobtrusiveness Minimal delay for (higher priority) local and grid jobs

2.Fairness3. Dynamic Resource Allocation4. Efficiency5. Robustness and Fault

Tolerance

Deployed as Koala Runner

Page 23: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Was it the System Designer?

• No• Mechanisms to inter-operate grids: DMM [SC|07], …• Mechanisms to run many grid application types:

WFs, BoTs, parameter sweeps, cycle scavenging, …• Scheduling algorithms with inaccurate information

[HPDC ‘08, ‘09, ‘10]• Tools for empirical and trace-based experimentation

• Yes• Still too many tasks• What about new application types?

April 20, 202323

Page 24: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202324

Agenda

1. Introduction2. Was it the System?3. Was it the Workload?4. Was it the System Designer?5. New Application Types6. Suggestions for Collaboration7. Conclusion

Page 25: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Cloud Futures Workshop 2010 – Cloud Computing Support for Massively Social Gaming 25

MSGs are a Popular, Growing Market

• 25,000,000 subscribed players (from 150,000,000+ active)

• Over 10,000 MSGs in operation

• Market size 7,500,000,000$/year

Sources: MMOGChart, own research. Sources: ESA, MPAA, RIAA.

Page 26: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

Cloud Futures Workshop 2010 – Cloud Computing Support for Massively Social Gaming 26

Massively Social Gaming as New Grid/Cloud Application

1. Virtual worldExplore, do, learn, socialize, compete+

2. ContentGraphics, maps, puzzles, quests, culture+

3. Game analyticsPlayer stats and relationshipsRomeo and

Juliet

Massively Social Gaming(online) games with massive numbers of players (100K+), for which social interaction helps the gaming experience

[SC|08, TPDS’10]

[EuroPar09

BPAward, CPE10]

[ROIA09]

Page 27: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202327

Suggestions for CollaborationSuggestions for Collaboration

• Scheduling mixtures of grid/HPC/cloud workloads• Scheduling and resource management in practice• Modeling aspects of cloud infrastructure and workloads

• Condor on top of Mesos

• Massively Social Gaming and Mesos• Step 1: Game analytics and social network analysis in Mesos

• The Grid Research Toolbox• Using and sharing traces: The Grid Workloads Archive and

The Failure Trace Archive• GrenchMark: testing large-scale distributed systems• DGSim: simulating multi-cluster grids

Page 28: October 23, 2015 1 Grid Computing: From Old Traces to New Applications Fribourg, Switzerland Alexandru Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed,

April 20, 202328

Alex Iosup, Ozan Sonmez, Nezih Yigitbasi, Hashim Mohamed, Dick Epema

Thank you! Questions? Observations?

More Information:• The Koala Grid Scheduler: www.st.ewi.tudelft.nl/koala

• The Grid Workloads Archive: gwa.ewi.tudelft.nl

• The Failure Trace Archive: fta.inria.fr

• The DGSim simulator: www.pds.ewi.tudelft.nl/~iosup/dgsim.php

• The GrenchMark perf. eval. tool: grenchmark.st.ewi.tudelft.nl

• Cloud research: www.st.ewi.tudelft.nl/~iosup/research_cloud.html

• Gaming research: www.st.ewi.tudelft.nl/~iosup/research_gaming.html

• see PDS publication database at: www.pds.twi.tudelft.nl/

email: [email protected]

Big thanks to our collaborators: U. Wisc.-Madison, U Chicago, U Dortmund, U Innsbruck, LRI/INRIA Paris, INRIA Grenoble, U Leiden, Politehnica University of Bucharest, Technion, …

DGSim