March 2019 - Ssio-2018 - mailman3.cels.anl.gov

Re: [Ssio-2018] ASCAC talk: suggestions for one slide
by Glenn Lockwood 22 Mar '19

22 Mar '19

I agree. I'm not sure what else could be added without adding another level of detail/complexity. Glenn On Fri, Mar 22, 2019 at 1:43 PM Mohror, Kathryn via ssio-2018 < ssio-2018(a)lists.mcs.anl.gov> wrote: > Hi Rob, > > > > I think you covered the major points there. Anything I can think of now > can be fit into those bullets. > > > > Kathryn > > > > *From:* ssio-2018 <ssio-2018-bounces(a)lists.mcs.anl.gov> *On Behalf Of *Ross, > Robert B. via ssio-2018 > *Sent:* Friday, March 22, 2019 12:30 PM > *To:* Settlemyer, Bradley Wade via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov> > *Subject:* [Ssio-2018] ASCAC talk: suggestions for one slide > > > > Hi all, > > > > I've been iterating on the slides today. I'm pretty happy with the set as > a whole, which have been uploaded to Google here along with a separate file > of narrative: > > > https://drive.google.com/drive/folders/1vjbzVGmBkeOwkSBb8GVvCxhUwIrIiQXY?us… > > > > Of course I'm happy to hear suggestions/corrections on any of the slides, > but I'm particularly interested in a suggestion on how to improve the HW/SW > / Deep Hierarchy mashup (important bits attached so you don't need to go > searching). I feel like there ought to be one more point I want to make on > this slide, but I'm a little at a loss. Thoughts? > > > > Thanks again for taking the time to help with the language of the research > opportunities yesterday; I think we've boiled it down especially well at > this point. > > > > Best, > > > > Rob > > --- > > *Improving efficiency and integrity of data movement and storage through > architecture of systems and services* > > - Lots of change has occurred outside of HPC in the area of different > storage service architectures. > > - One trend outside our community is that of stitching together a set of > different simple services to provide a more complex capability. > > - This is seen as a promising approach for HPC, but the concept needs to > be adapted to the networks, storage devices, programming models, and > science use cases in our domain. > > - With the continued diversification and specialization of storage > devices, it's increasingly important to develop more capable methods of > placing data to make best use of the collection, both from a user- and > facility- centric perspective. > > > _______________________________________________ > ssio-2018 mailing list > ssio-2018(a)lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018 >

1 0

Re: [Ssio-2018] ASCAC talk: suggestions for one slide
by Mohror, Kathryn 22 Mar '19

22 Mar '19

Hi Rob, I think you covered the major points there. Anything I can think of now can be fit into those bullets. Kathryn From: ssio-2018 <ssio-2018-bounces(a)lists.mcs.anl.gov> On Behalf Of Ross, Robert B. via ssio-2018 Sent: Friday, March 22, 2019 12:30 PM To: Settlemyer, Bradley Wade via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov> Subject: [Ssio-2018] ASCAC talk: suggestions for one slide Hi all, I've been iterating on the slides today. I'm pretty happy with the set as a whole, which have been uploaded to Google here along with a separate file of narrative: https://drive.google.com/drive/folders/1vjbzVGmBkeOwkSBb8GVvCxhUwIrIiQXY?us… Of course I'm happy to hear suggestions/corrections on any of the slides, but I'm particularly interested in a suggestion on how to improve the HW/SW / Deep Hierarchy mashup (important bits attached so you don't need to go searching). I feel like there ought to be one more point I want to make on this slide, but I'm a little at a loss. Thoughts? Thanks again for taking the time to help with the language of the research opportunities yesterday; I think we've boiled it down especially well at this point. Best, Rob --- Improving efficiency and integrity of data movement and storage through architecture of systems and services - Lots of change has occurred outside of HPC in the area of different storage service architectures. - One trend outside our community is that of stitching together a set of different simple services to provide a more complex capability. - This is seen as a promising approach for HPC, but the concept needs to be adapted to the networks, storage devices, programming models, and science use cases in our domain. - With the continued diversification and specialization of storage devices, it's increasingly important to develop more capable methods of placing data to make best use of the collection, both from a user- and facility- centric perspective.

1 0

ASCAC talk: suggestions for one slide
by Ross, Robert B. 22 Mar '19

22 Mar '19

Hi all, I've been iterating on the slides today. I'm pretty happy with the set as a whole, which have been uploaded to Google here along with a separate file of narrative: https://drive.google.com/drive/folders/1vjbzVGmBkeOwkSBb8GVvCxhUwIrIiQXY?us… Of course I'm happy to hear suggestions/corrections on any of the slides, but I'm particularly interested in a suggestion on how to improve the HW/SW / Deep Hierarchy mashup (important bits attached so you don't need to go searching). I feel like there ought to be one more point I want to make on this slide, but I'm a little at a loss. Thoughts? Thanks again for taking the time to help with the language of the research opportunities yesterday; I think we've boiled it down especially well at this point. Best, Rob --- Improving efficiency and integrity of data movement and storage through architecture of systems and services - Lots of change has occurred outside of HPC in the area of different storage service architectures. - One trend outside our community is that of stitching together a set of different simple services to provide a more complex capability. - This is seen as a promising approach for HPC, but the concept needs to be adapted to the networks, storage devices, programming models, and science use cases in our domain. - With the continued diversification and specialization of storage devices, it's increasingly important to develop more capable methods of placing data to make best use of the collection, both from a user- and facility- centric perspective.

1 0

Re: [Ssio-2018] Call today: status and our task
by Mohror, Kathryn 21 Mar '19

21 Mar '19

Hi Rob, I’m slow to get to my homework this week and was going to try to craft something up (the 200 word blurb you asked for) before the meeting. Should I target the 4 bullet list below or the original 5 bullets (so for me that would be “leveraging deep storage hierarchies” vs “Architecting to improve efficiency…”) Thanks, Kathryn From: ssio-2018 <ssio-2018-bounces(a)lists.mcs.anl.gov> On Behalf Of Ross, Robert B. via ssio-2018 Sent: Thursday, March 21, 2019 8:47 AM To: Settlemyer, Bradley Wade via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov> Subject: [Ssio-2018] Call today: status and our task Hi all, Alright, I've got an ASCAC talk next week, and the slides are coming together pretty well. Here are the current draft slides and associated narrative: https://drive.google.com/drive/folders/1vjbzVGmBkeOwkSBb8GVvCxhUwIrIiQXY?us… Here's the thing: we have a few different versions of our "research opportunities" (new term), and we need to downselect to one version, and we need to do that ASAP so that I can put the finishing touches on these slides (and also so we can quit discussing this bit :)). There are a bunch of places that these appear: executive summary, trapezoid slide, topics of slides in ASCAC (etc.) talk, and the brochure. ASCR has provided feedback that we should tune up the executive summary, and I'm attaching the latest to this email. Here's the version of the "opportunities" from that new section: • Reducing time to scientific discovery through in situ and streaming data analysis • Improving understandability and reproducibility through rich data formats, metadata, and provenance • Architecting to improve efficiency, durability, and better use of deep storage hierarchies • Understanding SSIO for resilience, performance, and control For reference, here is the previous version from the report: · Improving storage systems through better understanding · Reducing time to scientific discovery through streaming data analysis · Architecting for improved efficiency and durability of scientific data · Understanding science data and improving reproducibility through enhanced metadata · Enabling in situ and streaming workflows · Leveraging deep storage hierarchies Today I'd like us to focus on the language for these things. I think the new, 4-bullet version is good and the right level of detail for what we're trying to accomplish, but I would like to hear suggestions and criticism both on the grouping and on the language. We'll use this Google Document as a place to discuss and wordsmith: https://docs.google.com/document/d/1M8SLzBQlG1O5C4yJLxYijbpbuKzYJAUstL4ccEK… I look forward to chatting with you all later today! As a reminder, our call is at 1pm eastern, and here's the call-in information: https://bluejeans.com/177339282 Phone Dial-in +1.408.317.9254 (US (San Jose)) +1.866.226.4650 (US Toll Free) Meeting ID: 177 339 282 Rob

1 0

Call today: status and our task
by Ross, Robert B. 21 Mar '19

21 Mar '19

Hi all, Alright, I've got an ASCAC talk next week, and the slides are coming together pretty well. Here are the current draft slides and associated narrative: https://drive.google.com/drive/folders/1vjbzVGmBkeOwkSBb8GVvCxhUwIrIiQXY?us… Here's the thing: we have a few different versions of our "research opportunities" (new term), and we need to downselect to one version, and we need to do that ASAP so that I can put the finishing touches on these slides (and also so we can quit discussing this bit :)). There are a bunch of places that these appear: executive summary, trapezoid slide, topics of slides in ASCAC (etc.) talk, and the brochure. ASCR has provided feedback that we should tune up the executive summary, and I'm attaching the latest to this email. Here's the version of the "opportunities" from that new section: • Reducing time to scientific discovery through in situ and streaming data analysis • Improving understandability and reproducibility through rich data formats, metadata, and provenance • Architecting to improve efficiency, durability, and better use of deep storage hierarchies • Understanding SSIO for resilience, performance, and control For reference, here is the previous version from the report: · Improving storage systems through better understanding · Reducing time to scientific discovery through streaming data analysis · Architecting for improved efficiency and durability of scientific data · Understanding science data and improving reproducibility through enhanced metadata · Enabling in situ and streaming workflows · Leveraging deep storage hierarchies Today I'd like us to focus on the language for these things. I think the new, 4-bullet version is good and the right level of detail for what we're trying to accomplish, but I would like to hear suggestions and criticism both on the grouping and on the language. We'll use this Google Document as a place to discuss and wordsmith: https://docs.google.com/document/d/1M8SLzBQlG1O5C4yJLxYijbpbuKzYJAUstL4ccEK… I look forward to chatting with you all later today! As a reminder, our call is at 1pm eastern, and here's the call-in information: https://bluejeans.com/177339282 Phone Dial-in +1.408.317.9254 (US (San Jose)) +1.866.226.4650 (US Toll Free) Meeting ID: 177 339 282 Rob

1 0

Re: [Ssio-2018] ASCAC slide, caption
by Ross, Robert B. 19 Mar '19

19 Mar '19

Hi all, That was meant to go to Brad, but I'm happy to hear suggestions from others. I love what ANL has done to mailing lists... Sorry for the distraction. Rob > On Mar 19, 2019, at 2:03 PM, Ross, Robert B. via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov> wrote: > > Hey, > > I'm refining slides for ASCAC -- boiling them down, putting interesting sidebars on them, etc. This is a merge of your HW/SW and Kathryn's deep hierarchy slides. Hopefully you're not too insulted by the simplification. > > I'm hoping you could help me put together a story to go with the figure that you provided :)? And was it used in a publication anywhere? > > Thanks! > > Rob > --- > <PastedGraphic-3.pdf>_______________________________________________ > ssio-2018 mailing list > ssio-2018(a)lists.mcs.anl.gov > https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018

1 0

ASCAC slide, caption
by Ross, Robert B. 19 Mar '19

19 Mar '19

Hey, I'm refining slides for ASCAC -- boiling them down, putting interesting sidebars on them, etc. This is a merge of your HW/SW and Kathryn's deep hierarchy slides. Hopefully you're not too insulted by the simplification. I'm hoping you could help me put together a story to go with the figure that you provided :)? And was it used in a publication anywhere? Thanks! Rob ---

1 0

Fwd: [Mochi] ssio-brochure-draft - Somehow need to bring metadata and pr...
by Ross, Robert B. 17 Mar '19

17 Mar '19

I don't think we want to be creating new sections at this point, as adjusting these topics implicitly adjusts the PRDs. So it would be better to fit it into the existing organization? -- Rob Begin forwarded message: From: "Quincey Koziol (Google Docs)" <comments-noreply(a)docs.google.com<mailto:[email protected]>> Subject: ssio-brochure-draft - Somehow need to bring metadata and pr... Date: March 15, 2019 at 5:00:35 PM CDT To: <rross(a)mcs.anl.gov<mailto:[email protected]>> Reply-To: Reply <d+AORGpRdXmTdFUNT9SR6ZjrbUatbnvjYyHZ7Dhg0SydnZJyCeI1nFoakWoss3FOs-t3_shzcbtOO0LG11TqUmeBSgZbhuqdM02VXb51z5ESapLnQoYmcutlg(a)docs.google.com<mailto:d+AORGpRdXmTdFUNT9SR6ZjrbUatbnvjYyHZ7Dhg0SydnZJyCeI1nFoakWoss3FOs-t3_shzcbtOO0LG11TqUmeBSgZbhuqdM02VXb51z5ESapLnQoYmcutlg@docs.google.com>> Quincey Koziol replied to a comment in ssio-brochure-draft<https://docs.google.com/a/lbl.gov/document/d/1MK_lZVR6OF4pWHyztWaInHcHu4Q0E…> [Rob R.] Rob R. Capitalizing on advances in hardware and software Key question: How do we maintain productivity of scientists while leveraging complex and multi-layer storage systems? Somehow need to bring metadata and provenance back in, either here or in the workflow discussion [Quincey Koziol] Quincey Koziol I can try to add it to either section, but I would lean toward a new section that talks about "science reproducibility" and add it in there. What's your preference? Open<https://docs.google.com/a/lbl.gov/document/d/1MK_lZVR6OF4pWHyztWaInHcHu4Q0E…> Google LLC, 1600 Amphitheatre Parkway, Mountain View, CA 94043, USA You have received this email because you are a participant in this thread.Change what Google Docs sends you.<https://docs.google.com/document/u/103628923157011229096/docos/notify?id=1M…>You can reply to this email to reply to the discussion. [https://www.gstatic.com/images/branding/googlelogo/2x/googlelogo_tm_black54…] _______________________________________________ Ssio-services mailing list Ssio-services(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-services

1 0

Re: [Ssio-2018] more on brochure
by Ross, Robert B. 15 Mar '19

15 Mar '19

This might be a nice way to work it. First we need to see if we can drum up good images for the other programs, I guess. -- Rob On Mar 14, 2019, at 12:54 PM, Carns, Philip H. <carns(a)mcs.anl.gov<mailto:[email protected]>> wrote: I thought that the report cover page was really nice. If we were to do a mosaic, maybe one approach would be to feature that image again, but this time as the largest pane in a mosaic, with the smaller panes each having one example of a scientific visualization? That would emphasize the storage (and get more mileage out of that graphic) while also giving a nod to the range of science that it enables. thanks, -Phil On 2019-03-14 12:27:51-04:00 ssio-2018 wrote: I think a mosaic is definitely doable if we target science domains (one per office?) that are driven by I/O and have one visualization from each. I see that extreme heterogeneity is predominantly a tape silo, many wind farms, and many printed circuit boards. If that's the bar, we could totally surpass that and get brownie points for representing the other SC offices. Glenn On Mar 14, 2019, at 08:06, Ross, Robert B. via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]>> wrote: Hi again, I had a nice chat with Tom Peterka last week regarding getting the brochure together from a layout/etc. perspective. He suggested that we also have a look at the microelectronics brochure: https://drive.google.com/open?id=1ksdbBfzXw9_EpiZeggiRC8d74VCTKXli Would it make sense for us to use our nice image on the front, as a single image, rather than trying to drum up a dozen or so images to go all over the cover? My concern is that typically it's tough to find good storage and I/O related images. Thoughts? Rob _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018

1 0

Re: [Ssio-2018] more on brochure
by Ross, Robert B. 15 Mar '19

15 Mar '19

Ok, so if we were going to do this, then we would need images from BER, BES, FES, HEP, and NE? Did I forget someone? -- Rob On Mar 14, 2019, at 11:26 AM, Glenn Lockwood <glock(a)lbl.gov<mailto:[email protected]>> wrote: I think a mosaic is definitely doable if we target science domains (one per office?) that are driven by I/O and have one visualization from each. I see that extreme heterogeneity is predominantly a tape silo, many wind farms, and many printed circuit boards. If that's the bar, we could totally surpass that and get brownie points for representing the other SC offices. Glenn On Mar 14, 2019, at 08:06, Ross, Robert B. via ssio-2018 <ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]>> wrote: Hi again, I had a nice chat with Tom Peterka last week regarding getting the brochure together from a layout/etc. perspective. He suggested that we also have a look at the microelectronics brochure: https://drive.google.com/open?id=1ksdbBfzXw9_EpiZeggiRC8d74VCTKXli Would it make sense for us to use our nice image on the front, as a single image, rather than trying to drum up a dozen or so images to go all over the cover? My concern is that typically it's tough to find good storage and I/O related images. Thoughts? Rob _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018

1 0