Fixing 911 Overload

Every year, NENA – The National Emergency Number Association – estimates that there are over 240 million calls into 911 call centers (known as Public Safety Answer Points or PSAPs). A common problem among all PSAPs continues to be non-emergency calls arriving into the center on Their specialized 911 trunks that exist specifically for emergency calls. The quantity of these trunks is typically limited in each center and and the amount has been carefully engineered to handle the normal volume of 911 calls from the community served by that agency along with a few spares and diverse CO routing where available. The problem is that when they become flooded with non-emergency traffic, legitimate emergency calls could be blocked.

This design also makes the PSAP more susceptible to SWATTING and DDoS attacks by practically anyone and from anywhere. In some areas, emergency calls into a center that is busy may overflow to an adjacent PSAP. While this seems like a logical idea from a backup perspective, let’s examine the bigger picture here.

Not only does this expand the attack face from a SWATTING and DDoS perspective, it is actually quite useless unless there is a Mutual Aid and MOU agreement in place with the other agency. Without the proper authority and the radio comms infrastructure, as well as access to the CAD, there really isn’t much that those agencies can do other than answer the call and write down the information. They then need to figure out a way to get that incident to the agency that can provide service. Typically, if there is no access to the radio network of the adjacent community, or no visibility to the computer-aided dispatch or CAD system, there is little they can actually do with the information they have. Also, remember that if the call has overflowed to them in the first place, they might not even have a way of reaching the original agency using conventional methods, Where is the problem?

The Core Problem: Dedicated 911 Trunks

The existing 911 network in the US is dependent on specialized CAMA trunks. Queries for location use the Automatic Number Identification (ANI) received with the call to determine the location of the caller, and a direct peer-to-peer relationship exists between the PSAP and the local exchange carrier 911 Central Office using these single-purpose trunks.

Since these trunks are limited in number, when a non-emergency call arrives on an emergency trunk, that line becomes tied up for the duration of that call, even if the resource that handles the call is a non-emergency resource. This creates a traffic engineering problem, because now the number of trunks reserved to handle emergency calls are taking the additional non-emergency traffic, something that totally Skews the Erlang calculations used to engineer the number of circuits.

311 to the rescue?

Many believe, and cities like NYC and Philadelphia have implemented, a localized 311 non-emergency service designed to offload non-emergency calls from 911 Call centers and call takers. The call handling technology used to deliver a 311 service is similar to that in the 911 center. This allows the 311 facility itself to become a natural choice for a disaster recovery location or if a facility is needed to house a temporary relief workforce for the 911 center due to capacity or physical damage.

The fact that a 311 center exists though, does not itself provide a solution to the problem, a little more is required. The root cause of the overload issue noted earlier is that people dial 911 when they should have dialed 311. In other words, their call is arriving on the wrong network, and that wrong network has limited resources from a trunking perspective.

Are rubber bands the fix?

While fixes for critical emergency communications from the public should never use duct tape and rubber bands, “elasticity” does bring significant value. Looking back on legacy trunking, of nearly any kind, there is the limitation of a physical circuit on a pair of copper wires. If I need one, I order one. If I need 10, then I order 10. If I need 10, but I only can get 8, then I am short by 2, and there is not much I can do about it. SIP trunking, on the other hand, is delivered over a data facility. The actual bandwidth on that facility can often be dynamic and Is commonly referred to as “a pipe.”

When thinking about the characteristics of data, often it equates very nicely to water. If I need to deliver more water, I need to get a bigger pipe; if I need to deliver more data, the same concept applies. That being said, an inherent benefit of a “data pipe” is that often the delivery medium is the same regardless of the capacity or size. Now a request to the carrier can turn-up or turn-down service capacities through software or configuration. Because of this, the size of the pipe becomes elastic and flexible to my current trends and needs.

Re-engineering with new capabilities

With this new elasticity capability in our network, let’s re-engineer things a little bit to take advantage of its capabilities.

With all of these circuits moved to intelligent SIP trunking, I now have flexibility and sizing capabilities that allow me to be more dynamic with emergency and non-emergency call center call routing. The initial overall pipe capacity reflects best guess estimations for ALL TRAFFIC, and a new control layer of communication between my premise and the carrier networks exists to communicate any changes Required in the specific trunk route sizing required.

Sending these real-time statistics and state changes to the carrier allows real-time elasticity of IP trunking size to realize the most efficient use of resources. Calls to 911 are automatically flagged as such and routed to the 911 CPE where call takers answer. Similarly, calls to 311 are automatically flagged as such and routed to the 311 CPE, where call-takers also answer those calls. By doing this, the number of simultaneous calls to each type of service is controlled by the CPE.

Should a 911 call end up being a non-emergency situation, and the call gets transferred to the 311 center for assistance, a signal is sent to the carrier to add additional bandwidth to the inbound 911 trunk group to compensate for the non-emergency call.

How bandwidth is calculated and allocated is something that now becomes totally under the control of the receiving agency. For example, let’s go back to our disaster recovery scenario. A significant natural event is impacting the local area. An anticipated  20% increase of 911 call taker staff will require 15 additional call-taker seats. In the 311 Center, 15 seats get flagged as auxiliary 911 positions and get staffed by 911 personnel. An increase in carrier bandwidth allows for the additional call volume expected.

This scenario is just another example of why the nation needs to move to NG 911 quickly. The Legacy 911 network in the US uses analog CAMA trunks that are special-purpose and fixed in their capacity. Increases must be pre-engineered and may take weeks or months to implement or sit idly unused, causing unnecessary charges to stack up to municipalities. The technology to accomplish this architecture already exists in nearly every commercial market in existence today. We should heed the lessons learned over many years and provide the same level of innovation to our most essential call centers, ones that save lives.

Follow me on Twitter @Fletch911
Check out my Blogs on: Fletch.TV

---

© 2020, All Rights Reserved, Mark J. Fletcher, ENP
Reuse and quote permitted with attribution and URL