Cell Phones and Nature
09.01.2011 by Andrew M. Seybold
What happened during the earthquake was that everyone reached for their phones at once. The networks worked perfectly during the aftermath of the quake but they were simply overloaded on both the voice and the data side. Calls could not be made or received, calls were dropped, video taken of damage could not be sent, and SMS messages did not get through.
The East Coast has suffered a double whammy as we all know. First was the 5.8 earthquake followed by Hurricane Irene, which was not as bad as was feared but still bad enough that the damage will take a long time to repair. Both of these events caused problems for the commercial wireless networks but in very different ways, pointing out the major differences between network overload and cell site failures.
In both of these cases there were network issues. During the earthquake the problem was simple: The networks stayed up but they were overloaded and could not process all of the requests for service. This is the same scenario that has been experienced with landline phones for years. Remember how difficult it used to be to get a dial tone on Mother’s Day? Perhaps you remember when after an earthquake in California or during the wildland fires you could not get a call through to your relatives using the wired network?
While the cause of wired and wireless phone system overloads are different, the results are the same. The network is up and running but the number of people trying to make calls simply overwhelms the network. In the case of wired phones, the reason is that after your dedicated line reaches the nearest central office your call is joined with all of the other calls on a cable or microwave link. This link transfers the requests and the calls overloaded the link since all of these systems are built on the premise that not all phone users will want to make a phone call at exactly the same time. Therefore, the wired phone systems were designed to handle a normal, expected traffic load with extra capacity for peak call periods, but they were not designed for times when demand is unusually high. The lines and switches were jammed and people could not get dial tone and had to wait until the demand subsided.
The difference between wired and wireless network overloading is that in the wireless network the overloading happens when too many people are trying to use the network in a small area. Each cell site is typically made up of three sectors, each covering a 120-degree portion of the surrounding area (see diagram below).
This diagram depicts three cell sites with each site divided into three sectors. Each of the sectors has the same capacity as the others.
Each sector can handle a maximum loading within it. For the sake of simplicity, let’s assume that within each sector the maximum number of voice calls that can be handled is 100. A sector’s normal traffic load might be thirty calls at the same time, peaking at sixty calls in a single cell sector during busy periods. Good cellular design dictates that reserve capacity be built into each cell sector so that others entering that sector from another have capacity on the new sector and are not disconnected as they move from sector to sector.
The sector becomes overloaded when demand for service exceeds the maximum number of calls that can be processed in that sector, in this case 100, so if there are 120 people within the sector some will not have network access. The way you gain access to the network is that your device (or the network in the case of an incoming call) sends a request on what is typically called the signally channel. This channel is not only used to request a call but also for the network to track the location of the device so it can be found during an inbound call as well as to facilitate the hand-off to the next sector when the phone is moving. In some networks this signaling channel is also used for SMS traffic, which uses some of the capacity of the signaling channel.
If there are too many devices trying to access the network within a cell sector, the signaling channel becomes overloaded and some customers’ requests will not even reach the network (this is one reason priority access for public safety is not a viable option). So there are two issues, the total number of calls a sector is capable of handling, and the amount of traffic on the signaling channel. Even if more spectrum is allocated to a cell sector, while the number of calls that can be handled by that sector increases, there is still a finite number the sector is capable of processing and completing.
On the data side, even fewer data sessions per sector are normally supported. In normal usage, data bursts to and from the device will permit more customers to make use of the broadband data side of the system. However, if a number of customers are streaming video up or down, the total number of broadband data users is diminished greatly. Even in normal times we have seen the results of cell site sector overloading. AT&T had this type of problem as the iPhone took off a few years ago and many of its customers started using a lot of data services. It is possible that one sector or multiple cell sites are completely overloaded due to demand but calls can still be made and received a few miles away where the demand is less.
What happened during the earthquake was that everyone reached for their phones at once. The networks worked perfectly during the aftermath of the quake but they were simply overloaded on both the voice and the data side. Calls could not be made or received, calls were dropped, video taken of damage could not be sent, and SMS messages did not get through. No matter how much spectrum we have or how robust the commercial operators build these networks, we will have network overloading during major events.
This is not a new problem. You might recall that during the Oklahoma bombing the radio and TV stations were telling people within the affected areas not to use their phones so the commercial systems could be used to augment the public safety channels. During the earthquake, I am not aware of a single cell site failure so the bottom line is that in this instance, the problems experienced were network overloading and this will never be solved no matter how much spectrum we throw at it and no matter how many more cell sites are built. It is not possible for anyone to build a commercial wired or wireless network that will not reach saturation at some point, due to some type of major incident. The same is true, by the way, with the Internet for all of you who plan to rely on it and store all of your data in the cloud.
One advantage to the commercial wireless networks is that the network operators can do some on-the-fly network management. Especially the newer 3G and 4G networks have tools built in that enable pro-active traffic management by changing antenna patterns to shrink the radius of a cell site, to overlap cell sectors in a given area, and to try to balance the load. However, even with all of this new technology there comes a point where a cell sector, and possibly many cell sectors, will be overloaded and this will happen over and over again. It is more severe during an event such as an earthquake because once the event is over, everyone reaches for their phones at once. During a longer incident, say a hurricane, the traffic does not usually peak as quickly and therefore the networks are generally able to handle the additional traffic.
The other advantage to a natural disaster such as a hurricane is that there is advanced warning. In the case of Irene, you can review all of the press releases from the network operators and see that they were all preparing for the worst. They moved equipment around, made sure batteries and generators were operating and had their maximum capacity, and pre-dispatched people and spare parts to areas where the predictions were for the major damage from the storm.
From all of the reports I have seen, the commercial networks, for the most part, withstood what the hurricane threw at them. There were, according to the FCC’s records, a number of outages but they were not network-wide and were limited to cell sites that were damaged or flooded, or where the connection between the site and the network was destroyed. The result was that most of the East Coast was able to use the commercial wireless networks. I have not heard of any network overloads simply because the storm was both predicted and lasted so long in most areas.
The sites that went down went down because of wind damage or flooding, or as mentioned, because the link between the cell site and the network was broken. Today, many cell sites, but not all, have battery back-up and many have both batteries and generators. The number of sites with generators depends in large part on the network. Some networks won’t build a major site without a generator, others build out the network with key sites having both batteries and generators, but some sites only have battery back-up. Some sites are equipped with battery back-up and provisioned so that a portable generator can be driven to the site and connected. Some of the smaller picocell sites don’t have any back-up power at all and if you have a femtocell in your house or office and you lose power, you will probably lose the picocell as well.
It would not matter if every cell site in the United States had massive generators on them. First of all, generators do not operate underwater (one of the big problems during Katrina). Secondly, even if the generator continues to keep the site up, if the link between the site and the network is down then the site, while on and operating, is not functioning and might as well be off. The network operators do the best they can and are very responsive to restoring sites that go down, but sometimes they have to wait for the wired phone company, the fiber company, or even the power company to restore the link to the site before they can bring it back online. As you know, there are still some people without power in various parts of the East Coast and the power companies are working overtime to restore power.
Two different acts of nature caused incidents resulting in two different types of commercial network issues. During the earthquake, the networks stayed up but were overcrowded, a situation that will be repeated regardless of what we do, and the hurricane saw more spot outages due to power and communications links problems. In both cases these types of problems cannot be fixed by an FCC inquiry or a change in the rules, they will continue to happen. There is no such thing as a network that can withstand overcrowding or wind and flooding.
We have all come to rely on our wireless devices, and these incidents underscore our reliance on them. We have to learn to live with the fact that such disruptions will continue to occur. Mother Nature is to blame, not the network operators. In the meantime, what we can learn from this is to not rely 100% on a single form of communications. For my part, I keep four family radio handhelds with good batteries in them for local conditions within my family, and I have been a licensed amateur radio operator since my teens and our local organization provides emergency communications during disasters. We train and we prepare. Our slogan is that “When all else fails there is Amateur radio.” It worked during Katrina, in Haiti, and during the recent tornadoes; I will guarantee you that there were amateur radio operators on the air right after the earthquake and again during the hurricane.
Infrastructure-based communications systems will become overloaded or damaged, that is a fact of life. It is also one reason that the public safety community relies on what is known a simplex, or off-network communications (or peer-to-peer for IT types). If their infrastructure is down they are still able to communicate unit-to-unit no matter where they are, over distances of several miles in most cases. Their networks are designed to support this mode of operation and it is one of the reasons commercial cellular networks are not able to provide the type of communications needed by public safety. They simply cannot afford to be without the ability to communicate so they can continue to be effective during any type of emergency
We rely on our wireless devices but incidents such as these should make us more aware that we might not always be able to communicate with them. It is no one’s fault, it is a fact of life we have to learn to live with—part of smart disaster planning should include at least one other form of communications for times such as these.
Andrew M. Seybold