Through the Strategic Technologies for Cyberinfrastructure program, NSF has granted funding for a collaborative project between the Fermilab and the Information Sciences Institute. Project members will develop a resource provisioning system, building on work previously done at each of these institutions. Burt Holzman, Associate Department Head of CMS Computing Facilities, is a principal co-investigator and will manage the project.

In short, grid resource provisioning is a way to provide access to computing resources without requiring the user to know details about the underlying grid infrastructure (i.e., on which sites they have permission or enough resources to run their jobs). Fermilab's glideinWMS work management system, originally developed by Igor Sfiligoi and now maintained by Holzman, is a tool that provides this function on the Open  Science Grid (OSG). As a user submits their grid job, the Condor software underneath matches the job with the computer that will execute it. As part of Fermilab's work on the project, glideinWMS will be extended and integrated with Corral, its TeraGrid counterpart, to produce a more comprehensive tool called CorralWMS.

In addition, the grids will collaborate on facilitating the running of MPI jobs. The ability to run these jobs—which use the Message Passing Interface specification to manage communications between parallel tasks—is essential to many of the research communities across both OSG and TeraGrid.

The overall goal of the project is to provide an integrated, robust, scalable, and flexible service that will allow users from either OSG or TeraGrid to access both provisioned (previously negotiated) resources and “just-in-time” available resources for a single set of user jobs. CorralWMS will have a common interface that will enable access to new types of resources (such as clouds) and to common monitoring capabilities, tailored views for monitoring grid jobs in real time, and a comprehensive summary of statistics.

Besides contributing project management and extending glideinWMS, Fermilab will package and distribute the new tool through the Virtual Data Toolkit (VDT), the OSG software bundle, and will perform scalability and performance testing. A term position is open for doing this work under Burt's leadership.

~Marcia Teckenbrock and Burt Holzman

Configuration Management

Within the ITIL framework, Configuration Management is the discipline that identifies, records, controls, and reports on the IT infrastructure components— hardware, software, documentation, services, personnel, and any other items that an organization wants to track (called Configuration Items or CIs). Configuration Management helps to ensure the completeness and correctness of these items.

For example, say you're tracking web servers. Configuration Management might require current records of everything about the server, from its physical location to its security settings. In addition, the process of keeping this information up to date and measuring performance must be determined.

This month, several members across CD participated in a workshop, beginning to plan for this crucial process area. Members of the Core Team defined Configuration Management policies, roles and responsibilities, the types of CIs to track initially, the process and procedures required to change a CI's attributes, key performance indicators (KPIs, which allow you to measure performance), and other items.

Further planning and deployment of Configuration Management will continue over the next couple of months.

Change management

Also as part of our adoption of ISO20000, we are currently implementing the Change Management process— probably the most culturally significant process, since it changes the way we do some of our basic activities. The biggest difference you'll notice once Change Management is adopted is increased communication regarding the type of changes that will occur and the impact they will have on existing applications. For example, how many times have you noticed an outage, only to find out later it was because someone in another department changed something in their application that broke how yours worked? Or how often have you tried to upgrade your application while someone in another department wanted to repair the server that your application resided on?

Change Management is designed to accomplish a few goals. Change Management…

•  Ensures that standardized procedures are used for efficient and prompt handling of all changes

•  Minimizes the impact of change-related incidents on the quality of service

•  Improves day-to-day operations

The end result provides these benefits:

•  Improved alignment of IT services to the experiments' requirements and lab related operations (Do the experiments and the lab really need this change to occur?)

•  Increased visibility and communication of changes to both research and service support staff (Not only do our users know about the change, but the other teams will know about it too.)

•  Improved risk assessment (Have we really covered all of our bases and made sure that we can successfully implement this change?)

•  Increased productivity of end-users (Users will know when we're making changes, so they can plan around work done on the systems.)

•  Reduced adverse impact of changes (Better planning and less conflicts mean that changes will be less painful after they're rolled out.)

We all work hard supporting the lab and the science we produce. Change Management is designed to help us to provide a consistent high-quality of support.

Training sessions for both the Change Management process and the Remedy tool are upcoming.

Everyone is expected to participate in the fire drills we will be holding in Wilson Hall and FCC during the week of October 5th.

Wilson Hall will run several staggered drills on Thursday, October 8, 2009 (rain date is Friday, October 9th). Drills will be held in phases at the times below:

FLOORS

•  15, 14, 13 - 8:30 am
•  12, 11, 10 - 9:00 am
•  9, 8, 7 - 9:30 am
•  6, 5, 4, 3, 2, Atrium, Mezzanine, and Ground Floor - 10:30 am

Wilson Hall residents should leave the building and gather in the established assembly areas (LINAC, Cross Gallery and Transfer Hall sidewalk/lawn) across the circular drive behind the Auditorium. Supervisors or delegated individuals submitting a drill critique can print out critique forms, which need to be submitted to John Kent (MS 214) by Thursday, October 15, 2009.

FCC is scheduled to have a fire drill on Friday, October 9, 2009 at 10:30 am, and again, everyone is expected to participate.

If you need to be excused from participating in the drill for medical reasons, please contact the Medical Office for a Medical Evaluation Form. You or your supervisor should inform your emergency warden of this and arrange a time with them to review emergency procedures.

~Amy Pavnica