In my review of oBIX's new life in the OASIS I now have a better understanding of the true power of Project Haystack http://project-haystack.org/.

Although I have always been a strong supporter of this open source movement, I like many did not completely understand that as well as embodying a methodology for tagging data to describe its meaning the community has also developed a protocol to move this named data.

In several emails to the industry I have learned;

- oBIX and Haystack are both excellent to make data integration easy

- Haystack is good for tagging data so we can understand it

- BACnet is very expensive and difficult to use for data integration (at least from IT perspective)

To enable more streamlined communications of data between systems and software applications being based on the vendors own proprietary brew of IT standards we need to guide industry to use the evolving standards of either the open source Project Haystack or OASIS oBIX. This is as important to the industry as the adoption of open field bus standards like BACnet and Lon.  If not controlled the customer will need to become part of the vendors specific data communication approach and the data is not truly free. I am sure that no client wants to enter into a service contract to maintain data flow of a proprietary brew of IT standards yet that is where they are heading.

Our growth as an industry plus our acceptance as an industry of adding value to our real time data on its journey to the cloud depends on us supporting only a few data standards.

In the Haystack HTTP protocol

 - the data model is table based versus previous XML based approaches

 - table data is much easier to map to CSV, Excel, and other reports making it a much simpler and more flexible data model

- Haystack provides very simple solutions to common problems like querying points lists, modeling sites/equip, history data synchronization

To move ahead we need to specify or add to our mandatory requirements that vendor systems support either oBIX or Haystack and preferably both.

A test procedure needs to be developed by both to insure compliance.

There is a guide spec available developed based on Haystack community input:

http://www.skyfoundry.com/file/50/Guide-Specification-for-Data-Modeling-Standard.docx

In email discussions with Paul Ehrlich, PE President Building Intelligence Group LLC and the orignal chairman of the oBIX committe I gleened this insight:

I think that we are going to continue to see BACnet and perhaps LonTalk as the field bus protocols, oBIX will be used as a data exchange option to share data between systems.  But the main communications between systems and servers will be based on the vendors own proprietary brew of IT standards. 

A few comments in this

- first oBIX was done a long time ago so it isn't surprising that new technologies are coming into favor. Ideally if there is adequate support it can evolve.

Second oBIX was intended not for controls (that is where BACnet and other protocols shine) but for communications between controls and IT.

Finally there is some continuity - Brian Frank did most of the development of oBIX when at Tridium and has now developed Haystack while at SkyFoundry.

As always, however, the adoption is really about business and not the technology.

Paul

[an error occurred while processing this directive]From down under RAV PANCHALINGAM | Director R&D website: www.vaegroup.com.au

•    With the introduction of a more relaxed and better defined (via best practice) framework in Haystack, oBIX has lost a few step-holds on where it’s considered to be placed in the market. Hierarchy structure for asset management can be much more easily performed in Haystack then it can in oBIX.
•    oBIX is specified around XML and (in the software world at least) XML is slowly moving out of fashion as JSON is moving in. Haystack supports JSON where oBIX, in it’s immediate form, does not.

From Toby Considine this update of oBIX

Most of the other “new features” of the Haystack Protocol (improved formats for telemetry, support for JSON, etc.) are included in the update, with the difference being that they actually have formal definitions for conversion between formats, and those definitions can be machine implemented.

The current team on OBIX includes active participation from DOD ACER CERL, from IBM, from Vienna Institute of Technology, and is linked to numerous ETSI projects, including ones multiple open source projects on smart energy and the Internet of Things.

Project Haystack resources on AutomatedBuildings.com web site

John Petze  Partner at SkyFoundry Haystack data modeling techniques can be used with virtually any type of system data. It's not tied to any vendor, or communication protocol. It can be used with legacy system data and with more modern systems that allow tags to be defined in the end device. It can also be used with file data – like csv files, and Excel files. "Project Haystack. This eight minute video provides a great overview describing what it is about and why it is needed."  Link: http://youtu.be/5C6GwLbYqTw

http://www.automatedbuildings.com/news/dec13/articles/haystack/131126113003haystack.html

http://www.automatedbuildings.com/news/nov12/interviews/121016105005frank.html

http://www.automatedbuildings.com/news/jul13/interviews/130625021404petze.html

http://www.automatedbuildings.com/news/may11/interviews/110428100202frank.html

http://www.automatedbuildings.com/news/jan13/interviews/121218030303petze.html

[an error occurred while processing this directive]
[Click Banner To Learn More]

[Home Page]  [The Automator]  [About]  [Subscribe ]  [Contact Us]