As if Connectivity Week were not enough fun, CABA's Intelligent Buildings committee met on Friday after the conference. CABA was instrumental in the original founding of the oBIX initiative. Below is a summary of my report.

oBIX has been an OASIS Committee Specification since December. Delays in the IP management process as well as the review/comments/response cycle made progress slow last year. It is now in the winter of discontent when all adoption and movement appears glacial.

There are shipping applications with oBIX now. UNC is today running 70 buildings using oBIX. I am told the Department of Defense is using it in high end specialized situations, but things being what they are, I cannot say where or how. Interested parties can download the specification at:

http://www.oasis-open.org/committees/download.php/21462/obix-1.0-cs-01.zip

or can download the open source implementation at:

http://sourceforge.net/project/showfiles.php?group_id=148480

What is more interesting is the interest being shown to oBIX external to the buildings domain. The National Building Information Modeling Standard (NBIMS) and the oBIX committee have met to explore relationships between the standards, focusing on whether Energy Models developed directly against NBIMS models can be compared to live data from control systems read by oBIX to produce live models (or instantaneous commissioning). The Emergency Management Technical Committee, promulgators of the Common Alerting Protocol (CAP) would like invoke oBIX contracts directly from CAP alerts. The Open Geospatial Consortium would like to invoke contracts granting situation and spatial awareness to emergency responders. These conversations are the beginning of enterprise interactions with intelligent buildings.

During the report, I was asked about BACnet-WS and its relationship and competition with oBIX. Today, most implementers are still focused on point-to-point communication between systems using REST. Under REST, the difference between the two standards is small. As applications move toward interactions with systems outside the building system domain, wherein pervasive security, Federated identity management, cross-domain applications, long-running processes, and service orientation become more important, then the value of the SOAP binding offered by oBIX becomes of greater relative value. The applications cited above are good examples.

oBIX is currently recruiting members for the technical committee, both to flesh out errors and omissions (1.1) and to begin defining standard contracts based on the oBIX object model (2.0). oBIX is also considering forming an oBIX implementation committee. CABA members are invited to participate in either or both.

Building Controls don’t work because they were never designed. That was the recurring theme, both in and out of sessions, for Tuesday at BuilConn. Speaker after speaker repeated this, both on and off the podium.

Alan Edgar, of the executive board of NBIMS, came to BuilConn yesterday. He attended the Buildings 2.0 sessions in the morning, and was very interested; I did not, as I was talking to an “Experiences with Web Services” session. He seemed to enjoy it, and though it needed to be a perspective in NBIMS. The complement, of course, is that currently, controls are not.

I have long relied on a prop to explain the problems of control systems design. At any moment on the UNC campus, there are numerous capital projects under way. On any day, I can walk into the construction plan room and pull out one of the current plans. I turn to the Mechanical controls page – the one with three panes on it. There is one sheet for each floor, and the three panes are (1) a tag list for the controls, (2) a schematic for the controls and (3) a sequence of operations for the controls.

This is a reliable prop, because, I have never yet had to go to a second building. Flipping through the floors, it become obvious, even to cursory inspection, that these are not right. The tag lists will prove to have been developed for the 4^th floor and cut and paste onto the others, even onto the radically different first floor and lobby. Perhaps the Sequence of Operations will describe some system that is clearly not the one in the schematic. There is always some gross error – I have never had to go to a second building.

But perhaps the University of North Carolina is uniquely cursed? Last night, I talked In conversation with an engineer who was doing a retro-commissioning project at another university. After collecting data on 47 buildings, he began trying to understand the pattern to the building’s consistent bad performance. The common thread? Not one of them had more than a partial design.

There is no excuse for this. A system design for a control system should modeled up front as part of the design, The schematic, the SOO, and the tag list should be 3 views of the underlying model. If a “designed” system does not meet this low bar, that of internal consistency, one has to wonder if it was designed at all.

If we do not design the buildings we send out to be built, we will never get to the next bar of designing them well. If we do not design them well, then LEEDS, Green Buildings, and Zero Carbon Facilities are a sham. If they are a sham, the design profession has been overcharging a lot of people.

Consider that the next time you walk into a “Green” building.