How to ensure authorities can implement the most efficient and cost-effective traffic management and ITS for their city and region is always a major concern for governments, both central and local. And yet many argue that transport authorities’ options are being limited by the lack of interoperability between various suppliers’ equipment and that this can effectively lock them into a single supplier and elevated pricing.
More than a decade ago, a series of common standards and protocols were introduced to address this situation by ensuring a basic degree of interoperability between various suppliers’ equipment.
In the case of traffic light control within the US (and some other regions) this is US National Transportation Communications’ ITS Protocol (NTCIP). It lays down standards for the operation of traffic lights – for instance how long the green period should be – in a series of Management Information Bases (or MIBs).
A prime example, according to Q-Free America’s VP of urban solutions, Tom Stiles, is standard traffic light controller MIBs. These have been adopted throughout the US (and elsewhere) but they are not comprehensive and do not cater for the technical advances that are constantly made to the systems.
“Traffic light controllers have a bunch of settings that an engineer would configure at the roadside box to set, for instance, how long the green, yellow and red periods are for each direction and the order of movements. This may include some logic such as progressive green phases to create a green wave or how to contend with larger traffic volumes.”
Each of these settings is an individual ‘object’ for configuring the device and modern traffic lights have around 2,000 such objects stored in a local database that enables the system to run in a standalone capacity.
Proprietary information
With technical progress moving at such a pace, regulators not only struggle to keep up - but also the standards and protocols they introduce can be less than all-encompassing soon after (or even before) they come into effect. In the case of traffic light controllers, Stiles says NTCIP-compliant MIBs cover only about 15-20% of the device’s communication object definitions: “The remaining MIBs are manufacturer-specific and proprietary.”
He describes the proprietary information as including the object’s operational logic, description of operation and how the object interacts with other objects within the controller. This can include how pedestrians are handled, which light rests at green when the intersection is empty and so on. Other proprietary information covers the protocols used for remote access and configuration of signals from central control rooms and includes the ‘language’ through which devices can communicate with one another.
For ITS to reach its full capability across a wide area, the roadside devices need to reference multiple databases – many of which will be located elsewhere, such as at the next intersection or in the control room.
While each supplier will have a means of remotely controlling networked devices of their own manufacture, unless these protocols are openly available, authorities may be able to achieve that degree of control only through using the OEM’s device and systems. The consequence of this situation, according to Stiles, is that bringing a number of systems together from different manufacturers to achieve a ‘smarter city’ may not be possible.
Q-Free has agreements to share MIBs and protocols with two of the biggest suppliers to the North American market, which Stiles applauds, but adds: “We can’t move to a smarter world if we are stuck with this lock-in from one vendor to another.”
Furthermore, he believes such five- to 10-year ‘lock-ins’ mean there is little or no incentive for competitors to develop new and smarter products because they cannot gain access to existing systems.
He also says such lock-ins impact pricing - and prices can rise by 300% when an agency is effectively restricted to one supplier and a roadside controller. This could cost cities supporting multiple vendors $1,200 – but can cost ‘locked in’ authorities $4,500. “There may be a few additional features but not enough to warrant that level of premium,” he says.
Free availability
In response to this situation, Q-Free has made all its MIBs freely available and launched a campaign (#FREEtheMIBs) to encourage others to do likewise. Aimed at the US market (although similar problems occur elsewhere), it has attracted 25 active advocates to date. According to Stiles, 10 agencies have spoken about supporting it although, at the time of writing, only two have joined: Oregon DoT (which joined immediately) and Utah DoT which got on board a month or so in.
Utah DoT’s traveller information manager Lisa Miller, told ITS International the authority joined the campaign because it believes free MIBs are the right thing to do: “It sits well with Utah DoT’s mission to innovate transportation solutions to strengthen the state’s economy and improve quality of life. And we also want to foster an integrated transportation network.”
She has experienced the impact of what she characterises as ‘jurisdictional boundaries and data ownership’. “We have to facilitate the safest and most reliable transportation network and data jurisdiction or silos hinder our ability to be responsive,” Miller says. “We’ve been working for some time to streamline the data-sharing opportunities and it is really important to have a common language through these MIBs to enable our traffic management systems to communicate.”
She highlights the work of the Automated Traffic Signal Performance metrics project and says Utah works with industry partners to create suitable products. “We want a lot of options when it comes to procurement, not a monopoly situation,” Miller insists.
Her view is that public agencies are paying for the data and consequently it should be open and available for agencies to manipulate in order to optimise the system: “Having only 15-20% of the communication objects within those MIBs open and available really prevents us from having fully interoperable traffic management equipment. It handicaps us in trying to use the best traffic management solutions from different manufacturers.”
Viable business
In regards to a price premium on MIBs-locked products Miller says, “Nobody wins. Nobody will want to buy [that equipment] at an exorbitant price and many authorities’ procurement rules would prohibit the purchase of anything that is two or three times more expensive than a comparable product.”
That’s not to say that vendors cannot have viable businesses. “We always purchase the most appropriate products for the situation and some vendors’ equipment may have functions that better suit our requirements,” she explains. “And the challenges we face are not exactly the same as those in San Antonio, for instance, so we will probably end up buying different equipment. Budget concerns also mean most authorities are using equipment that is past its end of life so there are already interoperability issues. Adding [proprietary] datasets on top of that just adds another layer of complexity for us to deal with.”
Utah DoT uses a variety of datasets and open-source software in mobility and traffic management and Miller says its traffic signal contracts require vendors to share MIBs with the central traffic management system vendor within 30 days of an update. “We always have to ask ourselves, ‘are we getting the best value for the taxpayer and can we grow this system as we need in the future?’ There have been times when we’ve had less ability to manipulate datasets or we don’t own the code,” she continues.
“It’s a lot of money to put into something you may or may not be able to grow. We are trying to set ourselves up for success not only now but well into the future and open data principles are a really good way to ensure it will be easier to do – although it doesn’t solve all the problems.”
Security concerns
Not everybody is convinced of the wisdom of releasing the MIBs; some highlight security worries - including one of the biggest names in the US traffic light sector, McCain (see McCain voices safety and security concerns). Stiles is adamant that such concerns are unfounded, saying: “MIBs simply provide a common communications language. Freeing them has nothing to do with the security equation.”
While acknowledging that security is vital, he describes traffic signal control MIBs as “a dictionary” and points out that the NTCIP protocols used to control the red, yellow, green, flashing red and other phases are currently publicly available on the internet. Which means “a cybercriminal could [already] potentially wreak havoc on a region by forcing all signals to go red or worse, flashing red or dark.”
He says the MIBs that vendors are withholding relate to advanced functionality which improves operational efficiency. Without access to manufacturer-specific MIBs, providing full-range functionality when connecting to a competing vendors’ equipment is, he argues, “impossible”.
Returning to the security topic, Stiles says the NTCIP recognises the need for additional protocols to secure communication between ITS devices and is updating its standards to require support of the SNMP v3 protocols to address this issue.
He highlights what he sees as other, more imminent, security threats including vendors uploading database configuration and other files to and from traffic signal controllers using File Transfer Protocol (FTP) which he describes as “old technology and an extremely insecure one”.
Furthermore, he finds that some users do not change the default credentials on their network and IT equipment and adds: “Security can be managed appropriately by any agency using encryption provided by their IT departments.”
Unintended consequences
Could the law of unintended consequences come into play in another way and put manufacturers releasing their MIBs at a competitive disadvantage?
Stiles says: “It is a risk because now every vendor out there has our protocols and one of them - who was previously considering an exchange of protocols request - has said they will no longer consider sharing with us. This means they can integrate our controllers and could pitch for a system job that may previously have been awarded only to us. But we think it is a risk worth taking, because as traffic engineers we know that we can’t get to a smart city world without interoperability.”
Q-Free made a similar move in adopting the open-source Linux-based operating system for the hardware in its field-based controllers. “This is deployed by every manufacturer now,” Stiles adds. “But in the beginning, we were the only people doing this and a lot of [authorities] didn’t want it because we were the only ones. In the end, because we led the movement, we had the biggest presence.
Releasing the MIBs opens up the market in a similar way, he thinks: “We are leading that effort to change things. This is open and legitimate but it could backfire if all the other suppliers work together and stonewall us – but I don’t see that happening.”
Ultimately this campaign will have been successful when either all the manufacturers release their MIBs or all the DoTs sign up to say they won’t purchase systems which have proprietary MIBs. But Stiles doesn’t expect that to happen: “Some vendors will never cooperate and the US market is so big, there will always be room for them to make money.”
However, the campaign is gaining traction and Stiles expects another manufacturer, or perhaps two, to join this year.
Now there’s a prediction for 2020.
ABOUT THE AUTHOR:
Colin Sowman is a technical journalist with many years’ experience in the transport industry and the ITS sector
Kapsch considers global perspective
Others agree, at least in part, with Q-Free’s analysis. Joerg Rosenbohm, global traffic solutions expert, and his colleagues at Europe-based supplier Kapsch, believe that in NTCIP markets vendor-specific data elements can make the integration of signal controllers difficult. He explains: “If a vendor does not share the deployed MIBs, it impedes the deployment of the best solution for a city.” However, he adds a caveat: “Third-party vendors that do not sell signal controllers in a particular market are normally able to implement agreements to obtain the MIBs.”
Kapsch views the 15-20% figure for NTCIP-standardised MIBs as ‘a bit low for most North American signal controller vendors’ and the increasing use of NTCIP in other countries is viewed as positive. It still finds confusion among authorities about the standard, which was developed for phase-based controllers and feels that developing a NTCIP standard for stage-based signal controllers or adapting the British or South African standards might be a logical step forward.
That said, Rosenbohm points out that there are currently no standardised data elements in NTCIP for some common functions, such as the 1/10 seconds data collections as defined in the Idaho/Perdue algorithm or adaptive signal control.
The suggestion is that authorities draft their purchase specifications to either eliminate the use of proprietary data elements or to require that any such elements become part of the data elements used by the purchaser (effectively public domain). In the latter case, Rosenbohm says the purchaser must use these in future purchase tenders for additional controllers and enforce the use of these now-open data elements to ensure interoperability.
Furthermore, compliance with the details of the communications interface, including the data elements in the specified MIBs, will have to be verified during the testing: “Simply function testing is not sufficient.”
He goes on: “That way, vendors can continue to invent and push the state-of-the-practice forward while ensuring that market competition remains open.”
While not denying that some authorities may be paying more for their controllers, Kapsch believes this is not simply a case of profiteering. “Each non-incumbent supplier evaluates whether to bid, and if they learn that the incumbent vendor uses a lot of proprietary data elements, they might decide not to bid. This reduces the level of competition that would otherwise be available to the purchasing agency.”
While Kapsch has not joined the #FREEtheMIBs campaign (it does not sell signal controllers in North America), internationally the firm is a proponent of open interface standards such as NTCIP and UTC. It also points out that some vendors have developed extensions for UTC systems that have not been made public (and would cost a lot of money to reverse-engineer), which have made authorities dependent on that vendor’s products.
McCain voices safety and security concerns
McCain, one of the biggest vendors in the US, is not only uncomfortable with the #FREEtheMIBs campaign’s aims - but rejects its underlying rationale. In a written response to our questions, it told ITS International: “We do not believe that MIBs are the limitations for a DoT to deploy technology they identified as the best solution or value.”
Furthermore, it went on to sound a note of caution saying “Freeing the MIBs presents real threats to public safety and potential litigation, and we believe that a more thoughtful industry approach on this important safety topic is warranted.”
It explained its concern as follows: “Making MIBs freely available increases risk to the DOTs by introducing liability exposure associated with one supplier misinterpreting another supplier’s MIBs – leading to the potential for significant safety hazards to the traveling public. Agreements between suppliers and agencies offer a path to reducing risk.”
The company stresses that it embraces and participates in all published standards (such as ATC Cabinet, Controller, NTCIP), which it sees as the framework for open procurement options for DoTs. While not commenting on the wider industry, it added: “We have a proven track record of providing our products and services at fair market value.”
In conclusion, McCain said: “At this point in time, we firmly believe that this campaign exposes too much risk to our industry and our DoT partners. We highly encourage all DoTs to think through the potential risk it introduces with respect to safety, cybersecurity and liability.”
