October 2005 - Ag-tech - mailman3.cels.anl.gov

Access Grid Townhall Reminder - October
by Thomas D. Uram 03 Oct '05

03 Oct '05

The October Townhall is scheduled for Tuesday, October 4, at 10:00AM CDT (15:00 UTC). It will be held in the Argonne National Laboratory venue on the institutional venue server (Go to https://ivs.mcs.anl.gov:9000/ Venues/default, then Government Institutions Lobby, then Argonne National Lab). This month we will have general updates from Argonne and the community. Bring your discussion topics: townhall meetings are intended as a forum for open discussion of items of interest to the community, be they status, praise, or gripes. There will be a second session held for those in the Asia-Pacific region, at 6:00PM CDT (23:00 UTC). Of course, anyone is welcome to attend this session, if it's more agreeable with their schedules. ANL will be running bridges in the Argonne venue for this meeting.

1 0

Townhall Tomorrow?
by Todd Zimmerman 03 Oct '05

03 Oct '05

Hey All, I don't remember hearing any notice... is there a townhall tomorrow?? Todd

1 0

Node Listing pages
by Thomas D. Uram 03 Oct '05

03 Oct '05

The node listing on the AG portal is quickly becoming the standard reference for AG nodes worldwide. If you wish to register your node, you should do so on the AG portal, at http://portal.accessgrid.org/nodes . Questions about doing so should be directed to the portal forums. The existing node listing pages at accessgrid.org (http://www.accessgrid.org/community/nodes/nodes.html) are being deprecated. The accessgrid.org node listing pages will no longer accept new listings or revisions, and will eventually be removed from the site. Questions about the node listings being retired should be directed to ag-info(a)mcs.anl.gov. Thanks, Tom Uram

1 0

Multicasting
by Jeff Miller 03 Oct '05

03 Oct '05

Hello All, Putting together a project "snapshot" for our new grid room . The constraints I have listed are; Multicasting, Bandwidth and Encryption (or lack thereof) Are the constraints for Multicasting a matter of "routers" outside of my institution or is it something else? Will I need a VPN JUST for a room node? We have a 100mb pipe out to the commodity internet. But I'm assuming that AG uses mostly I2. Your thoughts! Jeff Jeffrey L. Miller Manager, Academic Media Resources Educational Technologies & Services Office of Information Technologies University of Notre Dame 153 DeBartolo Hall Notre Dame, IN 46556 574-631-6850 574-261-1895 Cell Jeffrey.Miller.376(a)nd.edu "It's a privilege to work at Notre Dame..."

1 0

Access Grid based seminar
by Lee Margetts 03 Oct '05

03 Oct '05

PLEASE CONTACT jon.gibson(a)manchester.ac.uk for further details ESNW/MRCCS Seminar Announcement "A new algorithm for finding minimal sample uniques for use in statistical disclosure assessment" Friday, 7th October 2005, 2-3p.m. BST Venue: Room 1.10 (ESNW Access Grid) Kilburn Building http://www.esnw.ac.uk/n-events.shtml --- Dr. Anna Manning The Centre for Novel Computing Department of Computer Science University of Manchester Improvements and innovations in computer processing power, disk storage and networks have led to dramatic increases in the ability to accumulate and analyze personal data. However, if personal data is made available, even in an anonymized form, there is a risk of individuals being identified using statistical disclosure through the matching of known information with the anonymized data, resulting in material specific to those individuals being revealed. This work focuses on the identification of individual records with a high risk of disclosure, a process otherwise known as Statistical Disclosure Assessment. The records belonging to certain individuals have a significant chance of being identified as their contents, or attributes, are unique and therefore have the potential to be matched directly with details (including names and addresses) from another dataset. An illustration of a `risky' record of this type is a sixteen-year-old widow in a population survey. The ability to comprehensively locate and grade such records leads to more efficient Statistical Disclosure Control (SDC) of released data. A sequential algorithm, SUDA (Special Unique Detection Algorithm), has been designed and implemented specifically for this problem and is currently used by the Office for National Statistics (ONS) in London. Only unique attribute sets without any unique subsets --- Minimal Sample Uniques (MSUs) --- are considered in order to avoid the use of redundant information and to keep the classification process as focused as possible. Although SUDA has greatly increased the depth of risk assessment possible, the demanding levels of execution time required to find all MSUs mean that it is restricted to small datasets, particularly in terms of the number of columns that they possess. This talk presents SUDA2, a recursive algorithm for finding MSUs. SUDA2 uses a novel method for representing the search space for MSUs and new observations about the properties of MSUs to prune and traverse this space. Experimental comparisons with SUDA demonstrate that SUDA2 is not only several orders of magnitude faster but is also capable of identifying the boundaries of the search space, enabling datasets of larger numbers of columns than before to be addressed.

1 0