Www.cs.wisc.edu/~miron Condor Team 2008. Welcome to Condor Week #10 (year #25 for the project)

Preview:

Citation preview

www.cs.wisc.edu/~mironCondor Team 2008

www.cs.wisc.edu/Condor

Welcome toCondor Week

#10(year #25 for the project)

www.cs.wisc.edu/Condor

Goodbye GridsWelcome Clouds!

we survived grids, weare ready for clouds and

prepare for what will follow

www.cs.wisc.edu/Condor

Subject: PSFrom: Ian Foster <foster@anl.gov>Date: Thu, 4 Dec 2008 19:40:28 -0600To: Miron Livny miron@cs.wisc.edu

Miron:

I visited CSIRO today and they told me that they are deploying a "desktop cloud." I asked them what "cloud software" they are using, and I was pleased to hear that they were using Condor. So Condor is officially cloud software, at least in Australia. I thought you'd like to know.

Ian.

www.cs.wisc.edu/Condor

Stay away from

the hype and stick to the

principals!

www.cs.wisc.edu/Condor

High ThroughputComputing

became a well established and widely used paradigm

www.cs.wisc.edu/Condor

Condor-GGlide-ins

Matchmaking

carry the load of EGEE and OSG

www.cs.wisc.edu/Condor

Virtual Machines

fit our principals

(they look like a job, they ‘come and go’, they

checkpoint, they have many attributes,

the can be matched, … )

www.cs.wisc.edu/Condor

Innovation through widely

adopted technology

partnerships with users in

academia and industry, and IT providers

Miron LivnyCenter for High Throughput Computing

Computer Sciences DepartmentUniversity of Wisconsin-Madison

The Condor Project – an Experiment in Experimental

Computer Science

www.cs.wisc.edu/~miron

Why am I here today?

Because I believe that: the Condor project is ‘different’, we do not have many (any?) other such Computer

Science projects, it is important for the future of our science to have more

(many?) such projects, we can and should learn from what we have done so far

what it takes to build and sustain such projects and CISE can and should play a leadership role in changing

the dynamics of our field so that more Computer Science departments will have such projects

www.cs.wisc.edu/~miron

“Why are you leaving academia and taking a job

in industry?”

“I want to have

impact!”

www.cs.wisc.edu/~miron

In the words of Mike Carey

“I left academia for industry because I was drawn to the idea of getting more direct access to real problems - from customers and challenges encountered while building commercial-grade software - because I felt like I was in somewhat of a mode of inventing and solving problems, at least w.r.t. some of the things I'd been working on.  Sure, that was leading to many written/submitted/accepted papers, but it was somehow less than satisfying after awhile.”

www.cs.wisc.edu/~miron

It is all about Software

I argue that ‘we’ do not know how to develop, maintain, support and evolve

(dependable) software, to integrate software components into end-

to-end (dependable) capabilities, to evaluate the quality of software, to estimate the cost (effort) related to

software and to treat software as infrastructure

www.cs.wisc.edu/~miron

(my) terminology

› Experiment - an act or operation for the purpose of discovering something unknown or of testing a principle, supposition, etc.:

› Technology Adoption – to select a technology as a means to meet an ends of significant importance/value

› Real users – individuals or groups who adopt (and use) a computing technology

› Experimental Computer Science – advance the state of the art of computing (new frameworks, new technologies, new abstractions) through experiments that involve real users

Two new Institutes on the UW Campus - MIR & WID

16

The MIR Three-Plane Model

17

Societal impact through education, research and technology

Human, physical and cyberinfrastructure

Summit Plane

Integration Plane

Foundation PlaneScientific Research Areas

Delivery avenues

Intellectual environment

Knowledge Discovery Base

Figure 1 The MIR Three-Plane Model

Computer Systems and Operations

Clemson Computing and Information Technology

Using Condor with BlackboardSam Hoover

IT Systems ArchitectComputer Systems and Operations

CCIT, Clemson Universityshoover@clemson.edu

Computer Systems and Operations

Clemson Computing and Information Technology

Problem Statement

• Courses in Blackboard need to be archived

• Blackboard provides a batch archive script• Over 6,200 courses are currently active• The course list was split into equal

numbers across 5 servers• Archives took 60 hours to complete

Computer Systems and Operations

Clemson Computing and Information Technology

Implementing Condor with Blackboard

• Blackboard is a “heavy” Java Enterprise application

• We wanted to take advantage of the multiple cores per server

• Security is a requirement• Performance of the application during

processing MUST be maintained

Sam Hoover: shoover@clemson.edu Sam Hoover: shoover@clemson.edu

Computer Systems and Operations

Clemson Computing and Information Technology

Benefits of using Condor with Blackboard

• Job scheduling ensures load balancing and enables higher throughput through use of all available hardware resources

• Post processing with DAGMan streamlines operation

• Jobs are suspended and resumed automatically to protect user performance

• Archive time has been reduced by 65%• End of Semester “full” archive is now feasible

Computer Systems and Operations

Clemson Computing and Information Technology

Credits

• Clemson could not have achieved this innovation without the contributions of the following people and groups:• Sam Hoover, CSO - Principal Innovator• Randy Martin and Matt Garrett, CSO• NSF CI-TEAM award (RENCI principal)• Sebastien Goasguen, School of Computing• Blackboard Operations Team (CCIT-wide

representation)

•  

www.cs.wisc.edu/~miron

Thank you for building such

a wonderful community

Recommended