atomicBIM: Splitting Data to Unleash BIM’s Power

This article by John Tobin addresses the problem of the massive, singular datasets created by current BIM applications. Not only are these getting rapidly unmanageable, they also do not address the need for distributed and ever-larger, greater granularity of data that is going to be vital for the future, collaborative process. As a solution, Tobin proposes the concept of atomicBIM—BIM in small, discrete pieces of data that would provide granularity and rapid access so that subsets of BIM information could be more easily accessed without a massive download.


11 thoughts on “atomicBIM: Splitting Data to Unleash BIM’s Power

  1. I would agree with the concept of small discrete data sets but should we consider who is to deliver what information to the integrated BIM solution. 90-95% of relevant construction and building management data will be defined and delivered by those who are not BIM model users. The CAD BIM is not the centre of the project world but only a different view of the data. Most data will be populated into the central project database or distributed database as alpha numeric data.What we need to understand is the stake holders in the supply chain that deliver information at different times during the project delivery. The nature of the temporal delivery means that different information or changes to information will occur and be implemented by different stakeholders and they may need to take ownership of data delivered by previous stakeholders. This is a complex problem and the CAD/BIM deliverer may not be responsible for the propogation of the data at delivery.

  2. This is a brilliant, forward-looking article. I enjoyed reading it and it makes a lot of sense. However, the commercial implications of this vision could be a serious stumbling block.

    Commercial vendors are still pitching present-day BIM tools and hoping for greater adoption. Only those users who are further along on the BIM adoption curve have started to feel the pain of “model bloating.” In other words, it is not yet a sufficient pain point for commercial vendors to start developing solutions based on the atomicBIM idea. Besides, I seriously doubt if any of these vendors would plunge head-on and support a vision such as this with the IFC being the focal point.

  3. Thanks Mervyn, I think it’s valuable commentary. One point I was trying to emphasize in the article is that authoring will be performed by more than just AEC types – that owners and non-AEC people will also use their own, industry-specific interface to enter relevant data.

    I do believe, however, that the AEC group will initiate the BIM process for the foreseeable future – at least until owners and planners become more sophisticated in creating the BIM “context” before they hire anyone – but that might change quicker that I think. Ultimately, however, the BIM model will indeed be just one of many views of the data.

    Some start-ups such as are already looking at how to filter out the mass of BIM data so only that which owners need is mined.

  4. Richard, Thanks for you comments. I’m glad you enjoyed the article.

    I tried to find the right words in the article to frame how IFC is our best candidate for a BIM atom, without being too optimistic. Of course it’s always possible that a commercial company could come up with an alternative ‘atomic’ file format, but it will be hard to duplicate the excellent schema development of the IFC format.

    I actually do feel that the BIM software market is going to split into two types – authoring and management. To some extent this is already happening. There already are people out there looking seriously at model servers.

    Actually if I were a betting man, I’d put my money on a company like Oracle over traditional CAD companies like Autodesk to be the final BIM master, because it’s really a data management issue, not a design software issue. The commercial payoff for people like Oracle is not the data file format, which is often open standard, but the back-end hardware and software that manages data distribution.

    As a practitioner dealing with multiple large files over a WAN connection, I can see this issue looming large pretty soon, and right now IFC is the closest format. The crucial piece we need is a robust and scalable data server.

    Thanks again.

  5. I think that this is a very good article, Mr. Tobin., and that there are already a few companies that are working on the model that you are talking about.

    I definitely agree with Mervyn in that the end user of the BIM data certainly shouldn’t have to be trained in a 3D modeling environment. Especially in certain places in the public sector. However, the modeling environment is essential for certain complex building and analytical functions of the process.

    I also think that the database would probably be the ideal place to store BIM data that should be shared across communities of practice and even across models (such as geospatial, utility data, emergency planning and response data, etc.) Oracle’s 11g format and safesoftware’s work with saving the IFC into a database seem to really be heading in the right direction.

    My company is focusing a great deal of attention on this Atomic BIM, and/or Big BIM little BIM, or lonely BIM vs. social BIM industry opportunity. Just glad to see other folks are thinking the same way.

    BTW, I don’t know if you have seen what these guys are doing, I think that they are certainly heading in the right direction: They seem to be extracting just the necessary BIM objects from the IFC file format for master planning (and a bunch of other purposes) to integrate, manipulate, and update the original model: extremely powerful, and I think this software and service architecture has the potential to deliver what’s needed to who needs it, enabling collaboration on an unprecedented scale. (They are saving BIMs in MySQL with PHP as the middleware.)

  6. A brief comment in response to John’s thesis that a company like Oracle might emerge as the integrator or final BIM master. A couple of years ago, Oracle announced a Collaborative Building Information Management (CBIM) platform initiative for the AEC industry to do something similar to what John is proposing — see their press release from April 2006:

    But we never heard of this initiative subsequently, and I assumed Oracle lost interest. But yesterday, the news broke that Oracle acquired Primavera: While this has nothing to do with BIM yet, could it be a sign of Oracle making a renewed attempt to get into the AEC space?

    It would be interesting to have the BIM vendors weigh in on John’s ideas of atomicBIM.

  7. To achieve atomic BIM we need a periodic table. In an increasingly complex industry, open standards based tools supporting IFC are the way to make this happen. John’s article on Atomic BIM is a very good explanation of the elements that are needed. Apart from the technology and standards needed, a cultural shift needs to happen in the industry, and that is really up to all of us to make that happen. Standards are going to be in constant development, but there is a good foundation already to work with and build upon. It is possible today to start achieving many of the workflows that John wrote of in the article. It is not 100% done, but neither is the Internet and we use it every day. The important thing is to jump in and drive the change.

    Recent BIMStorms used iPhones to create and access BIM data at the atomic level:

    BIMstorms have included thousands of people, many different stakeholders, from a wide range of the industry to collaborate in real time. This was supported by many software tools, and a BIM model server. The oxygen of the BIMStorm were open standards based IFC models and Open Geospatial Consortium based Geographic Information Systems. We invite the industry to jump in.

  8. I agree with Kimon that many of the elements of atomicBIM are present in some fledgling – or even more mature – form today. As someone who started using BIm software before it was “ready,” I feel now is good time for many of us to look further into IFC software to see what it possible.

    I originally had included a section on Kimon’s OPS in the article as an example of atomicBIM, but for space reasons it was not in the final version. However the remotely accessible, centralized arena of BIM is the kind of interface we’re after, with a back-end for data extraction.

    To Lachmi’s point on Oracle CBIM, the last report I saw which mentioned it was published by a European consortium and it was dated March 2008. But I somehow feel the material was older than that date. It is definitely an intriguing product concept and the thinking is quite sophisticated. The last active usage of it I heard was that a German firm was using it in some test capacity. I don’t think we’ve heard the last of it, but it sure is pretty quiet right now.

  9. Timely article, John, for those of us who have been wrestling with desktop BIM for a while already. And good to see Kimon here as I would agree that OPS is the most promising development to date in expanding BIM past the desktop. A few comments:

    My attempts to use IFCs to exchange data while preserving any degree of high-level model integrity—like importing an ArchiCAD model into Revit—have been an abject failure. I know I’m not alone here. This could be fixed of course. But while the vendors play lip service to IFC advancement, the current software development model does little to encourage it. I hope the demands of those outside the immediate AEC sphere will drive this. But I also see Autodesk moving to own interoperability and exchange like they have for every other facet of the software workflow.

    With regard to the basic concept of a granular data managed in a wholistic manner, this offers the potential for specialized players to enter the market with thin and light tools, like current iPhone apps. The possibility of designing a workflow built of only those modular parts your organization needs is very exciting. But again, it is a direct challenge to the expensive and bloated toolboxes that constitute the status quo.

    I too envision the day when the model moves to the cloud and is available in real-time to all the stakeholders. Your suggestion of an established player like Oracle building the underlying database is intriguing since I agree it will take a company with that kind of muscle and track record with open formats to stand up to the historically (hysterically) closed world of the CAD vendors. It’s a BIM of Dreams—if you build it… If it’s compelling enough… then even Autodesk will have to work with it.

    Lastly, a semantic commentary. As good as atomicBIM sounds, I would argue that MolecularBIM is more appropriate. Atoms are akin to materials like glass, wood, metal. We tend to deal with compounds like doors, windows and appliances.


  10. Ha ha, Geoff. Good point!

    I actually was gravitating towards the term MolecularBIM as I realized I was really mixing an “atom” of geometry with certain “atoms” of data to create various groups…. .. But then I just plain preferred the purity of AtomicBIM and left it. Artistic license, I guess.

    Also, it played better with “Atomic Bomb” which is what I often think BIM is anyway!

  11. Very pertinent article. I find myself in the quandary John presents early on. How the original workflow setup has choked downstream utilization. The adage of having to create content once is destroyed, because it can’t be gotten to in a usable format. I agree with one of John’s comments that there is a need for BIM i-management applications in addition to authoring software.

    I am looking at a project in a procurement focus that has in excess of 1000 linked/referenced drawings in the model. It appears that although the A/E has used a naming convention, the drawings are a composition of 2D and 3D content and include Revit Architecture, Revit Structure, Revit MEP, AutoCAD, and Civil Land Development. They have been ftp, archived, and transferred at least twice and still link to an original central file on the A/E server. Just to find a usable 3D existing topo for construction site utilization, much less any comprehensive building plans to establish construction sequencing, is daunting to say the least.

Comments are closed.