Skip to content

Open espi meeting minutes, 16 april 2012

ahendry edited this page Apr 23, 2012 · 9 revisions

Open ESPI Meeting

16 April 2012

9PT/10MT/12:00ET US

Attendees:

Deepak
Anne Hendry
Dave Krinkel
Jay Mitra
Jeremy Roberts
John Teeter

Agenda:

  1. OpenESPI source code architecture and repository (continued from 2 weeks ago)

  • Spring framework, OAuth

  • Developing an architecture to support porting (common library framework, interfaces), porting requirements (.net, persistence), ....

  • Populating the db with SDG&E data

  • Time zone rationalization

  • Ids, links, organization of data blocks

  1. Connectivity Week plug-fest / demo

  • partners

  • system support

Discussion:

  1. Connectivity Week Plug-In demo

  • System support

    There is a meeting each Friday at 9:30 PDT for those interested in participating in the Plug-In activities. John will send out the meeting information for others who would like to attend those meetings. To participate in the plug-in you must be a CW sponsor. Population of the GB part of plug-fest is fairly light so far – not certain if there are any committed players. Our participation would be limited because right now OpenESPI is not focused on upload and download of data files. We might look at doing that.

  • Partners

    Can discuss more on OpenADE call tomorrow and we can all think about other organizations we might be able to partner with.

  1. OpenESPI source code architecture and repository (John)

  • Spring framework, OAuth

    In terms of getting consistency of class structures, John has found some additional tools that use the Roo scripting language to create xsd and generate classes from xsd. This will result in a full re-population of the repository when complete. At that point will have a fully generated system, including annotations that will take us to the actual Atom interactions. This is quite exciting so glad we've held off doing the hand coding. Will have in place shortly.

  • Developing an architecture to support porting (common library framework, interfaces), porting requirements (.net, persistence), ....

    We need to work on building a strawman around a common library framework and see how this might move us closer to achieving cross-framework / cross-language support. Roo generates Java code (an inherited property from the Spring framework). It might be able to generate other code, but have only seen it referenced from within Spring. The OMG is generating multiple frameworks, so will speak to them at the upcoming F2F this June to see if their approach might provide a long-term solution for C++ and others.

  • Populating the db with SDG&E data

    Ron is hoping to get a sample script that shows the example of data population – what is there now is a skeletal (incomplete) schema. John will be generating the final structures for this in the next few days. Ron can be done with the xsl transformation a few days after that. We may wan to think about a no-sql generation as well – what is there now is an sql script and we may want to go beyond that.

  • Time zone rationalization

    SDG&E sample data start times were local time which brings up the question of what time zone is specified as 'standard'. Ron exported this file from the MS Excel into xml and assumed it was in UTC. If this needs to be changed, that capability could be built into the tool. Dave K noted that they get thousands of interval data files in different formats. A vast majority have a gap at the start of daylight savings and an overlap and the end. Most of the time the gaps/overlaps are averaged. Expected behaviour around time zones is still an open issue. A time zone should be added and start times should be UTC. This includes interval data. Dave K. mentioned that the problem with using local time zones is that you can't then effectively integrate multiple loads (for example, to run analysis) and can't aggregate. John agreed this will be problematic for provider that cross time zones. AI: We can raise this in the OpenADE call tomorrow -- whichever resolution comes through, the sample data we'll be using in May will inherit that decision.

  • Ids, links, organization of data blocks

    The interval readings and blocks were fairly large in the sample data. We need a general way to present more flexible structures. The thought is that if we requested time T1 to T2, we'd wrap that in one interval block. In this implementation we haven't gotten that far yet. Sure this will be resolved in that exercise.

  1. Other

    John will be traveling for the next two weeks; expecting to do code check-in mid-week.

Related links

Energy OS home page: http://energyos.org/

OpenESPI home page: http://www.openespi.org/

General OpenESPI information: https://github.com/energyos/OpenESPI/wiki/

Ubuntu Virtual Machine distribution: http://www.openespi.org/vm.html

VM setup instructions: https://github.com/energyos/OpenESPI/wiki/Virtual-Machine-Setup

OpenESPI source code repository: https://github.com/energyos/OpenESPI

GitHub issues: https://github.com/energyos/OpenESPI/issues

Relevant OpenADE issues: http://osgug.ucaiug.org/HelpDesk/Lists/servicerequests/GreenButton.aspx

GB Initiative: http://collaborate.nist.gov/twiki-sggrid/bin/view/SmartGrid/GreenButtonInitiative
GB Home:
http://www.greenbuttondata.org/