On-line computing via hardware and software support.
Fast turnaround computing to support experiments during data taking.
Batch processing computing for analysis of data acquired at Fermilab and for Monte Carlo calculations.
General-purpose batch and interactive computing for program development and analysis, project management and tracking, and document preparation.
Extensive local and wide-area networks to facilitate communication.
Support of many local workgroup computing activities.
The Memorandum of Understanding
After a proposed experiment receives Fermilab approval, the experiment and the Computing Division reach an understanding about how much and what kind of computing resources the experiment will use. The memorandum of understanding spells out this understanding, and the Computing Division allocates resources according to their availability and in line with laboratory priorities. Formal meetings or consultations are used to schedule major production activities; frequent consultation between the experiment and the division deals with major development projects. The division also provides some general computing resources for smaller efforts on a fair share basisóthat is, the division doesnít favor an experiment in the use of computing, except at the level of experiments running versus experiments not running.
The division and the experiment periodically review the MOU. The spokesperson should schedule a review of the experimentís computing status and needs at least every three months.
The Computing Division aims to provide the level of reliable computing your collaboration needs. If problems or shortfalls arise, you can assume that the division will help to resolve them within the capabilities of its human and financial resources.
The offline area of the Computing Division, the online area, and the experiment all provide liaisons. The liaisons facilitate communication between the division and the experiment, identifying needs and allocating resources.
Online Liaison. During development of the MOU, the Computing Division assigns one or more online liaisons to the experiment. This liaison provides the primary day-to-day contact between the experiment and the online support arm of the Computing Division. Online liaisons actively connect an experiment to the large body of expertise in online and data acquisition hardware and software in the Computing Division.
Offline Liaison. Many experiments, including all running experiments, have an offline liaison assigned from the pool of physicists and computer scientists in the Division. The liaison keeps the experiment informed about Computing Division activities, policies, resources, limitations and plans; and communicates to the Computing Division the experimentís needs, plans, problems and concerns. The liaison keeps the experiment and the division working together to accomplish the experimentís offline computing tasks. The liaison reviews and advises on all major requests for resources, passes them for implementation to the relevant departments, and follows up to ensure that they have been satisfied promptly and to good effect.
Experiment Liaisons. For its part, the experiment designates an online contact person to serve as a coordinator for the experiment in matters relating to online computing and data acquisition. In addition, each experiment appoints a person generally available at Fermilab and familiar with the offline computing of the experiment to serve as a liaison from the experiment to the division. The experimentís offline liaison provides information about computing to the experiment and channels feedback from the experimenters to the division. The experimentís major requests for resources must come through this person.
Liaisons make possible clear and consistent communication. The experiment offline liaison helps the experimentówhich may have more than 100 people actively engaged in data analysisóspeak to the division with one voice. By the same token, the Computing Division has almost 200 people; and, while experimenters can certainly address their requests to people other than the Computing Division offline liaison, we do ask that experiments review any major commitment of resources with the Computing Division offline liaison in advance.
The experiment offline liaison and the Computing Division offline liaison negotiate major requests for offline resources. If the request doesnít conform to the general groundwork in the experimentís MOU, it may be necessary to write an addendum to the MOU or otherwise come to written agreement. Small requests can go through normal channels, such as request forms; the Computing Division will address any questions about the request to your experimentís offline liaison.
The Computing Division maintains a central pool of electronics (PREP) and computing equipment for use in experiments. All equipment purchased by the Computing Division as part of an experimentís MOU comes from this pool.
Experimenters receive equipment either by over-the-counter issue or on-site installation. The Computing Division issues instruments and small equipment items directly to the experimentís representatives. Note that the division issues over-the-counter equipment only if the MOU calls for it; thus, experimenters should make sure they request enough equipment to prevent shortfalls. Computer systems are usually pre-configured and delivered in place by the various hardware groups. The MOU process specifies details of such installations.
Where to Find the Services You Need
Most Computing Division staff offices are on the sixth and eighth floors of Wilson Hall, and the second and third floors of the Feynman Computing Center.
Computer Accounts. To get an account on the central mail server (FNAL), the central UNIX Cluster (FNALU), the CDF central UNIX or VAX Cluster (CDFSGA and FNALD), and the DZero UNIX and VAX Clusters (D0CHA and FNALD0), fill out a form from the office on the eighth floor crossover of Wilson Hall. Account request and other forms are also available on the Internet at http://www.fnal.gov/ cd/main/forms.html
This form requires your Fermilab ID number, your signature, and the signature of your supervisor or experiment spokesperson. Return the form to the eighth floor crossover, and you will receive your account on the same working day, or, at the latest, the next working day.
Other forms allow you to make changes in resource allocation (such as FNALU disk quota changes), networking requests, and PREP equipment requests. If you or your experiment needs to vault magnetic tapes, you can obtain a code for external labeling. Some request forms are available online in INFO (see below) on the FNALU cluster.
Help When You Need It. During business hours, the Customer Support Help Desk, (ext. 2345) serves as the primary point of contact for general and offline computing issues. Computer and Communications Repair Services can be reached at ext. 4373. After hours, the operations supervisor (ext. 2746) deals with Central Facility operational or emergency issues. The operations supervisor can page on-call support personnel during off hours. Potential computer security incidents should also be reported immediately to the Help Desk (discussed more fully in Chapter 2 of this publication).
Send non-urgent questions, suggestions, or problems by electronic mail to firstname.lastname@example.org HELPDESK is read regularly on weekdays.
Call your experimentís online liaison (the Computing Division office at ext. 3690 can tell you who it is) when you have questions and concerns about online and data acquisition systems. Departments within the Computing Division provide help on detailed technical questions as the need arises.
Computing Division Library. The Computing Division library, WH8NE, has limited local documentation in self-service cabinets. The primary source for documentation is the web, starting either from the Computing Division home page (http://www.fnal. gov/cd/) or from the documentation page to do detailed searches (http://cddocs.fnal.gov/ cfdocs/productsDB/docs.html). The stockroom carries some UNIX manuals and other commercial books. You must order other manuals, including manuals for most proprietary products, from the vendor. In many cases the librarian can provide ordering information.
News and Information. The Computing Division publishes a regular newsletter on the World Wide Web to announce new features, programs, changes, ES&H issues and other timely information, as well as general or educational material. Current and back issues can be viewed at http://www.fnal.gov/ cd/CDN/
Each central computer system has a news service, updated frequently, with much of the same information. New announcements are presented at system login. Some information is also posted to Usenet news groups, but virtually all information is posted on the appropriate Web pages
Software and Documentation. The Physics Analysis Tools and Online Systems Departments and the UNIX Applications Support Group support a large selection of general-purpose and physics-related software, and many of the products are supported on all of the major platforms. Usersí guides to the various central computers describe how to access the software on that computer. The Computing Division library has a list of the available software and local documentation.
The Computing Division library also distributes online and data acquisition documentation, again all available on the Web.
A system named UPS, for ìUnix Product Support,î supports versions of programs that run on UNIX platforms; you may obtain them by running UPD (Unix Product Distribute). The document ìUNIX at Fermilabî has information about this system.
Connections to Networks. The Computing Division provides and operates an extensive Fermilab local and wide-area network, and regulates its access and use. The Computing Division must approve and coordinate any attachment to the network. Because of the importance of networks to the laboratoryís operations, Fermilab reserves the right to protect the integrity of the network.
If an experiment plans to use the network for intensive local data transfer, such as workgroup computing using local area clusters or YP protocols, or other intensive client/server transactions, the Computing Division will provide and manage an appropriate filtering device (bridge or router). Early consultation allows the division enough time to procure and install appropriate equipment.
The Computing Division maintains network name and routing tables for the laboratory. The division assigns network addresses, to guarantee uniqueness; and node names must be registered with the Computing Division to make them available to other users. You can request a form for this purpose by sending an electronic mail message to email@example.com or filling out the form on the Internet.
Completing a connection request form, available on the Web and at the Wilson Hall eighth floor crossover, starts the process of attachment of systems to the network. The Computing Division online or offline liaison must review major requests. Again, we encourage early consultation with the Computing Division, so that we can procure and install appropriate equipment and cabling.
Magnetic Tape Management. A tape retention policy aims to maximize the accessibility of tape data actively being used on systems in the Feynman Center and to provide archival storage for data that may be needed later. To do this, we must get rid of redundant and obsolete tapes. The policy gives a vault quota to each experiment; the number of unique raw data tapes and the number of physicists on the experiment determine the quota, run by run. This quota is halved three years after a given run and and goes to zero after five years. For the details, refer to the document MR0002 available from the Computing Division library.
Acquiring PCs and Workstations. If your experiment decides to acquire personal computers and workstations through the Fermilab Purchasing Department, there is an additional form that must be filed together with the purchase requisition. The eighth floor crossover of Wilson Hall has the guide you need, ìFermilab Guide for Acquiring Personal Computers and Workstations,î along with the format for the required Abbreviated Implementation Plan.
Members of a workgroup and the Computing Division negotiate support for the workgroup cluster. Support varies from occasional consultation, to system support for upgrades, to very intensive involvement. In all cases, however, the workgroup must provide a local system manager to carry out routine functions such as backups and account management. To take full advantage of the capabilities of the Computing Division, members of workgroups should work closely with Computing Division personnel on hardware, networking and software issues as they develop their systems.
Local systems are bound by the same rules for computer use and security as the central system. The local system manager has the responsibility to implement Fermilab rules on the local system. Fermilab may monitor the local system to ensure compliance.
General Consulting. The Computing Division provides general consulting to all Fermilab experimenters. However, if you need help with system or hardware issues on a distributed platform, the consultant will probably refer you to an appropriate support person.
System Support. Distributed workgroup systems have local system managers as well as backup system managers assigned from the Computing Division. If you have a problem that appears to be system- or hardware-related, call your local system manager first.
Software Support. The Computing Division provides support for hundreds of software packages. Many come from vendors, many come from other laboratories, and many are in-house developed. Some are general-purpose; many are specific to the HEP community. If the division supports the product you need, and if licensing allows for distribution to your computer, you can arrange to have it installed.
Hardware Support. The Distributed Computing Hardware Group in the Distributed Computing Department provides hardware support for distributed workgroup computing.
Online and Data Acquisition Support
Online and data acquisition support for experiments covers maintenance and repair of electronics, computer hardware and software installation and support, and development and support of specific hardware and software components, as set forth in the MOU.
The division also serves as a resource of knowledge and expertise for experimenters as they commission their data acquisition and online systems. Many groups in the Computing Division offer outstanding services; you can reach all of them through your experimentís online liaison.
Equipment Maintenance. If the MOU provides for it, Fermilab maintains both instruments and computer systems. Users usually carry malfunctioning instruments into the PREP issue window for exchange and repair by the Online Equipment Support Group.
Computer maintenance usually comes from the Computing Division, via the Distributed Computing Departmentís Distributed Hardware Group, an outside contractor, the Online Data Acquisition Hardware group or the Online In-house Computer Maintenance group. You can call a centralized 24-hour computer trouble line at ext. 4373. During accelerator running times, you can get after-hours support as well.
You can get hardware consultation services in support of CAMAC, NIM and FASTBUS devices via the Online Data Acquisition Hardware Group.
Software Installation. Experimenters should make sure to request appropriate software tools and licenses in the MOU. The experiment has the responsibility to ensure that all necessary software licenses are registered with the Computing Divisionís license coordinator, to meet all legal obligations. Backing up the online liaison, software consultants can help with software and systems areas that the division supports.
Software and System Support. The MOU spells out agreements for software support. The experiment liaison then serves as a channel for this support. The division provides consultation on most software and systems-level products. Experiment representatives usually handle systems management. The extensive networking throughout the laboratory makes most services available via mail utilities or direct contact with the responsible groups.
Responsibilities of Those Using Fermilab Computers
Proper Computer Usage. Lab policy on appropriate computer use and a summary of guidelines is given in Chapter 2 of this publication. The up-to-date lab policy is available on the Internet at http://www.fnal. gov/cd/main/policies.html
Data Integrity. All files containing data or programs that affect the success of the high-energy physics program must have protection against loss or improper manipulation. All users should take precautions to enhance the probability of data recovery in case of unanticipated loss or alteration of their data. Central systems are backed up by the Computing Division; users of PCs or local systems should make sure they understand the backup strategy for their systems.
Care should be taken not to introduce viruses and other such malicious software onto Fermilab computers. Software should only be downloaded from reliable sites. The PC Support Department in the Computing Division can suggest appropriate virus protection tools.
Certain types of data, including some personnel data, proprietary data, financial, procurement and inventory data, may have special confidentiality requirements and require special treatment. Before assembling or accessing such data, consult the CPPM to determine if such special treatment is necessary.
Licensed Software. Computer users at Fermilab often make use of licensed or proprietary software. Most of this software carries with it explicitósometimes implicitólegal restrictions for its use. In general, if the software is on the system in libraries supported by the Computing Division and you use it in the advertised way, its use on that local system is unrestricted. You should assume that you may not make or take a copy away with you unless you have permission from the maintainer of the program, who should understand any restrictions. Most software now has legal protection under patent or copyright law. Employees and users are expected to honor the terms of all licenses (see Chapter 2).
Usersí Responsibilities. Fermilab employees, visitors, and guest users may use the computer systems in support of the high energy physics program. We ask users to refrain from actions that may interfere with use of computers by others. Donít share your computer accountsóyou are responsible for any use that others make of them. Application for and assignment of a Fermilab computer account carries with it acceptance of the responsibilities and rules described in Chapter 2 of this publication.
Non-Fermilab-Owned Machines. You may connect computers not owned by Fermilab to the Fermilab network only with explicit permission from the Computing Division. Once you connect such a computer to any Fermilab-owned machine or local network, it is bound by all rules and policies that apply to Fermilab-owned computers described in Chapter 2. This applies to any computer connecting to Fermilab computing over a wide-area network or over phone lines for the duration of the connection. Fermilab has the right to monitor the use of the computer. Special restrictions concerning the use of licensed or proprietary software may apply to these computers. If such machines are removed from the Fermilab site, any Fermilab - provided licensed or proprietary software should be removed from the computer.