DM Monthly Status Report for December 2019


(Michelle Poland) #1

The DM monthly status report covering December activities has been posted to DocuShare, collection-972. For convenience, the High-level Summary is pasted below. Direct link to the full report http://ls.st/sbt

High-level Summary

Community Interactions, Meetings and Workshops

LDM-723, the Call for Proposals for Community Alert Brokers, was issued on December 27. All 15 groups that had previously submitted letters of intent to the project were invited to submit full proposals, which are due in June 2020.

In mid-December the Middleware Team presented the current status of the new “generation 3” middleware to the DM Leadership Team and interested others. They showed slides describing a number of new features and gave a live demo of the middleware supporting calibration products production on test data from the LATISS instrument on the prototype Batch Production Service at the LSST Data Facility. They then presented plans for ongoing middleware development in 2020.

Effort across the DM System Science Team and the Science Pipelines groups is focused on preparing for the Algorithms Workshop, which will take place in Princeton in March. A large number of expressions of interest in attending the workshop were received by the due date in mid-December; we are pleased that we will be able to accommodate everyone.

Fred Moolekamp of the Data Release Production team gave a well-received presentation on the current status of our deblending algorithms to the Dark Energy Science Collaboration’s (DESC) Blending Working Group.

Technical Progress

This month saw the release of a relatively complete draft of DMTN-135, the new sizing model and purchasing plan. Although work is ongoing, this document is now readable and forms an essential input to future procurement plans and budget discussions. It has already played a key role in developing LCR-2148, submitted in mid-December, which clarifies DM’s responsibility and budget for providing hardware to LSST Operations.

Other key documents updated this month include the DM System Requirements (LSE-61) and DM-Camera Interface Control Document (LSE-69) were updated. These now incorporate changes from a LCR-1933, LCR-1664, and LCR-836.

Over the last several months, we have developed a substantially revised set of DM milestones. These reflect lessons learned from construction to date and closer integration with the construction, systems integration and commissioning schedule of the project outside DM. These changes have been incorporated into the December project baseline.

All components of the LSST Science Platform (LSP) and the Engineering and Facility Database (EFD) are now running under the Argo Continuous Deployment Framework. These brings increased service maintainability, and decreases the effort required to bring up new instances of the services.

Version 19.0.0 of the LSST Science Pipelines was released this month. This release contains a range of new functionality and important upgrades, which are described in detail in the release notes. The release is accompanied by the characterization metric report DMTR-191.

A metric has been developed which can be used to assess the contribution of particular observations to developing models to mitigate the effects of differential chromatic refraction. The DM team will collaborate with scheduler developers to make this information available when planning the survey strategy.

The validate_drp system, which is used to track scientific performance metrics across the DM system, was updated to track the repeatability of galaxy photometry measurements, and to work with the new “FGCM” calibration system. We expect validate_drp to form a key part of ongoing verification and validation activities, and to be a vital tool in preparing for the Algorithms Workshop.

Qserv was deployed under Kubernetes for the first time at the Data Facility. Meanwhile, the NCSA test stand had EFD, Header Service, Archiver, Forwarder and Camera components, as well as a local Kubernetes cluster, installed; these are being used to test data transfers between the Camera and DM systems.

Finally, a range of new hardware has been received and is being deployed at the Data Facility. This includes 10 additional Kubernetes nodes, eight new Qserv nodes, 3 PB of data storage, and an AMD “Rome” system for testing.

LSST Atmospheric Transmission Imager and Slitless Spectrograph