SBInet's caretaker shares frustrations about delays, snafus
But Mark Borkowski remains hopeful that SBInet will be a success
- By Alice Lipowicz
- Feb 17, 2010
As executive director of the Homeland Security Department’s Secure Border Initiative, Mark Borkowski is in the unenviable position of overseeing SBInet, the border protection system that is as noteworthy for technical glitches and schedule delays as it is for its ambitious scope. In addition to fencing and vehicle barriers, the high-tech border surveillance system consists of cameras and other sensors strung on towers to be erected along the border with Mexico. A 28-mile prototype system was completed in 2008, and a 53-mile electronic surveillance system is now under construction in Arizona.
DHS has spent about $700 million to date on SBInet development and construction. In January, DHS Secretary Janet Napolitano requested a reassessment of SBInet, saying there were unacceptable delays. In addition, President Barack Obama is preparing a strategy for immigration reform that includes border controls — all of which has federal authorities wondering if SBInet is up to the task.
Borkowski, a retired Air Force colonel and former director of mission support at the Border Patrol, recently spoke with staff writer Alice Lipowicz about SBInet's latest developments.
FCW: What is the status of the project, and what is causing the latest delays?
Mark Borkowski: The first segment — [23-mile-long] Tucson-1 — is all constructed. We’ve got nine towers, along with the day and night cameras, radars, antennas for the ground sensors, and the communications with the command and control facility. It is all integrated and all networked.
What we have been doing is cleaning up and fixing the technical issues, mostly glitches with the radars and cameras, software bugs and security patching, and some hardware. We have cleaned up the vast majority. It is unfortunate but not unusual that as you do more testing, you find more issues. But they are decreasing in significance.
We have built it, done the fixes, and done a series of updates and corrections. Our goal was to complete the system acceptance testing and turn it over to the Border Patrol by January 2010. Now that has been pushed back to September 2010. That is a pretty significant delay.
Cleaning up additional issues has caused about half the delay. The other half is from a restructuring of the final system acceptance tests.
We initially anticipated about two weeks for the dry runs and two weeks to run the systems acceptance tests. The tests are basically a confirmation that what was built is what we ordered. The tests will occur before we turn it over to the Border Patrol for formal operational testing.
One of the requirements is that the SBInet system detect 70 percent of the intruders within range. To place the towers when SBInet was being designed, we asked: How do the intruders tend to come into this location? What are the paths — by foot and/or by auto? So in the Tucson-1 segment we outlined 55 routes. When we do the system acceptance tests, our plan was to send vehicles and people down those routes and calculate the percentage to see if the 70 percent requirement is met.
We initially planned two weeks for the dry runs and two weeks for the system acceptance tests. The problem is we need a statistically valid calculation to properly determine whether the requirement is met. We consulted with a statistician, who said we have to do the tests eight times on each of the 55 routes. So that is 440 tests.
We decided it is worth restructuring the tests to get a statistically significant number. We need these results because we will be making decisions about deploying this system more broadly in the future.
There was another element, too. The system was designed so that each operator has visibility to the nine towers but controls the sensors on two towers. So for Tucson-1, that is nine towers and five operators.
The Border Patrol has asked us to see if each operator can control three towers so that we would only have three operators for nine towers. It was not part of the design of the system, but the Border Patrol wants to know if it is viable. So we have added that to the system acceptance tests.
So the combination of the two changes significantly extends the schedule of the system acceptance tests by four to five weeks. Those tests will start July 31.
FCW: Was Napolitano’s call for a reassessment triggered by the extension of the testing deadline to September 2010?
Borkowski: Between December 2009 and January 2010, we knew we would have a delay, but we did not know how long it would be.
The secretary is very concerned about it. This project has had a series of delays. SBInet was supposed to be along a good part of the border by now, and it is not.
There is a certain urgency here. The delays have made it more significant. The secretary and the commissioner of Customs and Border Protection have had many discussions about the latest round of delays, which did, in fact, trigger the reassessment.
The question is, is [SBInet] viable? Will the delays go away, or should we give up?
Most of us believe it is viable, but the secretary wants to confirm that. We don’t want to be tilting at windmills.
Even if it is a viable system, it is still late. There is urgency in getting some IT contribution to the border. So even if we decide it is viable, what should we do near term to plug the borders?
We have 40 mobile solutions already deployed. We are looking at other solutions. It could include [unmanned aerial vehicles] or using the existing cameras and towers. The existing camera system is not networked or as powerful as SBInet, but it does give us some capability.
We are looking at more aircraft sensors, unattended ground sensors and other use of ground radars.
FCW: What is the schedule for the reassessment, and what will Napolitano be looking for?
Borkowski: There are three parts. The short-term part will be completed in [a matter of] weeks to look at immediate IT options for critical areas of the border, where there is no time for a comprehensive analysis.
Longer term, the assessment will look at the future of SBInet Block I. That will take about a year.
Block I is the entire Arizona border [with Mexico], which is 370 miles.
In SBInet, we have decision points. We will be completing the Tucson-1 and Ajo-1 segments this year for 53 miles. The next decision point is a formal yes or no on Block I.
So the question in the reassessment is: Do we proceed with Block I as planned, or do we re-evaluate whether that is the right plan? It is possible the department will place a higher priority on sections of the border in California, New Mexico or Texas than on completing the rest of Arizona. Part of the reassessment is whether we need to rethink the plan.
Initially, the plan was to spend $6.7 billion for the southwest border by January 2014, excluding a 500-mile segment of Texas. On the contract with Boeing Co. on SBInet, the contract value to date is about $750 million to $800 million, and we have spent $650 million to $700 million so far.
FCW: In the budget request for fiscal 2011, Napolitano asked for $574 million for SBInet, down from $800 million this year. What was the basis for that reduction? Was it due to the delays in deploying SBInet?
Borkowski: I submitted that budget request last spring. It includes funding for all border infrastructure, including SBInet, the physical fence and vehicle barriers, and roads.
In planning for SBInet, we had advance money for SBInet Block I and Block II in the budget. Some of that money is not needed in 2011. For example, there is an operations and maintenance fund of $30 million for Block I that is not needed because Block I is not built yet. So delays have contributed to reductions in the budget.
FCW: You became executive director of SBI in November 2008, about two years after Boeing was hired. How confident do you feel about Boeing and SBInet at this time?
Borkowski: We have had three or four years of going in the wrong direction. In hindsight, it was a bad direction. But we are turning the ship, although I am not always happy with the pace of turning the ship.
I am not fully happy with Boeing. I would like it to happen faster.
But to mitigate my frustration with Boeing, I recognize that a good part of the responsibility is with the government and with the clarity of the government requirements and management. We contributed to the problems we are experiencing.
We had an urgent need to get something deployed. The strategy was that we would buy something commercial off the shelf, and that was probably good enough. It was a reasonable assumption at the time. But how do you manage a project like that? Even if you think it is easy, you need to manage it like it’s hard.
What happened was, if there was a problem, they fixed the cameras or the radars or whatever device appeared to be causing the problem. But we needed to look at the innards and at the systems and integration. We needed to learn the systems as well as the people who built them.
For example, we had a problem with radars suddenly accelerating. In the past, there would be a workaround to fix it. This time, we had to force the engineers to go into the radars, which were Israeli. Some of the software code was actually written in Hebrew. We had another problem with the radars on the fixed towers. It was not happening on the mobile towers, only the fixed towers. It turned out the problem was the grounding scheme for the fixed towers.
The SBInet system seems to fundamentally work. I have 70 percent to 80 percent confidence; it’s not 100 percent.
We have stopped and fixed everything we have found, but then when we test, we find something new. We are trying to get to a point where the problems are trivial. We seem to be getting close to it.
After several cycles of this, is there another piece we have not yet looked at? I’m trying to get Boeing to go beyond the latest round of problems and to find out, is there anything else? That is what we are trying to do now.
We have done a lot of work to wring this out. The SBInet system that is in place is fundamentally working, but it is a little bit fragile.