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

SSIO 2018 Workshop Planning
by Ross, Robert B. 13 Jun '18

13 Jun '18

Hi all, I've chatted with all of you previously about this, but I've been lax in getting moving. Lucy has asked me to organize a workshop to help define priority research directions for scalable storage and I/O (SSIO), ideally leading to a call for proposals some time in 2018. Additionally, there is some interest from Almadena on the NSF side to restart an inter-agency activity in this space. We will be the organizers that drive the ASCR process forward. We'll see also how things develop on the inter-agency front and what role we might play in that process. Our goal is to organize our thoughts about future research directions, produce an initial document capturing that thinking and describing state of the art, plan and hold a workshop to identify a set of priority research directions, and produce a final workshop report capturing this information. My initial plan is based off the experiences from the old HEC-FSIO process, the more recent 2014 ASCR SSIO workshop (for which there is a 2015 report), and the most recent Extreme Heterogeneity workshop planning (for which we will have the workshop soon). All of this is up for discussion/negotiation. Summary and action items. - Think about what you think the big issues are in SSIO research for enabling DOE science, talk with your team and colleagues. - Start putting together the "virtual" summit materials. - Look for the Doodle poll for a summit date. - Look for a Doodle poll for a time we could regularly meet to discuss progress. Details follow. Thanks in advance for your help in making this a success! Rob --- Organizers. Here's our team: - Gary Grider (LANL) - Scott Klasky (ORNL) - Glenn Lockwood (LBNL) - Kathryn Mohror (LLNL) - Lee Ward (SNL) Lucy Nowell is our ASCR contact, and I expect that Laura Biven will be included and possibly take over as contact before we are done. We have a mailing list: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> for discussion. You have been added to the list. Please let me know ASAP if for some reason you cannot help. Topical areas and initial discussion. Five focus areas were discussed for further investigation at the prior workshop. - HW/SW architectures - Metadata, name spaces, and provenance - Supporting science data - Integrating with external services - Understanding SSIO systems I propose that we do a "virtual" summit (i.e., videoconference) for 3/4th of a day, rather than flying somewhere for a day to meet in person. Each organizer would present their view on challenges as relate to mission and discuss if/how they think we should adjust the topical areas in response to changes since 2015. I think of these as 20 minute presentations that are devoid of our own research specifics/propaganda and focus on these big picture items. From this we would select a set of topical areas used to organize the factual status document (see below). Factual status document. We are modeling our process after the BES "basic research needs" approach (see attachment), with adaptations to the scope of the area and ASCR needs. This process includes the generation of a factual status document (FSD) prior to the workshop. There is a separate group working down this path for an Extreme Heterogeneity FSD and workshop, too. The FSD can be thought of as the background sections of the final workshop document, and that's how we will approach it. With this in mind, we will organize the document around our initial guesses at topical areas, understanding that the topical areas may have to be adjusted in the final report as a result of discussion. The goal would be to complete the following for the FSD: - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps -- what we need to know that we don't know - Research Needs - Topic 1 - Definition/summary of area - Current status - Topic 2 ... - Appendix: References This is likely to be a 50 page document. This document would be partitioned into sections to be managed by individual organizers. Rather than requesting white papers from potential attendees, organizers would invite experts in respective areas to participate in the development of requisite sections. It might be easiest to develop this document using an online system such as Google Docs, understanding that eventually it would be turned into a Word document and finally into a PDF for publication as the workshop report. Position papers. In the Extreme Heterogeneity planning, we asked for position papers. We got over 100 of them. Parsing them has been a real challenge for the organizers. We should think about what approach we might employ for gathering input from the community prior to the workshop and what we would do with that material prior to the workshop vs. incorporating into the final report. Workshop. We will shoot for around 60 attendees, maybe 30/20/10 lab/university/industry. This helps keep the justification and planning fairly simple. We would plan for a 1.75-day workshop to facilitate covering five topical areas as separate, single-track discussions. If for some reason we end up needing only 1.5 days for the meeting proper, this would leave more time for the organizers to meet and discuss report completion. We will (mostly) assume that attendees have read the FSD prior to arrival at the meeting. Never the less, we will spend 1/2 day bringing the attendees up to speed, focusing on the background material from the FSD. This is similar to the opening plenary session portion of the BES workshop structure. The remainder of the meeting could follow the model from the 2014 SSIO meeting: 90-minute moderated discussion sessions, with notes captured live in PPT. With 1.75 days we can accommodate 5 total discussion topics. We will place a topic of broad interest last to encourage attendees to participate for the full meeting. These sessions will identify current status in the area, its CS challenges, its opportunities, and its research needs, reflecting the breakout session portion of the BES workshop structure. Because the meeting is single track, there is no need for outbrief sessions (i.e., the participants will have been present for all discussions). Report. The final report will be generated by augmenting the FSD with material summarizing the outcomes of the workshop. We will target a 75-100 page document (the 2015 report was approximately 90 pages). The outline for the report will look something like this: - Executive Summary - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps (need to understand what this typically includes) - Research Needs - Topic 1 - Definition/summary of area - Current status - CS challenges and opportunities - Priority research directions (3-5) - Topic 2 ... - Priority research directions

8 44

Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning
by Mohror, Kathryn 08 Mar '18

08 Mar '18

Hi Scott, I see your point - thanks. I'll add two columns then. Kathryn From: Klasky, Scott A. <klasky(a)ornl.gov> Sent: Thursday, March 8, 2018 3:39 PM To: Mohror, Kathryn <mohror1(a)llnl.gov>; Ross, Robert B. <rross(a)mcs.anl.gov> Cc: ssio-2018(a)lists.mcs.anl.gov Subject: RE: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning The problem is that that would be a huge amount of people working on ML, visualization, ..... I know that the area I am suggested is funded by the current SSIO program so it seems appropriate Scott ________________________________ From: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> Date: March 8, 2018 at 6:32:07 PM EST To: Klasky, Scott A. <klasky(a)ornl.gov<mailto:[email protected]>>, Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> <ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: RE: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Hi Scott, What about a single category of "Analytics" to cover both of those ideas? I think it's a bit of a stretch for compression, but "good enough" so that we are covering the bases for expertise and we don't end up with a laundry list. Thanks, Kathryn From: Klasky, Scott A. <klasky(a)ornl.gov<mailto:[email protected]>> Sent: Thursday, March 8, 2018 3:13 PM To: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>>; Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning I would add 1. In situ infrastructure - coupling..... 2. Data reduction/compression Scott ________________________________ From: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> Date: March 8, 2018 at 5:58:31 PM EST To: Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> <ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Hi all, Here's a link to the worksheet we use to sort out the attendee list: https://docs.google.com/spreadsheets/d/16oeV2PAaOdeuKzjWJ1M8tF45mgzSSMuxzCg… A thing I was not sure of was: what categories do we want to use to ensure that we are getting a good spread of expertise across attendees at the meeting? To start, I chose: * Storage Systems * I/O models * Applications/Use cases Do those seem appropriate? Can you think of others? If you think other categories are appropriate as you are adding people to the list, I suggest creating a column with a heading for your new expertise category, or at the very least put something in the "notes" column. For each attendee that you add to the worksheet, put an 'x' in the column(s) that best represents the attendee's expertise. Then we can tally them up to see if we have a good balance of people. Thanks, Kathryn From: Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Sent: Thursday, March 8, 2018 1:43 PM To: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Thanks! -- Rob On Mar 8, 2018, at 3:34 PM, Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> wrote: Hi Rob, I'll volunteer to manage the attendees spreadsheet. I'll send out a link to the list in a bit. Kathryn From: ssio-2018 <ssio-2018-bounces(a)lists.mcs.anl.gov<mailto:[email protected]>> On Behalf Of Ross, Robert B. Sent: Thursday, March 8, 2018 8:49 AM To: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Forgot the attachment...classic... On Mar 8, 2018, at 8:02 AM, Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> wrote: Hi all, Scott asked a good question that I will summarize as "what is going on?". Here's what I know: Virtual Summit. Lucy has asked for permission to have our virtual summit, back on the 21st. I will double-check with her on status of that request. Meanwhile we should all gather our thoughts for the summit: - Think about what you think the big issues are in SSIO research for enabling DOE science, talk with your team and colleagues. - Fill out this poll; we're optimistic that we will have permission to do this soon: https://doodle.com/poll/mvfxb8fcqucimtim Scope. We've been asked to think about the data and learning applications in addition to traditional simulation. I don't think this is a surprise, but it does mean that we need to think about those communities as we think about who to include in the discussion/workshop and what to have in our initial report. Organizers. Gary has asked to take a step back but is happy to chime in as needed. Here's our revised team: - Brad Settlemyer (LANL) - Scott Klasky (ORNL) - Glenn Lockwood (LBNL) - Kathryn Mohror (LLNL) - Lee Ward (SNL) I think everyone probably knows Brad. I'm sure he'll do a great job. Gary will remain on the mailing list, too. Meeting Dates. Lucy and I have chatted on possible dates, and we want to get some dates on the calendar. We're looking for two days in July/August. Here's a short set of dates that work for me and Lucy; we're hoping to get lucky and get a hit. Note that we didn't include Mon. or Fri. in these so people can travel during the week: - https://doodle.com/poll/nw4xsqpx2agd3dhh Attendees. We should start thinking about attendees. I've attached the list of people from the SSIO workshop back in late 2014 as a starting point. Can someone volunteer to create a Google spreadsheet that we could use for organizing possible participants. Note that we don't have permission to have the meeting at this point, so we're really just discussing who we think key thinkers are in the area. I'm also attaching the EH and ML pre-workshop documents as examples of the kind of report we need to prepare. Thanks, Rob --- On Mar 1, 2018, at 4:17 PM, Ward, Lee <lee(a)SANDIA.GOV<mailto:[email protected]>> wrote: Rob tasked me with trying to kick this off. I had something different in mind but I suspect a call is a better idea. Thanks! Sent from my iPhone On Mar 1, 2018, at 2:34 PM, Klasky, Scott A. <klasky(a)ornl.gov<mailto:[email protected]>> wrote: Hi Rob, Has there been a doodle poll for this? Scott From: ssio-2018 [mailto:[email protected]] On Behalf Of Ross, Robert B. Sent: Friday, January 12, 2018 1:18 PM To: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] SSIO 2018 Workshop Planning Forgot the attachment, of course. -- Rob On Jan 11, 2018, at 10:32 AM, Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> wrote: Hi all, I've chatted with all of you previously about this, but I've been lax in getting moving. Lucy has asked me to organize a workshop to help define priority research directions for scalable storage and I/O (SSIO), ideally leading to a call for proposals some time in 2018. Additionally, there is some interest from Almadena on the NSF side to restart an inter-agency activity in this space. We will be the organizers that drive the ASCR process forward. We'll see also how things develop on the inter-agency front and what role we might play in that process. Our goal is to organize our thoughts about future research directions, produce an initial document capturing that thinking and describing state of the art, plan and hold a workshop to identify a set of priority research directions, and produce a final workshop report capturing this information. My initial plan is based off the experiences from the old HEC-FSIO process, the more recent 2014 ASCR SSIO workshop (for which there is a 2015 report), and the most recent Extreme Heterogeneity workshop planning (for which we will have the workshop soon). All of this is up for discussion/negotiation. Summary and action items. - Think about what you think the big issues are in SSIO research for enabling DOE science, talk with your team and colleagues. - Start putting together the "virtual" summit materials. - Look for the Doodle poll for a summit date. - Look for a Doodle poll for a time we could regularly meet to discuss progress. Details follow. Thanks in advance for your help in making this a success! Rob --- Organizers. Here's our team: - Gary Grider (LANL) - Scott Klasky (ORNL) - Glenn Lockwood (LBNL) - Kathryn Mohror (LLNL) - Lee Ward (SNL) Lucy Nowell is our ASCR contact, and I expect that Laura Biven will be included and possibly take over as contact before we are done. We have a mailing list: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> for discussion. You have been added to the list. Please let me know ASAP if for some reason you cannot help. Topical areas and initial discussion. Five focus areas were discussed for further investigation at the prior workshop. - HW/SW architectures - Metadata, name spaces, and provenance - Supporting science data - Integrating with external services - Understanding SSIO systems I propose that we do a "virtual" summit (i.e., videoconference) for 3/4th of a day, rather than flying somewhere for a day to meet in person. Each organizer would present their view on challenges as relate to mission and discuss if/how they think we should adjust the topical areas in response to changes since 2015. I think of these as 20 minute presentations that are devoid of our own research specifics/propaganda and focus on these big picture items. From this we would select a set of topical areas used to organize the factual status document (see below). Factual status document. We are modeling our process after the BES "basic research needs" approach (see attachment), with adaptations to the scope of the area and ASCR needs. This process includes the generation of a factual status document (FSD) prior to the workshop. There is a separate group working down this path for an Extreme Heterogeneity FSD and workshop, too. The FSD can be thought of as the background sections of the final workshop document, and that's how we will approach it. With this in mind, we will organize the document around our initial guesses at topical areas, understanding that the topical areas may have to be adjusted in the final report as a result of discussion. The goal would be to complete the following for the FSD: - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps -- what we need to know that we don't know - Research Needs - Topic 1 - Definition/summary of area - Current status - Topic 2 ... - Appendix: References This is likely to be a 50 page document. This document would be partitioned into sections to be managed by individual organizers. Rather than requesting white papers from potential attendees, organizers would invite experts in respective areas to participate in the development of requisite sections. It might be easiest to develop this document using an online system such as Google Docs, understanding that eventually it would be turned into a Word document and finally into a PDF for publication as the workshop report. Position papers. In the Extreme Heterogeneity planning, we asked for position papers. We got over 100 of them. Parsing them has been a real challenge for the organizers. We should think about what approach we might employ for gathering input from the community prior to the workshop and what we would do with that material prior to the workshop vs. incorporating into the final report. Workshop. We will shoot for around 60 attendees, maybe 30/20/10 lab/university/industry. This helps keep the justification and planning fairly simple. We would plan for a 1.75-day workshop to facilitate covering five topical areas as separate, single-track discussions. If for some reason we end up needing only 1.5 days for the meeting proper, this would leave more time for the organizers to meet and discuss report completion. We will (mostly) assume that attendees have read the FSD prior to arrival at the meeting. Never the less, we will spend 1/2 day bringing the attendees up to speed, focusing on the background material from the FSD. This is similar to the opening plenary session portion of the BES workshop structure. The remainder of the meeting could follow the model from the 2014 SSIO meeting: 90-minute moderated discussion sessions, with notes captured live in PPT. With 1.75 days we can accommodate 5 total discussion topics. We will place a topic of broad interest last to encourage attendees to participate for the full meeting. These sessions will identify current status in the area, its CS challenges, its opportunities, and its research needs, reflecting the breakout session portion of the BES workshop structure. Because the meeting is single track, there is no need for outbrief sessions (i.e., the participants will have been present for all discussions). Report. The final report will be generated by augmenting the FSD with material summarizing the outcomes of the workshop. We will target a 75-100 page document (the 2015 report was approximately 90 pages). The outline for the report will look something like this: - Executive Summary - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps (need to understand what this typically includes) - Research Needs - Topic 1 - Definition/summary of area - Current status - CS challenges and opportunities - Priority research directions (3-5) - Topic 2 ... - Priority research directions _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018 _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018 <2018-01-15-eh-pre-workshop-document.pdf><ml-pre-workshop-doc-20180124.pdf>_______________________________________________ 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] [EXTERNAL] SSIO 2018 Workshop Planning
by Mohror, Kathryn 08 Mar '18

08 Mar '18

Hi Scott, What about a single category of "Analytics" to cover both of those ideas? I think it's a bit of a stretch for compression, but "good enough" so that we are covering the bases for expertise and we don't end up with a laundry list. Thanks, Kathryn From: Klasky, Scott A. <klasky(a)ornl.gov> Sent: Thursday, March 8, 2018 3:13 PM To: Mohror, Kathryn <mohror1(a)llnl.gov>; Ross, Robert B. <rross(a)mcs.anl.gov> Cc: ssio-2018(a)lists.mcs.anl.gov Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning I would add 1. In situ infrastructure - coupling..... 2. Data reduction/compression Scott ________________________________ From: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> Date: March 8, 2018 at 5:58:31 PM EST To: Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> <ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]>> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Hi all, Here's a link to the worksheet we use to sort out the attendee list: https://docs.google.com/spreadsheets/d/16oeV2PAaOdeuKzjWJ1M8tF45mgzSSMuxzCg… A thing I was not sure of was: what categories do we want to use to ensure that we are getting a good spread of expertise across attendees at the meeting? To start, I chose: * Storage Systems * I/O models * Applications/Use cases Do those seem appropriate? Can you think of others? If you think other categories are appropriate as you are adding people to the list, I suggest creating a column with a heading for your new expertise category, or at the very least put something in the "notes" column. For each attendee that you add to the worksheet, put an 'x' in the column(s) that best represents the attendee's expertise. Then we can tally them up to see if we have a good balance of people. Thanks, Kathryn From: Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> Sent: Thursday, March 8, 2018 1:43 PM To: Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> Cc: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Thanks! -- Rob On Mar 8, 2018, at 3:34 PM, Mohror, Kathryn <mohror1(a)llnl.gov<mailto:[email protected]>> wrote: Hi Rob, I'll volunteer to manage the attendees spreadsheet. I'll send out a link to the list in a bit. Kathryn From: ssio-2018 <ssio-2018-bounces(a)lists.mcs.anl.gov<mailto:[email protected]>> On Behalf Of Ross, Robert B. Sent: Thursday, March 8, 2018 8:49 AM To: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] [EXTERNAL] SSIO 2018 Workshop Planning Forgot the attachment...classic... On Mar 8, 2018, at 8:02 AM, Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> wrote: Hi all, Scott asked a good question that I will summarize as "what is going on?". Here's what I know: Virtual Summit. Lucy has asked for permission to have our virtual summit, back on the 21st. I will double-check with her on status of that request. Meanwhile we should all gather our thoughts for the summit: - Think about what you think the big issues are in SSIO research for enabling DOE science, talk with your team and colleagues. - Fill out this poll; we're optimistic that we will have permission to do this soon: https://doodle.com/poll/mvfxb8fcqucimtim Scope. We've been asked to think about the data and learning applications in addition to traditional simulation. I don't think this is a surprise, but it does mean that we need to think about those communities as we think about who to include in the discussion/workshop and what to have in our initial report. Organizers. Gary has asked to take a step back but is happy to chime in as needed. Here's our revised team: - Brad Settlemyer (LANL) - Scott Klasky (ORNL) - Glenn Lockwood (LBNL) - Kathryn Mohror (LLNL) - Lee Ward (SNL) I think everyone probably knows Brad. I'm sure he'll do a great job. Gary will remain on the mailing list, too. Meeting Dates. Lucy and I have chatted on possible dates, and we want to get some dates on the calendar. We're looking for two days in July/August. Here's a short set of dates that work for me and Lucy; we're hoping to get lucky and get a hit. Note that we didn't include Mon. or Fri. in these so people can travel during the week: - https://doodle.com/poll/nw4xsqpx2agd3dhh Attendees. We should start thinking about attendees. I've attached the list of people from the SSIO workshop back in late 2014 as a starting point. Can someone volunteer to create a Google spreadsheet that we could use for organizing possible participants. Note that we don't have permission to have the meeting at this point, so we're really just discussing who we think key thinkers are in the area. I'm also attaching the EH and ML pre-workshop documents as examples of the kind of report we need to prepare. Thanks, Rob --- On Mar 1, 2018, at 4:17 PM, Ward, Lee <lee(a)SANDIA.GOV<mailto:[email protected]>> wrote: Rob tasked me with trying to kick this off. I had something different in mind but I suspect a call is a better idea. Thanks! Sent from my iPhone On Mar 1, 2018, at 2:34 PM, Klasky, Scott A. <klasky(a)ornl.gov<mailto:[email protected]>> wrote: Hi Rob, Has there been a doodle poll for this? Scott From: ssio-2018 [mailto:[email protected]] On Behalf Of Ross, Robert B. Sent: Friday, January 12, 2018 1:18 PM To: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> Subject: Re: [Ssio-2018] SSIO 2018 Workshop Planning Forgot the attachment, of course. -- Rob On Jan 11, 2018, at 10:32 AM, Ross, Robert B. <rross(a)mcs.anl.gov<mailto:[email protected]>> wrote: Hi all, I've chatted with all of you previously about this, but I've been lax in getting moving. Lucy has asked me to organize a workshop to help define priority research directions for scalable storage and I/O (SSIO), ideally leading to a call for proposals some time in 2018. Additionally, there is some interest from Almadena on the NSF side to restart an inter-agency activity in this space. We will be the organizers that drive the ASCR process forward. We'll see also how things develop on the inter-agency front and what role we might play in that process. Our goal is to organize our thoughts about future research directions, produce an initial document capturing that thinking and describing state of the art, plan and hold a workshop to identify a set of priority research directions, and produce a final workshop report capturing this information. My initial plan is based off the experiences from the old HEC-FSIO process, the more recent 2014 ASCR SSIO workshop (for which there is a 2015 report), and the most recent Extreme Heterogeneity workshop planning (for which we will have the workshop soon). All of this is up for discussion/negotiation. Summary and action items. - Think about what you think the big issues are in SSIO research for enabling DOE science, talk with your team and colleagues. - Start putting together the "virtual" summit materials. - Look for the Doodle poll for a summit date. - Look for a Doodle poll for a time we could regularly meet to discuss progress. Details follow. Thanks in advance for your help in making this a success! Rob --- Organizers. Here's our team: - Gary Grider (LANL) - Scott Klasky (ORNL) - Glenn Lockwood (LBNL) - Kathryn Mohror (LLNL) - Lee Ward (SNL) Lucy Nowell is our ASCR contact, and I expect that Laura Biven will be included and possibly take over as contact before we are done. We have a mailing list: ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> for discussion. You have been added to the list. Please let me know ASAP if for some reason you cannot help. Topical areas and initial discussion. Five focus areas were discussed for further investigation at the prior workshop. - HW/SW architectures - Metadata, name spaces, and provenance - Supporting science data - Integrating with external services - Understanding SSIO systems I propose that we do a "virtual" summit (i.e., videoconference) for 3/4th of a day, rather than flying somewhere for a day to meet in person. Each organizer would present their view on challenges as relate to mission and discuss if/how they think we should adjust the topical areas in response to changes since 2015. I think of these as 20 minute presentations that are devoid of our own research specifics/propaganda and focus on these big picture items. From this we would select a set of topical areas used to organize the factual status document (see below). Factual status document. We are modeling our process after the BES "basic research needs" approach (see attachment), with adaptations to the scope of the area and ASCR needs. This process includes the generation of a factual status document (FSD) prior to the workshop. There is a separate group working down this path for an Extreme Heterogeneity FSD and workshop, too. The FSD can be thought of as the background sections of the final workshop document, and that's how we will approach it. With this in mind, we will organize the document around our initial guesses at topical areas, understanding that the topical areas may have to be adjusted in the final report as a result of discussion. The goal would be to complete the following for the FSD: - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps -- what we need to know that we don't know - Research Needs - Topic 1 - Definition/summary of area - Current status - Topic 2 ... - Appendix: References This is likely to be a 50 page document. This document would be partitioned into sections to be managed by individual organizers. Rather than requesting white papers from potential attendees, organizers would invite experts in respective areas to participate in the development of requisite sections. It might be easiest to develop this document using an online system such as Google Docs, understanding that eventually it would be turned into a Word document and finally into a PDF for publication as the workshop report. Position papers. In the Extreme Heterogeneity planning, we asked for position papers. We got over 100 of them. Parsing them has been a real challenge for the organizers. We should think about what approach we might employ for gathering input from the community prior to the workshop and what we would do with that material prior to the workshop vs. incorporating into the final report. Workshop. We will shoot for around 60 attendees, maybe 30/20/10 lab/university/industry. This helps keep the justification and planning fairly simple. We would plan for a 1.75-day workshop to facilitate covering five topical areas as separate, single-track discussions. If for some reason we end up needing only 1.5 days for the meeting proper, this would leave more time for the organizers to meet and discuss report completion. We will (mostly) assume that attendees have read the FSD prior to arrival at the meeting. Never the less, we will spend 1/2 day bringing the attendees up to speed, focusing on the background material from the FSD. This is similar to the opening plenary session portion of the BES workshop structure. The remainder of the meeting could follow the model from the 2014 SSIO meeting: 90-minute moderated discussion sessions, with notes captured live in PPT. With 1.75 days we can accommodate 5 total discussion topics. We will place a topic of broad interest last to encourage attendees to participate for the full meeting. These sessions will identify current status in the area, its CS challenges, its opportunities, and its research needs, reflecting the breakout session portion of the BES workshop structure. Because the meeting is single track, there is no need for outbrief sessions (i.e., the participants will have been present for all discussions). Report. The final report will be generated by augmenting the FSD with material summarizing the outcomes of the workshop. We will target a 75-100 page document (the 2015 report was approximately 90 pages). The outline for the report will look something like this: - Executive Summary - Introduction - Science Challenges - Technology Challenges - Knowledge Gaps (need to understand what this typically includes) - Research Needs - Topic 1 - Definition/summary of area - Current status - CS challenges and opportunities - Priority research directions (3-5) - Topic 2 ... - Priority research directions _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018 _______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018 <2018-01-15-eh-pre-workshop-document.pdf><ml-pre-workshop-doc-20180124.pdf>_______________________________________________ ssio-2018 mailing list ssio-2018(a)lists.mcs.anl.gov<mailto:[email protected]> https://lists.mcs.anl.gov/mailman/listinfo/ssio-2018

2 1