WP2RequirementsInstrument

From EUAGwiki

(back to Requirements Capture)

Contents 1 WP2 Requirements Instrument 1.1 What questions do you think we should ask the scientific communities? 1.2 How important do you think the following kinds of issues will be in the application areas you are interested in? 1.3 What would you say are the main obstacles for the wider uptake of e-Science in your country and in the application areas you are working with? How might they be addressed? 1.4 In your opinion, what are the main sustainability challenges and how could they be addressed? 2 Collection of questions from 21st May meeting 3 WP2.1 Survey of Requirements 3.1 First Draft 3.1.1 Personal Information 3.1.2 General Information 3.1.3 Use of Grid Technologies 3.1.4 Computing Resources 3.1.5 Network Connectivity 3.1.6 Data 3.1.7 Human Resources 3.1.8 Software 3.1.9 Finally 3.1.10 References

[edit] WP2 Requirements Instrument

[edit] What questions do you think we should ask the scientific communities?

• application requirements, including package/program, system and scientific library, hardware, data rate, size and growth, data management and operational issues. 2. The way how they collaborate with other research groups

• With whom you collaborate? Do you need to share data with your collaborators? Do you need periodically sharp (but not long) increase in computing resources available to you? Do you need interactivity or is batch processing fine for you? Do you have many small (short) jobs, many long jobs or few but large (long) jobs? And how many of your jobs is single/dual CPU and how many are really parallel (needing tens of CPUs)? A comment to the next list -- we are supporting different applications with different requirements so one table for all of the applications does not make sense. Therefore. I am answering for Computational Chemistry only

• 1. How do you know about Grid ?! 2. What is the application you want to deploy on Grid system which is best for your job ? 3. are you ready for Grid ? Do you intend to use Grid for your organization, how about your plan?

• Will having more computing power be helpful for them? If more computing power is helpful for their research, are they expecting any factors that may be bottle necks.

• Don't know, although I think a lot of education will be required if we are going to ask social scientists anything about EGEE. i know virtually nothing, and i am way ahead of most!

• What do they need ? What do they want ? What can they imagine ? What are the obstacles they experience ?

• What role do you think advanced ICTs can play in your specific research area? Have you heard about the concept of e-Social Science? Have you heard about the International Conference on e-Social Science? What kinds of datasets do you use in your research and where do these datasets come from? Do you make use of electronic datasets held at data archives? What computational methods do you use, e.g., data linkage, simulation, statistical methods? What barriers do you face in the uptake of e-Research methods?

• Will Web-based job submission portals simplify your research?

[edit] How important do you think the following kinds of issues will be in the application areas you are interested in?

Here are the categories used in the kick-off survey and the responses we gave ourselves:

critical	very important	important	not relevant	Response Count
technical (workload management)	36.4% (4)	9.1% (1)	54.5% (6)	0.0% (0)	11
technical (data management)	45.5% (5)	36.4% (4)	18.2% (2)	0.0% (0)	11
technical (interoperability)	9.1% (1)	54.5% (6)	27.3% (3)	9.1% (1)	11
technical (other)	0.0% (0)	50.0% (5)	40.0% (4)	10.0% (1)	10
user friendliness	36.4% (4)	45.5% (5)	9.1% (1)	9.1% (1)	11
organisational	18.2% (2)	36.4% (4)	45.5% (5)	0.0% (0)	11
ethical/legal (e.g., confidentiality, privacy)	36.4% (4)	18.2% (2)	27.3% (3)	18.2% (2)	11
licensing	9.1% (1)	45.5% (5)	18.2% (2)	27.3% (3)	11

[edit] What would you say are the main obstacles for the wider uptake of e-Science in your country and in the application areas you are working with? How might they be addressed?

Here are the answers we gave in the kick-off survey:

• 1. motivation/incentive for user community 2. the understanding of user community themselves that they are the driver to lead e-Science applications and collaborations

• Small very fragmented research teams; most of them located in the capital and not overly interested in collaboration within the country. Very limited knowledge of grids (and unwillingness to learn more). Using a simplified equation: Grid = EGEE = HEP community => grids are of no benefit for me as I am not associated with the HEP. Lack of interactive work with the (EGEE) Grid, too complex (and slow/inflexible) management and scheduling.

• cf the SHARE Roadmap: http://www.eu-share.org

• a) the grid doesn't exist b) no-one knows how to use it. How to address this should be obvious i.e. create 'it' and have better tools for access.

• Amount of people resources. Greater collaboration with international groups like this and dissemination of the information the the ANU and wider Australian research communities.

• While there are some early adopters in the social sciences, most researchers in this area are not familiar with the concept of e-Social Science, do not see it as important for their work or do not know how to operationalise it. There is also a perception that e-Science is just for 'big science' and that it has no relevance to the social science.

• Acceptance, Expertise, Resource and data access, organization policy.

• Information regarding the project is not yet widespread. Lack of computing resources from potential partners. Lack of trust among universities.

[edit] In your opinion, what are the main sustainability challenges and how could they be addressed?

Here are the answers we gave in the kick-off survey:

• Only reliable services/infrastructure and ease-of-use interfaces would attract users to stay with. The cost to integrate users applications to gLite is also an issue.

• The funding :-) Acceptance (by broad scientific communities) that Grid provides a kind of the same commodity added value as they see in their use of networks. Understanding at the funding agencies level that building a coherent infrastructure is much less expensive than dissolving the money among individual research teams and letting them to buy and manage the equipment. The Grid community accepting that the "Big science" is not the only potential benefiter from Grid technology and changing the way how small really distributed teams are supported (both by the available technical solutions, man power required to join and the actual user support level and quality)

• Involvment of ever and ever large scientific sectors, to be reached making use of the EGEE vituous cycle Involvment of enterprises, disseminating top-down and with virtuous cycle Convince government people, using ad-hoc dissemination, mainly events to convince them about utility of grid and relative uptakes.

• Deploying for end user and let them interested then invest themselve for Grid Sciencetific organizations with real need for current work.

• Challenges - funding, skills, uptake. Solutions - funding, training, community-building.

• I believe that this will be a set of evolving challenges. Initially in terms of getting the resources to the users in a form they can use. Utilising the massive consumer compute resources available on the web (examining the spectrum from distributed capabilities like Seti@home to the commercial systems like google and amazon). We should also evaluate the environmental consequences against the social and science benefits. What direct benefits of global social or scientific capability can this group demonstrate?

• e-Social Science is still largely dependent on the core funding stream made available by the UK's ESRC. ESRC have made it part of their core strategy but if we do not manage to achieve wider uptake across a number of disciplines within the social sciences, this position may be at risk. We need to establish a set of core applications with wider applicability. The ESRC e-Infrastructure project is seeking to achieve this and to provide portal-based access to national resources (datasets such as census data and computational and storage resources provided through the NGS)

[edit] Collection of questions from 21st May meeting

• What are the application trend of Grid technology in your country?
• Impact of grid technology in macroscopic level of your nation?
• Who are target group and beneficiary?
• What are the helps that needed from grid developer for those who want to apply grid technology ?
• Should there be any macro-policy to support grid rapid adoption?
• What type of project or activity will help broadening grid technology realization in your nation?
• what gaps exist in training provision
• what coordination structures exist in our country? Do you have a national grid service, an office of cyberinfrastructure, a local chapter of OGF or similar institutions?
• Is there a national or regional e-science conference that you would consider to be essential?
• top three potential application domains (and the leading organizations of each) to adopt grid technology
• current application domains that already making use of grid and doing international/cross organizational collaboration
• questions about details of any national grid service - middleware used, organisational structures, funding models etc.
• What prevent application developer to use grid?
• Is there a national roadmap?
• Is training provision coordinated / centralised?
• How is provision of infrastructure for research funded?
• If you have a national programme, does it connect up with different research disciplines?
• outreach and education vs. training
• Does grid scare you in anyway? If so, why?
• other than academic applications, how about the plan to industrial applications, any concern or issues there ?
• two types of users...application users.. and grid tech developers - need to see both application community and grid tech community developed...hand in hand
• do you have a coordinated community enagement programme to overcome problems with communication between users and technology or application developers?
• do you have experience with the EGEE infrastructure in particular?
• does anyone systematically collect information about uptake, barriers and enablers?
• how is the impact assessed?
• Are there supercomputers and their users groups in your country?
• as human resource is also a precious resources in grid, shall we collect the expert list of each country as well, categorized by grid technology, and domain experts (subgrouped by each domain)

[edit] WP2.1 Survey of Requirements

[edit] First Draft

Dear Colleague,

the EUAsiaGrid project aims to promote the use of advanced information technologies, in particular the EGEE Grid, in the AsiaPacific region. We would like to ask for your kind help in establishing requirements for the implementation of resources in the partner countries and for the necessary coordination policy and community engagement process.

If you are unfamiliar with grid computing technologies or any of the terms used in this questionnaire, please have a look at <insert link here> the grid introduction page <end link> to familiarise yourself with these concepts.

Best wishes,

Alex Voss (for the EUAsiaGrid consortium)

[edit] Personal Information

The information in this section is optional and will be treated as confidential. We will only contact you if you indicate explicitly that you are happy for us to do so.

• Name: ________________________________________
• Institution/Research Group: ________________________________________
• Address: ________________________________________________________________________________
• Telephone Number: ________________________________________
• Email Address: ________________________________________
• Please subscribe me to the EUAsiaGrid mailing list [ ]
• I would like to know more, please contact me by email [ ]

[edit] General Information

1. Research Domain: ________________________________________ should be a tick-list? okay. if it's too broad, maybe we can leave a blank to be more specific. -Rey
2. Can you please describe your specific research area? ___________________________....
3. Job title: Professor [ ] / Associate Professor, Reader, Lecturer [ ] / Assistant Professor, Research Fellow [ ] / PhD Student [ ] / Postgraduate Student [ ] / Other [ ]

[edit] Use of Grid Technologies

1. Are you currently involved in research projects using grid? Can you please provide a short description of these and pointers to websites where appropriate.
2. Do you think your research benefits or could benefit from the use of grids?
3. What kinds of resources are needed for your research? computation [ ] / data management [ ] / meta-data management [ ] / Other _________________________

[edit] Computing Resources

1. Do you use computing resources such as clusters, supercomputers or other compute resources such as Condor pools? _____________________....
2. Are such resources provided and supported in your institution? _______________...
3. Do you have specific requirements such as high memory capacity or support for long-running (weeks or months) compute jobs?

[edit] Network Connectivity

1. What kind of Internet connection does your institution have?
2. Do your applications have particular networking requirements (e.g., high bandwidth, low latency)?

[edit] Data

1. What is the nature of the data you process? Is it largely bulk data or fine-grained data in databases?
2. How much data do you process?
3. Is storage capacity an issue? Are access speeds crucial?
4. Do you need to process data in real-time?

[edit] Human Resources

1. How many colleagues do you have with experience in grid computing?
2. Do you have software developers specializing in parallel algorithms?

[edit] Software

1. How important is a robust yet user-friendly application to you?
2. Is there any algorithm that you would like to run on the grid?
3. Does your work require the use of commercial or other special license software?

[edit] Finally

• Do you have any other comments? __________________

[edit] References

• Adapted from http://engage.ac.uk/documents/ENGAGE%20Interview%20Template.pdf/at_download/file:

1. Could you briefly describe the type of problem that you are studying and what your scientific objectives are?
2. Do you write your own software packages or tailor existing software? Are these open source?
3. Who are the typical users of the software/services you provide?
4. What is the approximate number of users for your project?
5. Are you collaborating with other institutions? a. Do you share resources with other institutions? b. Are there any constraints involved when trying to share resources with these institutions?
6. Are you using computers as part of your research? a. Do you use a single PC, servers, clusters, other facilities? b. How often do you use these facilities? c. How long does it take to perform runs on these facilities? d. Which of the following is the most important factor when performing these runs: CPU/memory intensive computation, Gathering/processing large quantities of data, High throughput (data/jobs), Availability of remote resources, Security, Job scheduling? e. Which of the above are the limiting factors of your current set-up? f. Are you currently using Grid technologies for your research? i. What benefits have you found from using Grid technologies? ii. What are the main difficulties in using Grid technologies?
7. What do you think your future computing requirements will be: More CPU power, More storage capacity, Data Management tools, Data analysis/visualization tools, Have a virtual collaboration environment to work with remote colleagues, Single point of access to resources?
8. If you had an unlimited budget what would be your ideal system setup?
9. What are the main issues, in your view, restricting your uptake of e-Infrastructure?
10. Are there any issues, comments or remarks that you would like to make that haven't been covered by this interview?
11. Are you happy for us to publish the information you've provided (not including your contact details)?

• Adapted from http://web.fhnw.ch/plattformen/avross/papers-and-prensentations/AVROSS%20Paper%20Michigan%20v2-2.pdf:

1. the respondent's background, organization, and experience with e-Infrastructure;
2. the respondent's current or most recent e-Infrastructure project;
3. background about funding and results;
4. the respondent's views of catalysts (Seed funding from an outside agency, Seed funding from home institutions, Organizational incentives, Collaboration, Observation of successful projects, Computational requirements of your research, Contribution to interesting research, Support for teaching, Emerging standardization of available tools) and barriers (Lack of initial funding, Costs associated with e-Infrastructure development, Lack of information about usefulness, Lack of staff available to help with development, Insufficient applicability of existing technology to social science research problems, Problems with intellectual property rights, Lack of trust in sustainability, Problems with protecting confidentiality of data, Locked into other technologies) to the development and implementation of e-Infrastructure projects
5. further e-Infrastructure projects and people who might be able to provide interesting information

• Adapted from http://www.nesc.ac.uk/technical_papers/UKeS-2007-01.pdf:

1. sources of funding of the responders. The funding sources mentioned as other were either university or EU-related funding sources. The project goals varied from infrastructure providers (hardware, data sources, digital curation, etc.), science and research groups (gene expression, engineering, networking, bio-informatics, high energy physics, etc.), to middleware, portal or visualization development groups.
2. respondent's area of research, where each was asked to select all that might apply.
3. where the respondents felt their role was in a user taxonomy
4. importance to the respondent for different facets of system software tools and services. Jobs are presented as being more important in this set of responses, but this doesn't contradict the interviews in that this question isn't where are you having problems, but how important is it to your project. As seen in the interviews as well, there is still a focus on the basic functionality (job submission before monitoring, file access before provenance or replication). And as expected, very few aspects of this problem were considered of no importance at all.
5. "What are the main issues, in your view, restricting your uptake of the e-Infrastructure?"