ALMA Software Scheduling subsystem Planning for cycle5 onwards Alexis Tejeda atejeda@nrao.edu

Scheduling Team, 2016 (as of cycle4) Alexis Tejeda, NRAO 80%, lead dev. Rachel Rosen, NRAO 20%, dev. Ignacio Toledo, JAO, dev. Sergio Martin, JAO subsystem scientist Maria Diaz, deputy subsystem scientist   Cognizant Scientists Sheng-Yuan Liu, EA Gerald Schieven, NA Arielle Moullet, NA 3

Software delivered as part of online releases, usually two per cycle ONLINE and OFFLINE Scheduling Online Software delivered as part of online releases, usually two per cycle Interacts with the rest of the ALMA online software (critical to perform observations), e.g.: SchedBlocks full auto transitions Scheduling Offline Doesn’t interact directly with the rest of the ALMA software, but it is used to support Online operations, e.g.: DSA Has more releases per cycle, allowing to introduce changes more quickly 4

Full Auto SchedBlock state transitions Urgent problems fixed Scheduling Cycle4 Full Auto SchedBlock state transitions Urgent problems fixed Transitions are now based on QA0 DSA (Dynamic Scheduler Algorithm) was delivered as part of the offline software Algorithm adjustments affecting SB selection 18 tickets addressed so far, consolidated in 3 patch requests Technical performance/scalability problems affecting mainly to the scheduler GUI panels, WIP. Overall, scheduling in cycle4 seems to works better compared to cycle3, according to ADE PRTS KPI October Report. 5

Scheduling Cycle4, how everything fits together Scheduler Queue 4 ONLINE DSA GUI 1,3 6 Archive 2 OFFLINE DSA

F2F meeting held in Socorro NM, August 31st to September 2nd Scheduling Cycle5 F2F meeting held in Socorro NM, August 31st to September 2nd Key points discussed: Team, roles definition and collaboration model DSACore catch up Scheduling future 7

Collaboration model DSACore (an evolution of WTO tool), is now an ICT Scheduling deliverable, maintained by Ignacio Toledo (JAO Data Analyst) Software Engineering 8 ALMA Scheduler Data Analyst Science Requirements

Collaboration model DSACore (an evolution of WTO tool), is now an ICT Scheduling deliverable, maintained by Ignacio Toledo (JAO Data Analyst) Software Engineering Technical expertise 9 ALMA Scheduler What we do really need now Data and algorithms expertise Data Analyst Science Requirements

Requirements, Sergio Martin Roles definition Requirements, Sergio Martin Graphical Interfaces (panels) Rachel Rosen Online Scheduler Alexis Tejeda ONLINE 10 Offline Scheduler (DSA or a.k.a. DSACore) Ignacio Toledo OFFLINE

Rely on few data sources as possible DSA Catch Up Problems Rely on few data sources as possible currently, DSA is relying on different data sources to feed the algorithm: array configurations, pwv data, weather, etc, which is prone error and difficult to test APDM access, what DSA needs is in XML format, we are currently exploring different technologies to parse everything in an efficient way Constant evolution of requirements Use cases and documentation SPT (scheduler planning tool) as a script-tool, will not be included in cycle5 11

Rely on few data sources as possible DSA Catch Up Rely on few data sources as possible currently, DSA is relying on different data sources to feed the algorithm: array configurations, pwv data, weather, etc, which is prone error and difficult to test APDM access, what DSA needs is in XML format, we are currently exploring different technologies to parse everything in an efficient way 12 APDMs XML are parsed are kept in memory Testing oracle events and feasibility to use a nosql database just for DSA

Constant evolution of requirements Use cases and documentation DSA Catch Up Constant evolution of requirements Use cases and documentation SPT (scheduler planning tool) as a script-tool, will not be included in cycle5 More DSA simulations are needed in order to implement an initial working prototype Constrained by observatory requirements 13

Offline scheduler (DSA) Future Scheduling After a few discussions we converged to collaborate in an efficient way, we will keep this collaboration model Offline scheduler (DSA) Continue the work to support the constant evolution of the observatory requirements Simulations Towards to a working SPT APRC feedback 14

Online Scheduler (short term, possibly cycle5) Future Scheduling Online Scheduler (short term, possibly cycle5) Improve the GUIs usability by fixing bottlenecks and memory usage efficiency Streamline the scheduling working database (SWDB) Avoid the need to import and parse APDMs for the online system, there’s relational data in materialized views to work with Improve full auto to have a bugless/bug-free SBs transition (avoiding to lose EB statuses) All the work is now being focused on the DSA mainly, at some point, hopefully the online scheduler will be stable This allows to have more room to improve testing, integration and validation phases 15

Future Scheduling Improve observing efficiency, project completion rate, number of QA2 passes, and reduces AoD work load. 16

The Atacama Large Millimeter/submillimeter Array (ALMA), an international astronomy facility, is a partnership of Europe, North America and East Asia in cooperation with the Republic of Chile. ALMA is funded in Europe by the European Organization for Astronomical Research in the Southern Hemisphere (ESO), in North America by the U.S. National Science Foundation (NSF) in cooperation with the National Research Council of Canada (NRC) and the National Science Council of Taiwan (NSC) and in East Asia by the National Institutes of Natural Sciences (NINS) of Japan in cooperation with the Academia Sinica (AS) in Taiwan. ALMA construction and operations are led on behalf of Europe by ESO, on behalf of North America by the National Radio Astronomy Observatory (NRAO), which is managed by Associated Universities, Inc. (AUI) and on behalf of East Asia by the National Astronomical Observatory of Japan (NAOJ). The Joint ALMA Observatory (JAO) provides the unified leadership and management of the construction, commissioning and operation of ALMA.