December 2006 - ogsa-hpcp-wg

Good Bye
by Peter G. Lane 15 Dec '06

15 Dec '06

Hi All, Some recent circumstances have forced me to quit my job and become a stay-at-home dad for a year. I will probably monitor things from time to time, but at this point I don't have any plans to continue my involvement in the BES and HPCP working groups. Participating in these working groups has been a lot of fun as well as a useful experience. I learned a lot. It was nice meeting those of you who I did meet at OGF, and I regret missing many of the other HPCP demo participants at SC. Good luck to everyone and the future of open standards. Peter P.S. Please contact Stu Martin (smartin(a)mcs.anl.gov) if you need to ask any questions about the Globus Project's BES implementation/plans.

1 0

reminder: HPC Profile call today (in about 20 minutes)
by Marty Humphrey 15 Dec '06

15 Dec '06

Folks, Just a reminder: we will have an HPC Profile call today, starting in about 30 minutes. Our focus will be on the "security section" of the HPC Basic Profile, recently uploaded to Grid Forge. Toll Numbers: USA 1-203-480-8000 TollFree Numbers: USA 866-500-6738 Participant Passcode: 231867

1 0

draft "security considerations" for HPC Basic Profile
by Marty Humphrey 13 Dec '06

13 Dec '06

Folks, Thanks again for all of your hard work for the Supercomputing demo last month. I believe that we now have a new HPC Basic Profile that reflects our collective experiences, at least in regard to security. More precisely, I have uploaded a new draft of the "HPC Basic Profile" to forge.ogf.org. The *only* modification is the inclusion of a "Security Considerations" section based on our collective experiences of the SC2006 interopfest. I encourage you to read it at: http://forge.ggf.org/sf/docman/do/downloadDocument/projects.ogsa-hpcp-wg/doc man.root.drafts.hpc_basic_profile/doc13736/5 Note: I remind everyone that the "HPC Basic Profile" corresponds to the "Base Case" of the use-case/requirements document (http://www.ggf.org/Public_Comment_Docs/Documents/Aug-2006/draft-ggf-ogsa-hp cp-use-cases-02.pdf); correspondingly, the "security considerations" of this "HPC Basic Profile" doc essentially only covers this "Base Case" as well. The "Common Cases" are NOT specifically addressed. For those of you who cannot immediately read this, the essence of the HPC Basic Profile "Security Considerations" in this draft is as follows: R0501: An INSTANCE MUST support TLS 1.0, SHOULD support SSL 3.0, and SHOULD support TLS 1.1. R0502: An INSTANCE MUST support the FIPS-140 compliant ciphersuites. R0503: An INSTANCE MUST support TLS_RSA_WITH_AES_128_CBC_SHA. R0504: An INSTANCE MUST support service authentication using X.509 certificates using RSA cryptographic keys and the SHA-1 digest algorithm. R0505: An INSTANCE MUST support either client authentication using username/password credentials or X.509 certificates using RSA cryptographic keys and the SHA-1 digest algorithm. R0506: An INSTANCE must use TLS/SSL encryption key agreement based on the RSA algorithm. Diffie-Helman key agreement shall not be used. R0507: Client authentication based on username/password must use a password digest and conform to the Web Services SecurityUsername Token Profile 1.1. We encourage comments/questions on these 7 requirements/mandates, both on this email list and on the call this Friday (11am Eastern). However, *BEFORE* raising a question or concern on this list above, *PLEASE* read the document first, as the document has a fairly detailed (for a technical recommendation) explanation/justification. Your questions might possibly be answered in this doc. I also note that this is a *DRAFT*, so there's plenty of time/room for discussion on this! See you on the call on Friday. This security section will be the main topic. -- Marty

1 0

HPC Profile call this week on FRIDAY (focus: security section of "HPC Basic Profile")
by Marty Humphrey 12 Dec '06

12 Dec '06

Folks, In an attempt to ensure the maximum number of participants, let's go back to Friday this week at 11am Eastern. Let's see how many people we get. Tomorrow, I will send the draft of the "security section" of the "HPC Basic Profile". We have written a rough draft based on our conversations and discussions centered around the HPC Profile interop demos for Supercomputing 2006. By getting it out tomorrow, we'll have time to read it and have email discussions prior to the call on Friday. Thanks, Marty

1 0

HPC Profile Conf Call - 12/6
by Richard Ciapala 06 Dec '06

06 Dec '06

Attendees · Marty · Glenn · Chris · Alessandro · Rich Notes · Faults o We agreed with Chris's refactoring of faults per his earlier email. Need to validate in tooling o We agreed with Chris's list of faults and decided to add a generic invalid input fault. Still need to decide if we have more specific faults for common invalid input such as activity ID or just use generic invalid input fault. · FactoryAttributesDocument o We agreed with Chris's refactoring of factory attributes document. The main changes included: § Minor naming changes § Embedding BasicResourceAttributesDocument in FactoryResourceAttributesDocument (as opposed to having the latter derive from the former) § Added ResourceName § Subelements of ContainedResource must be BasicResourceAttributesDocument or FactoryResourceAttributesDocument § Considering using lexical representation for values of NamingProfile and LocalResourceManagerType · Range_Type o We agreed to mandate support for min and max processor count via LowerBoundRange and UpperBoundRange · Filter o Reviewed reasons for needing filter o Discussed adding Filter element to GetFactoryAttributesDocumentType with Any subelements to allow for filter extensions o Discussed working through a filter extension which will likely be a profile extension · Security o Will discuss next week Thanks

1 0

Changes to factory attributes document
by Christopher Smith 06 Dec '06

06 Dec '06

Hi all, I just wanted to send out a couple of sample XML documents for discussion in tomorrow's call. The first attachment is an example of a factory attributes document, as is reflected with the current schema (including all the issues we've seen having to do with multiple xsd:any, etc). I've also added some sample "extensions" as placeholders. The second attachment reflects some schema changes that I'd like to propose. One is just a re-factoring to avoid the use of XML Schema extension as discussed in last week's call. I've also added a proposed new element to the basic resource (resource name) which I'd like to see added. -- Chris

1 0