Subject: Ways to organize the proposal From: "John Rundle" Date: Tue, 1 Mar 2005 12:54:57 -0800 To: , "Marlon Pierce" CC: "Rundle, John" , "Louise Kellogg" , "Ken Joy" , , "Don Turcotte" , "Dennis McLeod" , "Dave Yuen" , "Bertram Ludaescher" , "Michael Gertz" Please refer to message below from Michael Gertz on ways to organize the iSERVO proposal Here are some thoughts regarding the components and structure of the proposal. 1) The project description should have an organizational chart/diagram at the beginning that shows: Participating US institutions: For each institution, the diagram should show major research efforts/projects that are already in place at that institution and that will be integrated into the proposed research/eduction. For each institution, we should indicate major education efforts, such as (televised/online) courses, organization of workshop, summer schools etc. For each organization, we should indicate with what international partners an institution already is collaborating. Clearly, not every institution is collaborating with every other institution and/or international partner. Links between institutions result from having more than one institution collaborating on existing projects. The idea here is that we all get a better picture in terms of who is already doing what, courses/educational efforts that are in place etc. Such a chart then is used to guide the writing/organization of the research plan and education plan. Such a chart might also turn out to be very helpful for the reviewer. Talking about all the relationships among institutions, research projects, educational efforts only in the text might not give a clear overall picture. 2) Research plans and objectives. Because this proposal/program is more about IT deployment than IT development, a natural question is "What are the research objectives?" Given that we want to build on projects and infrastructure that are already in place, major objectives should be * "scale-up" existing infrastructures (e.g., tools, services etc) * "Virtualization" virtualize existing tools and techniques to facilitate large-scale collaborative efforts. Let me use GEON or GeoStreams as an example: What infrastructure is necessary to design, implement, and deploy a large-scale scientific workflow in a collaborative fashion, involving some US and international partners? What tools and techniques are necessary to integrate and exchange vast amounts of heterogeneous and distributed remotely-sensed (or sensor) data? How can existing tools be used to allow scientists to design and utilize simulation tools and techniques in a collaborative fashion? Scaling up existing infrastructures and making them more virtual can be considered a research objective, I think. 3) Education plans and Objectives We need to provide extensive training to US and international students and faculty so that they can effectively utilize and extend the cyber-infrastructure (CI) we build in 2). Educational efforts include: * annual workshops (all institutions participate) * multi-institutional / project specific workshops, focusing on specific parts of the CI to be developed in 2). Here is another place where the chart can help. * summer schools: US students and faculty will visit a partner site, and vice versa. Summer schools and workshops should also be used as "all hands-on" meetings where groups of students/faculty focus on extending and evaluating CI components. * televised courses: that's why we have to have a list of courses in place, the names of the institutions that will offer such courses etc. Do our international partners have facilities to participate in televised courses? How do we "educate" people at partner institutions, how to they "educate" people at US institutions? * online-courses: such courses would focus on how to use CI components that are built in the proposed research. Setting up such courses is a major effort (time consuming). ==> budgeting for personnel to install, maintain, and monitor course sites is critical! ___________________________________________________ John Rundle Director, Center for Computational Science and Engineering Professor of Physics and Engineering University of California One Shields Ave Davis, CA 95616 Tel: (530) 752-6416 Fax: (530) 754-4885 http://cse.ucdavis.edu/~rundle/ ___________________________________________________ ----- Original Message ----- From: "Geoffrey Fox" To: "Marlon Pierce" Cc: "John Rundle" ; "Zhaojun Bai" ; "Steve Ward" ; "Rundle, John" ; "Naoki Saito" ; "Maggi Glasscoe" ; "Louise Kellogg" ; "Lisa Grant" ; "Ken Joy" ; ; "Don Turcotte" ; "Dennis McLeod" ; "David M. Manaker" ; "Dave Yuen" ; ; ; "Bertram Ludaescher" ; "Bernd Hamann" ; "Linda Potoski" ; "Roland Freund" ; "John Rundle" ; "Jim Crutchfield" ; "Jean-Pierre Delplanque" ; "Raissa D'Souza" Sent: Thursday, February 10, 2005 6:28 PM Subject: Some Thoughts on "New" Component\ > One way of broadening iSERVO and hence addressing NSF Requirement > of a new collaboration could be a "Realtime Grid track" with major Grids > signed up > to support real time science by linking together with a Grid Operations > Center GOC at IU > (IU currently runs Internet2 and Transpac networks as well as Grid > Operations > Center for "Open Science Grid" > > With GOC we could credibly support both new iSERVO installations and > notification > of Tsunamis etc as well as international access grid and sensor access. > > > Grid Components I see are > Scientists@iSERVO > SensorNet@iSERVO > CuratedData@iSERVO > GIS(Visualization)@iSERVO > HPC@iSERVO > DesktopGrid@iSERVO (BOINC as in climateprediction.net) > OutreachandEducation@iSERVO > Alerts@iSERVO (using communitysciencegrid enclosed) > > LEAD noted by Marlon fits this but it isn't Solid Earth > > We could have an IT part that is general to "Earth/Environmental Science" > but a Solid Earth as Science specialization? > > Any comments? > > -- > : > : Geoffrey Fox gcf@indiana.edu FAX 8128567972 http://www.infomall.org > : Phones Cell 812-219-4643 Home 8123239196 Lab 8128567977 > > -------------------------------------------------------------------------------- > We wish to define a model for science linking traditional > research to education and which uses a model which > appears to be succesful from e-commerce. This can help > broaden the participation in and appreciation for science. > > e-commerce is built around > a) End products like digital cameras that you might want > to buy > b) Places such as Amazon and eopinions that compare > products. Other places that list and otherwise > compile information about products. > c) Places like Froogle and pricescan that collect all > the different places you can buy and compare them. > Prices and reliability indicators can be given. > d) Places where you buy products with associated > catalogs, purchase and shipping units. > > Now to e-CommunityScience: > Imagine a world full of sensors producing science data. > Assume all sensors are enabled as Web (Grid) Services > These sensors could be traditional pieces of hardware > (seismometers, GPS Satellites, Weather, Video cameras, > accelerators ...) or the papers, Blogs and Wikis that > can be thought of as the "Web Service Interface to > Brain waves) > > Now the products in science are a mix of "not so > controversial real things" (the observed data) > and ideas/deductions. > We assume all products are generated by some sort of workflow > linking data to simulations to more data to deductions etc. > Just as in e-commerce, a given set of "facts" can have > many different organizations/people generating deductions. > We assume as in e-commerce, that we have other Web services > that collect related "scientific deductions" and presented > in a way you can sort in various ways. "Vendors" of "deductions" > could be rated in ways you choose > By number and quality of PhD's in team > Religuous correctness > Type of organization etc. > > End users are responsible for choosing scientific results > according to criteria they feel comfortable with. In particular > the current scientific methodology with roughly similar models > for evaluation of "vendors" can thrive. > The advantage of this approach is that it documents method of > selection of data and ideas and most importantly it allows > for a bigger net to catch ideas. Maybe most good results will > come from traditional sources but community science will > allow all to be potential participants. > > For example, consider the current tragedy in Asia and suppose > (as probably isn't the case now) that relevant seismic > sensors are set up as Web Services. Then naively many different > groups would be to apply different approaches and their > results can in real time be collated and made available > to those worried about Tsunamis. There need not be just one > system but multiple approaches; thus even if Tsunami occurs > when key people in one organization are not available, those > in another time zone could come up with the needed analysis. > > This suggests that Science Grids be built in a certain fashion > where at each stage of a possible science ideaflow or workflow > one sets up a framework that allows not only all people but all > resources to be harnessed i.e. we support both HPCC simulations > but also simulations on a P2P network in somebody's home. > > This approach seems to be an attractive way to bring research > both into K-12 education and to communities that today do not > have a rich scientific infrastructure and expertise. > One could help both education and research disadvantaged > organizations by linking them to the now transparent > traditional enterprise and by providing resources that > enabled their participation in it i.e. providing the needed > computing cycles for a local Tribal College to develop > custom agriculture models linked to the "traditional > enterprises environment, weather and climate data"