GGF18 Schedule
GridWorld/GGF18
September 11-14, 2006
Washington DC, USA

Monday, September 11
2:00 pm - 3:30 pm
Digital Sherpa: Custom Grid Applications on the TeraGrid and Beyond (90 mins)
Ron Price
View Participants

Digital Sherpa is a grid application for executing High Performance Computing (HPC) applications across many grid enabled HPC resources. It automates non-scalable tasks such as staging, job submission and monitoring, including recovery features such as resubmission of failed jobs. The goal is to allow any HPC application to easily interoperate with Digital Sherpa to become a custom grid application. Distributing the jobs across HPC resources increases the amount of computer resources that can be accessed at a given time. Success using Digital Sherpa has been found on the TeraGrid and there are many more applications of Digital Sherpa in progress.


Digital Sherpa is a grid application that allows any other application or user to have access to more resources, improve load balancing and automate tedious and time consuming tasks such as job submission and monitoring in multiple computer systems. Digital Sherpa takes its name from “sherpa” who are known for their great mountaineering skills in the Himalayas, expert route finders and porters. Like the “sherpa” in mountaineering, Digital Sherpa helps an application or user to find their route (available HPC resource that meet their needs), login for you via Grid Security Infrastructure (no more logging in to several machines and trying to keep an eye on several queues), carry some of your gear for you (stage data in), climb to the top (execute job and restart job if necessary) and carry gear out for you (stage data out). Currently Digital Sherpa has the ability to do everything except for choosing the best HPC resource, although it is planned to be implemented in the final stages of the project.


Another main feature of Sherpa is that is a program and not a job oriented application. Recently there has been an increase in the number of HPC programs that require more than one job to complete a task. This is because the total amount of resources that a job can execute is not given by the logical need of the program, but by HPC resource policies where the job is executed.

Benefits of Using Sherpa:

* Automation of login, data stage in and stage out, job submission, monitoring, and auto restart if the job fails.
* Having Sherpa distribute your jobs across various HPC resources can increase the amount of resources that can be used at the same time. This also can be seen as reduction the queue wait time by submitting jobs to several queues.

Digital Sherpa has been designed and planned to be scalable, maintainable and used directly by people or other applications. It is based on Web Services Resource Framework (WSRF) and it is implemented using Globus Toolkit 4.0 (GT4). We have successfully submitted basic jobs using Digital Sherpa across the following GT4 enabled sites: Purdue, NCSA Mercury, UC/ANL, Indiana and Utah.

Agenda:
Overview of Digital Sherpa

Reasons for Creating Digital Sherpa, Goal of Sherpa and Conceptual Details

Digital Sherpa Implementation: Choice of API, Past and Present

Digital Sherpa Application and Results: Basic Jobs to Real Science

Benefits of Using Digital Sherpa: Automation of Tedious HPC Required Tasks, Less Time Spent in the queue

Future Plans for Sherpa: Resource Discovery via MDS and Proxy Auto Renewal.

Q&A With Audience


Location: 154B
 
Rate This Session:
Rating: Comments:

 
    Slides:     Digital Sherpa: Custom Grid Applications on the TeraGrid and Beyond

> login   RSS RSS Contact Webmaster

OGFSM, Open Grid ForumSM, Grid ForumSM, and the OGF Logo are trademarks of OGF