UZH.14

SMSCG (phase 4)

Long Title: Swiss Multi-Science Computing Grid (phase 4)
Leading
Organization:
Universität Zürich
Participating
Organizations:
Universität Bern
Université de Lausanne
Eidg. Forschungsanstalt für Wald, Schnee und Landschaft (WSL)
Haute École Spécialisée de Suisse occidentale
SWITCH - Teleinformatikdienste für Lehre und Forschung
Domain: Grid
Status: finished
Start Date: 01.01.2013
End Date: 30.04.2013
Project Leader: S. Maffioletti
Deputy Project Leader: P. Flury
Website: http://www.smscg.ch

SMSCG provides shared computational resources and basic data management facilities for the Swiss scientific community to solve computational problems. This fourth phase will focus on operational costs, operational services and support.

(see also phase 1, phase 2 and phase 3)

Results

Component Description
SMSCG website The project webpage documents the services and points to the details on the available resources (monitor), their current usage, the supported application portfolio, the communities, and the code and software repositories.
Applications Continue growing of the dedicated HTC solutions mostly developed with the GC3Pie framework.
Increased usage of the infrastructure From this page it is possible to display detailed accounting information. Both number of jobs and used walltime can be classi ed per available resource and/or per Virtual Organization (only for registered users)
Sensitivities and uncertainties of modeled ground temperatures in mountain environments Publication by Gubler, S., Endrizzi, S., Gruber, S., and Purves, R. S. Geosci. Model Dev. Discuss., 6, 791-840, doi:10.5194/gmdd-6-791-2013, 2013.

We created a computational service to solve computationally intensive problems which is used by researchers throughout Switzerland. The service comprises computing and storage infrastructure, a support network of experts located at each participating institution, and a set of powerful tools for efficient and effective use of the infrastructure.
Researchers have access to a reliable and powerful computational service. Combined with our competent assistance and support we make them excel in their research; since 2008.
Project members have cooperated in building up and sharing knowledge. They benefit from access to this knowledge pool for their decisions on investments, on support models, and on technological choices.
Partners, like SwiNG and CRUS, benefit from the people network we established. We connect user communities, service providers, and funding agencies.

Researchers with computational problems often have no access to performant infrastructure or need to build up considerable knowledge to exploit the capabilities of such an infrastructure: this is a bottleneck for their research limiting the scope of and dictating the advance in their research.
Our service eliminates this bottleneck. SMSCG bridges the gap between the computational part of the research problem and the research question the researches is solving by offering consulting, support, and computational infrastructure. Researchers can again focus on what they do best: their research.


Goals

The SMSCG-IV aims

  • at the infrastructure level to further reduce operational costs by capitalizing on the experiences and on the knowledge gathered in the previous phases; this will be applied, for example, to optimize the processes and to compile supportive quality documentation. The project will continue evolving towards new technologies and models that can provide more effective services to the user communities.In this, emerging High- Throughput mechanisms for data analysis (e.g. Hadoop/MapReduce) as well as new application provisioning models (e.g. virtualization/cloud) will be further explored. This will increase the competence portfolio of the participating members and allow to perceive the infrastructure as an evolving system always able to intercept and react on the user needs.
  • at the support and operational services level to avoid effort duplication for our Swiss members that are internationally involved by alignment with international efforts (e.g. NGI-CH).
  • at the user and application support level by disseminating further the model where researchers benefit from a close collaboration with expert support teams for enabling their high throughput computational needs.

These goals altogether will foster a committed SMSCG community, which is sustainable beyond April 2013.

Benefits

The benefits for the two main stakeholders are:

  • User groups benefit from experienced IT consulting group for the porting of their application and of their usecases.
  • Infrastructure providers have the opportunity to work closely to research groups and learn how their usecases get mapped into infrastructure requirements. Some key ones are referenced to illustrate the impact that they have on the infrastructure level. In this way, the whole project is continuously learning and improving on how non-trivial usecases can efficiently be tackled from both the application integration and from the infrastructure point of view.

Steps

The work is divided into the following work packages:

  1. Operation and Infrastructure:
    • Reduce of operational costs of the SMSCG infrastructure by streamlining operational processes and provision of supportive high quality documentation.
    • Eliminate duplicate operation (and funding) of central services that can be outsourced to international bodies like NorduGrid ARC and EGI.eu.
    • Evaluate Hadoop and similar emerging high-throughput technologies for scientific applications.
    • Evaluate new infrastructure hosting/integration models (benefits/risks).
  2. User Support and Recruitment:
    In addition to standard user support SMSCG will further invest into consulting, expand in these activities by having more dedicated (not at least as a result of having reduced operational activities) and knowledgeable persons within the participating universities to actively engage with the local power communities.
  3. Project management:
    will focus mostly on the preparation for the new funding program:
    • Reaching out communities and institutions of interest.
    • Gather interests for potential future project proposals.
    • Discussing on the alignment for the new funding program (revision of sustainability plan).

Back