# Infrastructure design considerations in a war torn world?

Suppose that a company is charged with building a communications network with contemporary levels of technology in a country with lots of perennially war torn rural areas (an ongoing asymmetric, armed insurgency or civil war, not an open near peer conventional war) on a basically Earth-like world, although it has more extremes of weather in the target area like large glaciers, deserts, mountains, difficult to traverse rivers, frequently extreme storms, etc.

What sort of design and maintenance and security considerations could the company adopt to maximize the reliability of this infrastructure? What parts of that would be visually distinctive (i.e. suitable for cover art for a novel, or visually interesting in a graphic novel format)?

• Do you expect the construction crews and finished ground facilities (if any) to be likely targets of fighters or bandits? – user535733 May 22 '18 at 18:42
• What bandwidth must this infrastructure support? – user535733 May 22 '18 at 18:42
• @user535733 In the absence of some measures taken to pacify them or win them over, yes. The company isn't a party to the conflict but the existence of infrastructure facilities the functioning of the regime that the insurgents oppose. I don't know how to quantify bandwidth numerically, but enough to be the main telecommunications service for a country with widespread but not universal access to devices that can use it (maybe 20-30% of the population has access +/-). Feel free to make assumptions and demonstrate how they affect the conclusion. – ohwilleke May 22 '18 at 18:49
• FWIW, this question was inspired by Verizon ads about their network surviving extreme weather events that I've been bombarded with lately, although obvious this question is different. – ohwilleke May 22 '18 at 18:53
• Examples of this already exist (like Somalia). Generally, infrastructure construction/reconstruction awaits the end of the war...as infrastructure is almost always much easier to destroy, and it can be terribly difficult for unemployed civilians to pay a phone bill. – user535733 May 22 '18 at 21:19

I will take Legisey's answer and take it up a notch: aerial & spacial.

Satellite internet access is already a thing, although it has some disavantages:

• Geostationary satellites provide a high bandwidth but with high latency, so real time applications (conversations) are heavily hampered.

• LEO and MEO satellites have lower latency, but lower bandwidth. Also, since they do not keep station you need more of them to keep coverage at all times.

• As with any wireless technology, obstacles (which, depending of the frequency, might include rain) can affect the quality of the signal and lower the available bandwidth.

• Building the satellites and putting them in orbit may be expensive. OTOH, it might be still cheaper than building the infrastructure of a country from scratch, specially if population density is low.

If you go one step down and assume that your rebels do not have means to shoot down high flying airplanes (which usually is a reasonable assumption), there have been several proposals that might be more eye-catching.

In general, the lower you are, the better the latency, and the less atmospheric phenomena affect the signal. You also cover less area and population, which means that the same bandwidth is shared between less users. OTOH, you need more units to cover the country, you need those units to be able to handle the weather and avoid conflicts with other airplanes, and there is more risk of the guerrilla forces getting a way to shoot it down.

1Here there is a better image(third picture in the slide).

• I very much like this solution. Fewer dead contractors, and if the money dries up the drones and blimps can be repurposed. – ohwilleke May 23 '18 at 1:06

Wireless

Because of the ongoing rampant war, building and maintaining cables is too difficult:

• The building teams will get shot while laying the cables
• If you try to use tunnels, for instance across deserts, it will be immensely expensive
• Once the team is gone, the cable will be cut, hacked, sold for scrap, whatever suits the guerilla

So they would use a technology similar to radio, maybe VHF, but deployed as a computer network. Some advantages in the case you describe:

• The technology is not very complicated
• They can build some relay-stations and protect them much more easily than a cable
• They can mount relay antennas on trucks and move them before the storm
• They can build the relay station in a safe place, then drop it by helicopter, along with some guys/tank/machine guns to protect it
• They have instantaneous near global WiFi access
• They also can use balloons or drones to cover some areas.
• "They can mount antenna on trucks" I like the idea of a mobile infrastructure, whether on trucks or some other vehicle, a lot. – ohwilleke May 22 '18 at 19:04
• The "antenna on a truck" will probably be too low except in ideal conditions (e.g. you can move the truck on top of the only hill in the area). You need the antenna to be as high as possible, and that would make the truck very unstable and sensible to wind. Some sort of installation would be needed for it to work (of course, you may use the truck as the base of the installation, but then it would not be movable until you dismantle the installation). – SJuan76 May 22 '18 at 20:13
• See, for example, Somalia. – Mark May 22 '18 at 21:55

## Redundant Everything

Redundant radio towers, redundant transmission points, redundant paths...The design of the original requirements for the Internet by the US military offer a really good insight into the design requirements for this system. In fact, the requirements are exactly the same except for having to resist insurgency.

In order maintain service, the network must have multiple and redundant paths to move traffic. This requires redundant hardware and network topologies that support multiple transmission routes. The company will need to develop new network protocols to handle the needs of this all-wireless communications network.

## Costs

This is your primary concern. Building huge, heavily protected infrastructure takes a long time and is costly to maintain. Alternatively, deploying small, cheap, moderate range transmission units makes the network as a whole highly resistant to attack. Using small nodes also shifts the burden of attack from the government/company to the insurgents.

With a few small towers, the company can afford to be wrong only once or twice. The insurgents only need to be right once or twice. However, with small towers, the company can afford to be wrong a lot while the insurgents must be constantly killing new towers. You want to make this as expensive as possible for the insurgents.

Wired connections also represent the same kind of high value target that big static installations do. They must cover long distances and are very hard to defend. Wireless connections don't suffer from this problem.

• Self-powered by solar or wind. They should not require power to be piped in to them.
• Ultimately unrepairable. Think the worst kind of iPhone irrepairability. These should be cheap enough that if they break, just drop another one. Also, if they are too difficult to break down, the insurgency won't be able to repurpose the parts.
• Air droppable. This constrains the weight (and therefor cost) but also means that should the company need signal in a new area, only a few flights are needed before a strong network has been established. There's also a huge demoralizing aspect to this. Should the insurgents destroy a few towers, more can be dropped the next night. It's very discouraging to always be fighting an enemy that just doesn't stay dead.
• Self-installing. Once on the ground, the unit should be able to install itself and connect to the rest of the network. If it's the first node in a new area, it should behave appropriately.

## Nature of the Network

The network should be a mesh network to be the most resilient. However, there are scaling problems with this approach so some kind of tree or star topology would be most efficient. An added benefit to each node would be the ability to detect if the network is under attack then switch into mesh mode to ensure maximum resiliency.

## Ideal Circumstances

The network is so hard to kill that the insurgents start to use the network instead of trying to destroy it. This leads to all kinds of benefits.

• You could make your nodes printable. I assume your design is already wireless. – Henning M. May 22 '18 at 22:10
• @HenningM. Thanks for pointing that out. Yes, my solution is wireless too. – Green May 23 '18 at 0:44

In the real world this sort of thing normally falls into "write-off before starting" territory, the danger to the installation teams and the property loses are just too high for there to be an acceptable pay off at the other end. If I had to had to make it work I'd use tanks, the infrastructure units are assembled in "safe zones" far away from the conflict, preferably another country, they're armoured like main battle tanks, or maybe even dreadnoughts. These units are then transported in and drop on site, they're anchoured using a single large auger pile that can be driven in moments once the unit is on site from within the unit. Visually it looks like a solar powered bunker, because it is.

Also you build them using the most powerful signal gear you can run so that you decrease the "minimum necessary installations" number then put up several times that number because you absolutely know you're going to have downed units at times so you need a lot of redundancy, I'd say five fold redundancy should work, maybe more if your gear becomes a priority target.

• You can assume that someone (perhaps a third party international supporter of the regime) is willing to pay big \$ to make this happen. It doesn't have to be profitable based upon consumer charges alone while the conflict is ongoing (although if peace arrived, it would be nice if it could be self-supporting at that point on an operating cost basis). – ohwilleke May 22 '18 at 19:02
• @ohwilleke Yup my answer stands unchanged. – Ash May 22 '18 at 19:07

I thought of two options and only put the first down thinking someone else would get the other end but they haven't so:

Saturation

Build small, but rugged, relatively low signal power/range units. Build a couple of million of them. Drop them from cargo planes across the country during the wet season, so they self-bury, night ops will minimise civilian casualties. These units are the size of a baseball and have an individual range of 200m or so when buried in fields and irrigation ditches, they are pre-programmed to mesh into a comprehensive network. There are far too many to systematically remove them all from any large, and they're easy to replace if someone does. Individual units only have a battery life of 18-24 months but if you put out a bounty on units damaged or with dead batteries you can refurbish them and reseed spotty areas as needed at reduced cost. Signal life in a given area can be extended further if the mesh is programmed to maintain only the minimum units needed for coverage and put others into a standby mode awaiting a need.