status

location
Washington, DC
Subscribe to GeoRSS Subscribe to KML


Open Data Standards don’t apply to the Military?

Published in Government  |  13 Comments


USAFA IncompatibleLast night I came upon a new posting at FedBizOps for the US Air Force Academy’s ESRI Software License Renewal. The solicitation is a sole-source justification for license renewal of ESRI software for $25 million USD.

While government procurement makes things like sole-sourcing common as a mechanism to just renew license – it is really the supporting justifications for this sole-sourcing that are disconcerting.

From the solicitiation: “ESRI is the only source that can satisfy the needs of the government for the following reasons.”

  1. The Geospatial Information System section within the 10th Civil Engineering squadron has been using ESRI software since their initial development eight years ago.
  2. All of the past and present mapping and client specification have been developed using ESRI software products.
  3. The Dean of Faculty’s Geography department also uses the ESRI academic site license for teaching all Geographical Information System coursework to the cadets.
  4. Software standardization between the 10th CES, DFEG, and the entire USAFA is extremely critical.
  5. Compatibility allows GIS data sharing between all agencies on the USAFA will continue to support GIS development in the future.
  6. Award of this contract to another contractor would jeopardize the performance of our mission by making all of the existing GIS data non-usable.

Which of those reasons are legitimate to the missions of the government and defense, and which are indicative of a more endemic problem of vendor lockin?

The points of the critical nature of compatibility are very important. Data and information must freely flow between sources, analysts, consumers, observers, and archiving. In addition, there are definitely costs to retraining and maintenance that affects changing or introducing new software.

However, the majority of reasons provided by the solicitation point to legacy decisions, old implementations, academic education of specific vendors (definitely not uncommon), and the bold closing statement that the data from these software packages is not usable in any other tools.

It’s that last particular point that should be the most disturbing to the administration. Apparently all geospatial data being developed and utilized by the USAFA would be unusable without a sole software vendor. This causes concern over broader interoperability with other agencies and organizations, access to important national information, and archivability and retrievability.

The fault here isn’t on ESRI. They offer an interoperability suite that supports OGC, ISO, and other standards that agencies could utilize. The fault lies with the government contracting that justifies this type of reasoning of renewal and continuation because of single-vendor lockin. There is little excuse that open, compatible interfaces should be part of such a large contract.

I wonder when software licenses and interoperability spending will show up in the USASpending IT Dashboard?

Similar Posts


Responses

  1. Eric Wolf says:

    August 12th, 2009 at 10:59 am (#)

    A few points:

    1. There really is no expectation that DoD should be as open as other Federal Government branches. They’ve always acted under significant level of obfuscation. Right or wrong, they are the ones with the guns.

    2. $25M is a pretty small contract in DoD terms, The USAFA will spend probably 10X that much to retrain for another vendors product. The single-vendor status has allowed the USAF and the Academy to do things that no other school could. They don’t teach any GIS classes. ArcGIS use is expected in all classes.

    3. For the DoD, ArcGIS is a platform – not just a software applicaton. ESRI has developed specific extensions for DoD and extensions have been developed internally to DoD that rely on ArcGIS as a platform.

    But I have to agree with you that the vendor lock-in causing such extreme data interoperability and possible data inaccessibility is downright scary – especially considering how important GIS has become to the DoD.

  2. Andrew Turner says:

    August 12th, 2009 at 11:04 am (#)

    Good points Eric – and “open” doesn’t have to mean “public”. It’s obviously imperative that DoD be able to share information internally between appropriate operations and users. If this is at all impeded because of data formats, licenses, or incompatibilities is troubling indeed.

    At the very least I would hope for clauses that state that the data must be at all times available using industry standards for interoperability – and the most appropriate platform used based on capabilities (such as operation specific features and extensions).

  3. Eric Wolf says:

    August 12th, 2009 at 11:07 am (#)

    The DoD has standardized on ArcGIS – which provides considerable capability for interoperability. ESRI’s licensing is surprisingly easy to deal with when you have a broad site license – like what the DoD (and for that matter, the USGS or most universities) enjoy.

  4. jubal harpster says:

    August 12th, 2009 at 11:16 am (#)

    They are also probably using the SDSFIE data model in ArcSDE with a number of complex extensions built around it. Switching to another vendor, say Autodesk for example, would likely not make the data completely unusable but would create an enormous pain. $25M may not be much money for the US military but its a lot for the rest of us. Building military operations around a single vendor is indeed scary.

  5. James Fee says:

    August 12th, 2009 at 12:21 pm (#)

    @Eric

    The problem I have with this is that I’d like to see an open procurement process. I’m fine with proprietary if it meets the needs of warfighters, but I’m just not seeing this here.

    The issue here is implementation, an ESRI system should be as interoperable as any, but in the case of the Air Force, theirs is not. To me it sounds like there should be a solicitation to fix their existing system.

  6. Sean Wohltman says:

    August 12th, 2009 at 12:33 pm (#)

    Great call to the carpet, Andrew.

    Eric, your argument that “$25M is a pretty small contract in DoD terms” is sad. I’m a taxpayer – I don’t want the government spending needlessly because they can or because someone figures “well hell, at least its not another Osprey program!”

    You suggest that they would spend 10X $25M to train on another vendor’s product.

    Really – show me any one of the > 500 Million Google Earth users in the world that spent that much, if anything to “train” on using that software or any of the OpenStreetMap.org contributors, or geocommons contributors or users….

    There is a big difference between teaching GIS and teaching ESRI’s version of GIS – too many enter the field knowing how to push buttons, not understanding the algorithms underneath the tool.

    That leads to bad maps, and I don’t just mean cartographically – I mean bad maps because someone clicked a button in Military Analyst after guessing at parameters and now a decision maker has what looks like a compelling map in front of him to make his decisions off of.

    I’d hope students are taught to appreciate spatial analysis from all sources, not just one vendor’s take on how to solve a particular problem.

    I wonder why a site ELA of this magnitude is even needed given NGA’s mission and substantial investment in ESRI ELAs already. Shouldn’t the students be learning how to call and evaluate geoprocessing results from those big NGA ArcGIS Server implementations?

  7. Brian Levy says:

    August 12th, 2009 at 2:50 pm (#)

    Most of our customers in the DoD (and commercial space) share a similar requirement with ESRI as the backbone/legacy infrastructure. Normally a wrapper of custom applications, processing, and management tools exist that are built on top of this infrastructure year upon year for decades which “justify” #1 and #2 above. It is very difficult to move away from this infrastructure as this is maintained as core to the success of the operation, mission, or business (#5 above). These dependencies also get thrown into the overall budget for subsequent years.

    Recently, I’ve noticed the smaller and mid-size transitioning to open source and data interoperability to save money on infrastructure and do more with applications and problem solving. $25M (over 1 base 2 options) is quite a large contract in the geo market to spend on a single vendor but we’ve seen this before. It’s truly amazing where and how the money falls.

  8. Gregg says:

    August 12th, 2009 at 4:38 pm (#)

    Want to compete ? call John Boylan at Cassidy and Asssociates
    Washington DC for the federal amrketing team

  9. Virginia says:

    August 12th, 2009 at 4:46 pm (#)

    What I have noticed is that DoD gives lipservice to the concept of interoperability. ESRI gives lipservice to open standards and is as proprietary and closed as all get-out. You need to work in a completely ESRI environment for the data to work, which is what this procurement is saying.

    So what the Air Force is saying is that they’re blowing off the requirements of the Clinger-Cohen Act of 1994 to support a completely closed, proprietary set of software (using the proprietary nature of that software as justification to continue using it–go figure!).

    When is the Government going to get serious about forcing vendors to BE OPEN–not just use a very thin, shallow veneer of “open standards” like ESRI does. (Ask anyone how they can reuse ESRI data structures–they’re proprietary and cannot be reused, regardless of whatever people say.)

  10. Virginia says:

    August 12th, 2009 at 4:54 pm (#)

    What I mean about the ESRI data structures is that the ArcSDE metalayer creates the ArcObjects which contain the topological structures that cannot be exported except by ESRI. The structure of the ArcSDE metalayer is not published–so that if you have data in ArcSDE (or data outside of ArcSDE) you cannot fully extract the data from within or incorporate the data from outside EXCEPT THROUGH the metalayer. The personal geodatabase is the alternative concept–also locked down tightly. The “open” alternative is a shapefile–but that does not contain the topological information stored in the ArcSDE metalayer. So are they conformant to OGC standards? Are they interoperable? Are they open? I hardly think so!

  11. Tim Beermann says:

    August 12th, 2009 at 5:34 pm (#)

    A couple of points towards the original article:

    The referenced USAF Academy solicitation is not an actual solicitation for vendors to procure new spatial software. It was written by the 10th Civil Engineer Squadron specifically for the 10th Mission Support Group Contracting Office to explain why they (the contracting and procurement office) should use ESRI to renew the USAF Academy’s academic site license for ESRI products, and not another vendor to get an ESRI academic site license. The simple reason for specifying ESRI is they are the single organization that can offer an academic site license for ESRI products.

    Due to Federal contracting rules and regulations, that explanation is not good enough. You need to have a more compelling reason, which the author of this document is trying to do. All of the stated reasons for going with ESRI where written to satisfy a contracting official. Obviously, what works for contract officials does not work for geospatial professionals.

    A couple of points towards other comments:

    The USAFA is not “procuring” any new software. They are only renewing the annual (academic site) license for what they already use.

    The stated value of $25 million is incorrect (there is absolutely no way that a single base would be allowed to spend that much on software. The USAFA can’t get the $30 million they need to fix a leaking chapel, which is a national icon). I called the 10 CES, and they told me it is actually about $25 k. This is a pretty good deal for the academy to have virtually unlimited access to ESRI software for academic and operational purposes.

    ESRI is not the only spatial software used at the USAFA. Academic and research departments are using Google Earth (universally loved by Air Force members), Virtual Earth, WorldWind, PostreSQL/PostGIS, OpenLayers, GeoServer, and many other package, both professional, free, and/or open source.

    On the operational side of things at the USAFA, ESRI is the GIS mainstay (there is a lot of AutoCAD for engineering work). Switching out ESRI is not pragmatic. Besides the costs associated with purchasing/switching software, training personnel, and the immediate impact on being able to provide support to customers, is the fact that any new software going on to an Air Force network need to go through a rigorous, time consuming, and expensive procedure to be authorized for use before it can be loaded on a single computer.

    Air Force installations have had a standards based program, called GeoBase, which has advocated the use of OGC standards since 2003. In fact, the Clinger-Cohen Act of 1994 was explicitly used to help direct the need for these standards. The central mission statement of GeoBase is “One Base .. One Map”. It was not uncommon 10 years ago for installations to be mapped several times to meet the data requirements for multiple software systems. GeoBase aimed to create an authoritative mapping data source that could be accessed as a vendor neutral service for any system that required a mapping component. A lot of money has been saved with GeoBase mapping procedures.

    HOWEVER

    There is truth to virtually every comment made here in regards to the Air Force, and the DoD as a whole. I could go on for hours as to why this is the case, but the bottom line is that a lot of work needs to take place to break down the data and system silos that exist across the Air Force.

  12. Andrew Turner says:

    August 12th, 2009 at 8:51 pm (#)

    Thanks for the insight and clarification Tim.

    A couple of points I’d like to know more about:

    - If the license really is just $25k, then why does the solicitation still ask for $25m? This isn’t a doubling, or quadrupling, it is 3 orders of magnitude!

    - Are there examples of the various geospatial tools that are being used, how they’re operating together, and how they’re incorporated into the curriculum?

    - I can understand differentiation between actual requirements and solicitation requirements – but the stated reasons aren’t inadequate, they’re strongly indicative of deep problems or expectations. The last point in particular “making all of the existing GIS data non-usable” is not unambiguous.

  13. Tim Beermann says:

    August 13th, 2009 at 2:18 pm (#)

    I believe that the solicitation has a typographical error ($25,000,000.000 as opposed to $25,000.000). Either that or $25 million is the ceiling on what can be applied to the contract vehicle they are using (which has no real bearing on the actual amount used). I am more convinced of the former, since the numbers are so closing related. This is the standard price for ESRI academic licenses, with a couple of examples below:

    http://www.kars.ku.edu/user/schedule.shtml
    https://rcc.sr.unh.edu/ESRI/GeneralInfo.pm/

    The last statement, “making all of the existing GIS data non-usable”, is a bit strong, but is intended to end the internal debate between CE and contracting as to whether or not they should spend time looking for another vendor that can renew the academic site license (no one else can). But even if the USAFA was completely interoperable, they can’t use the data if their database software, regardless of who the vendor is, quits working because the license has expired.

    Your central theme is correct. While interoperability has been preached heavily, certain policy and historical factors have led to a virtual vendor lock-in in most cases.