Begin forwarded message:
From: "Linesch, Mark" <mark.linesch@hp.com> Date: 2 January 2007 16:04:56 GMT To: "David Snelling" <David.Snelling@UK.Fujitsu.com>, "Joel Replogle" <replogle@ogf.org> Cc: <scrumb@ogf.org> Subject: FW: comments on OGF Technical Strategy document
David, Joel,
Comments on the TSC document from Greg Astfalk at HP. Mark
Mark Linesch: Open Grid Forum (OGF): Hewlett Packard 281-514-0322 (Tel): 281-414-7082 (Cell) mark.linesch@hp.com : linesch@ogf.org
-----Original Message----- From: Astfalk, Greg Sent: Saturday, December 30, 2006 7:24 AM To: Linesch, Mark; Walker, Martin Antony (HPTC); Vickers, Paul Subject: comments on OGF Technical Strategy document
FWIW, in no particular order, and ranging from mega- to minutia...
- Section 1.2 is missing a section numberin the third sentence
- Section 1.3: "distributed computing at Internet scale" is a poor summation. It will, for many people, eliminate the more practical and valuable intra-organization grids.
- Section 1.3, second paragraph implies that your focus is only on collections of supercomputers. I certainly hope that is <not> the case. Otherwise why did you merge with EGA, and why am I funding OGF?
- Section 2: the stated goal is laudable, very. However, I have angst over the use of "defined" (see more comments below). The clock is ticking and 3 years is a genuinely short time-frame to get through the process described later in the document. For me this imples that you need to speed-up since pushing the date out is the wrong thing to do.
- In some places in the document you use the construct "...text - text..." this should always be "...text<em-dashh>text...". That is, use the em-dash character (Word has it) with no spaces on either side. There are some places where you did this.
- The bullet list in section 2 does not have the requisite emphasis on "product". OGF needs to be known for more than jsut paper specs. How do reference implementations evolve to "product"?
- Figure 1: I assume that the union of use cases will provide the gaps? This should be almost known already, especially including the commercial space. This gors back to my point above that 2010 is an aggressive goal.
- OGF has a lot of WGs, some more important and relevant than others. Is there a possiblity of re-factoring the human capital toward meeting the core set of WGs needed to meet your 2010 goal? I am aware that this is volunteer stuff.
- Section 4.3: Would this be more accurate titled if it were "Data Movement"
- Section 4.3: Change "...from this from this..." to "...from this..."
- Section 4.6: You need to add auditing to the required list, i.e., all three components of the so-called AAA (aka triple-A) are necessary.
- Section 4.6.1" Tyhw work of Wenbo Mao in HPL-Beijing is, IMO, especially useful here. It does, however, require the presence of TPMs.
- Section 5 in the sub-bullet list: You list "Product" and that it is supported. By who? This goes back to my point on the word "defined" in your goal statement. Organizations will not be able or ready to build operational grids without functioning code. It is axiomatic to me that if only paper specs exist then your goal will not be realizable. It is also nearly so that you can not develop "product" by then. To have "support" by 2010 is, IMO, impossible. Is this a solvable conundrum?
- Section 5: Change "Table 1 is not a complete list of OGF activity nor..." to "Table 1 is neither a complete list of OGF activity nor...". You know, the ususal neither/nor, either/or thing...
- Section 6: A really, really good mandate. Change "All OGF document must..." to "All OGF documents must...".
-- Take care: Dr. David Snelling < David . Snelling . UK . Fujitsu . com > Fujitsu Laboratories of Europe Hayes Park Central Hayes End Road Hayes, Middlesex UB4 8FE +44-208-606-4649 (Office) +44-208-606-4539 (Fax) +44-7768-807526 (Mobile)