Since the August Project and Community Workshop, Bob and I have been working to set up a process for international groups to obtain LSST data rights, following the guidance given to us by the US funding agencies earlier this year. This update is to let you all know how this will work in practice, so that we can all support our international colleagues. It’s quite a long post, but this is a weighty and complex topic - indeed, comments and questions are most welcome in the thread below.
Background
There will be two types of LSST data, prompt products (most notably a stream of alerts), and data release products, generated approximately annually. The alerts will be public, while the data releases will have a proprietary period of two years during which they will be available only to people with LSST data rights. All US scientists will have LSST data rights, as will non-US scientists covered by data rights agreements with the US agencies (or their designates). Prior to May 2019, international groups were invited to enter into data rights agreements with the LSST Corporation (LSSTC): the deal was that they would help pay for LSST operations, in return for data rights for some number of principle investigators (PIs) and their junior associates (JAs, 4 per PI). Some 35 MOAs were signed, promising data rights for over 200 PIs and 800 JAs. Earlier this year, the LSST’s US funding agencies, the NSF and DOE, moved to a different model, in which they cover the cost of LSST operations themselves, but seek to offset some of their costs and add value to the overall LSST enterprise via international groups making in-kind contributions in return for data rights, rather than financial ones. Instead of “pay to play,” this is a more collaborative model - and one which should help both the LSST facility and the LSST science community succeed (by reducing funding risk to the facility, and enabling international groups to do more with their money). You can read the US agency talking points here. LSST data rights will be established by new agreements with the NSF and DOE (or one of their designates) and international members. Watch out for a separate post later this week about LSST data policy.
While we are open to new proposals of in-kind contributions in exchange for LSST data rights for groups without a prior agreement with LSSTC, we will be focusing on the existing LSST science community (those with LSSTC MOAs) in the first instance. We are in contact with the current MOA holders, and will be working with them to help them propose meaningful in-kind contributions that “expand the resources available to the US science community.” We are also discussing with a small number of groups some particular in-kind contributions that could offset the cost of LSST operations. While we have a little under two years before the grace period set by NSF and DOE expires, we want to move quickly to help our international colleagues get new agreements in place well in advance of that.
Timeline
Here’s what we’ll be asking the international groups to do, and by when - with explanations below.
- Invitation to current MOA holders and interested groups, October 31 (this Thursday)
- Letters of Intent due November 22 (see below)
- Proposals due March 31
- Evaluation feedback provided by May 31
- Iterate to convergence on draft agreement text, summer 2020
There’s not much time before the November deadline, but we think that’s OK: most groups have been thinking hard since May about the kinds of contributions they could make, and on top of that, we really just need an early indication of the number of PIs each group is looking to get data rights for, and roughly what they could contribute in-kind. We’re seeing these LOIs as conversation starters. The US agencies will review each one, and once they confirm that a letter is appropriate for further development and likely to produce an agreement for exchange of data rights, we’ll work with that group to help them work out a more detailed proposal. By December we should be able to start making sure that each group is in contact with the appropriate facility team or science collaboration, who can help them better target their proposed contributions. International PIs will benefit from good interactions in their science collaborations.
Contribution Evaluation
We are establishing a special in-kind contribution evaluation committee (CEC) to evaluate proposed in-kind contributions, starting with the letters of intent. This committee will be appointed by the LSST Operations Director, with a charge and composition following recommendations from the LSST Science Advisory Committee (SAC). The CEC will include representatives of the eight LSST science collaborations as well as several at-large members, in order to ensure that each in-kind contribution is accurately targeted and then fully embedded, so that its benefits are real and far-reaching.
In its assessments, the CEC will be looking for proposed in-kind contributions that represent commitments to the success of LSST that are at least comparable to those that were made in the original MOAs with LSSTC (although note that in-kind contributions need not be sustained over the lifetime of the LSST survey in order to support PI data rights over the same time period). We will take steps to avoid conflicts of interest that CEC members might have.
The CEC will be a standing committee: it will monitor the in-kind contributions over time (collecting input from us in the facility as well as the science collaborations) and give feedback to an agency-level Resource Board that can help solve any problems that arise.
Acceptable In-kind Contributions
Contributions that could offset LSST Facility Operations costs will be dealt with by direct interaction between LSST and certain groups, and won’t be requested in our invitation. However, contributions that could enhance the LSST system for the benefit of the US science community may be proposed.
We’ll also be looking for in-kind contributions that expand the resources available to the US science community. Examples of the type of in-kind contribution that are likely to be acceptable following a positive CEC evaluation include:
- Observing time, dedicated to proposals led by US PIs, at key non-US facilities. (Observing time contributed only to subsets of the US community will not be accepted by the US agencies.)
- Access to surveys or proprietary datasets of high value to the US community, including (but not restricted to) datasets complementary to the LSST survey and which enable high priority LSST science.
- Dedicated software development effort, to be either embedded in one or more LSST Science Collaborations and assigned to a needed analysis pipeline, or focused on a particular enhancement to the LSST system.
- Dedicated and appropriately supported computing resources, made available to one or more LSST Science Collaborations for their analyses.
As a rough guide, and in order to achieve “comparable commitments to the ones that were made in the original MOAs,” we’ll assume that 1 FTE year of dedicated effort (at the appropriate skill level) will be approximately sufficient to obtain data rights for 1 PI for the duration of the survey (13 years, US FY22 through FY34). The equivalent cost of the US agencies providing the same 1 FTE year of effort is about $300k; we can use that number as a rough guide when considering the value of other resources, including observing time, again using US prices. For example, computing resources purchased in the US cost roughly $10k per million CPU hours, and $200k per Pb of disk storage, and so 30M CPU hrs, or 1.5 PB of disk space, would be approximately sufficient to obtain data rights for 1 PI for the duration of the survey (modulo CEC evaluation of their location and distribution in time). All these “exchange rates” are only approximate: we’ll ask the CEC to weigh in on the value of each contribution being proposed. Past contributions by PIs (and their groups) that would fit into one of the categories outlined above will also be taken into account.
Summary
This is the in-kind scheme above that we’re rolling out this week. There will be challenges at all stages, to be met by not only the international groups putting together proposals, but also the US-led community who will be evaluating those proposals. However, I hope you can see that we all have a great opportunity here: instead of having to pay for LSST operations, those of us outside the US will be able to use our local funding locally, while adding value to the whole LSST enterprise. If we do this right, we could all end up in a much stronger and better-supported science community.
The experience to date shows that people have a lot of questions about this topic! Please feel free to ask away in the comments under this post, and Bob and I will answer them as best we can, and also digest them into a FAQ for your reference.
Best wishes,
Phil