education and training

down to a modest fraction of the total cost of ownership as opposed to the current situation. Test beds should also enable tools for improved usability and programming productivity and avoid a potential setback in these re-gards due to expected changes in the architecture of computer and storage systems.

Education and Training of Active researchers: Education and training of active researchers should be organized differently from training of train-ers. Active researchers are likely to already have experience of the structure or use of resources similar to the ones available in the e-Infra-structure, or have good knowledge of models, methods, algorithms, soft-ware, data, etc. within the domain of research, but they may need to learn how to best use the e-Infrastructure and its associated services, or learn of alternative or new approaches better suited to the specific e-Infrastruc-ture.

Integration into academic curricula: For maximum impact, knowledge of the e-Infrastructure resources (hardware, software, data, services), how to access and use them effectively and associated methodologies, models, algorithms, programming systems etc., should be integrated into standard academic curricula, adapted as needed to the many disciplines using the e-Infrastructure. Though academic disciplines are responsible for their curricula, e-Infrastructure resource and service providers should facili-tate this integration. For example, education in parallel programming was initially predominantly offered by academic computing centres through

“Summer Schools” or code porting workshops but is now, at least for commonly used parallel programming paradigms, integrated into many standard academic courses in many disciplines at the undergraduate level.

Change is, however, a constant in e-Infrastructure, and new material con-tinuously needs to be introduced into academic curricula to ensure that the next generation of e-Infrastructure users are well prepared to effec-tively use and enhance it.

There are many known and well established approaches to effective edu-cation and training with their own advantages and disadvantages. Most are relevant for education and training the e-Infrastructure´s effective use and evolution.

Seasonal Schools (“Summer Schools”): These are common and have been found effective in many disciplines including the education and training of users and developers of e-Infrastructures. Seasonal Schools have been used very successfully by, for example, EGI, PRACE, and XSEDE in regards to e-Infrastructure and by HPC centers with, for example, PDC having trained over 1,000 researchers in its 2-week Summer School, entitled ”Introduction to HPC”.

Domain specific workshops: These enable an exposure to more advanced material with a focus on material that is of specific interest to a commu-nity, or the effective use of particular resources, software, data sets, or code porting techniques, etc. Like Seasonal Schools, domain-specific workshops have been successfully used in regards to e-Infrastructure by EGI, PRACE,

XSEDE and others and by individual resource provides, such as HPC centres.

Web based education and training: Though seasonal schools and workshops are very effective, and offer intangible benefits such as potential new collab-orations (possibly across disciplines), they also have their limitations with regard to accessibility by being available only once or a few times a year and at a few locations and with attendance limitations. Therefore, web-based training should be made available, both in an entirely self-learning mode and in an assisted mode.

Collaboration: Knowledgeable and effective trainers are in short supply and developing quality training materials is very time consuming. Typically the needs are common to many e-Infrastructures, resource providers and communities and, hence, development of training materials and the deliv-ery of training in the form of Seasonal Schools, Workshops, and the Web can be carried out as cooperative efforts. For instance, PRACE and XSEDE cooperate in both these aspects and PRACE is establishing a few advanced training centres focused on the development of training materials and co-operative delivery of training at sites other than the centre.

Outreach: An e-Infrastructure serving the Swedish research communi-ties is a significant resource and investment. Therefore, an outreach effort should be associated with it that not only ensures it is used to its full poten-tial for research, but that its potenpoten-tial to attract new users, the next gener-ation researchers that will move research forward and enhance the society, and generate public good is exploited. To serve such purposes the XSEDE

“Training, Education and Outreach Service” (TEOS) works with campuses and organizations to help instil digital services into the practices of faculty, staff and students. Activities with colleges and departments include: pro-viding support for incorporation of computational science resources and methods into the curriculum; recruiting and training Campus Champions to raise local awareness of digital services; helping campuses and organiza-tions to enhance their cyberinfrastructure resources; and visiting sites to provide digital services needed by the institution´s personnel. To deeply en-gage the community in sustained use of computational services, the TEOS team also works closely with individuals to understand and address their particular needs and challenges. These individuals may be faculty, research-ers, students, IT staff, or administrators. Deep engagement activities include training, workshops, internships, fellowships, and consulting sessions de-signed to address the unique needs, perspectives, and practices of diverse communities. PRACE also has extensive outreach efforts and, in addition, has established a PRACE Young Investigator Award to stimulate the interest of young researchers in the innovative use of its infrastructure. These ef-forts are viewed as effective.