I'll have a go at it. Sure, it sounds a bit SciFi but its basically where Air defense is going or is going to be going in the next decade or two or three.
Nodelessness, networking, mobility, modularity and comonality would be of utmost importance.
While there would, of course, be a chain stretching from high air defense command till the actual radars and shooters - command network (radio lines and other datalinks, various landlines where available etc) would also go between each individual sub unit (vehicle). Basically, each sub system would continuously report about its status to every other sub system in the whole network - like a gigantic peer to peer network.
Sub systems like missile carriers would report stuff like their location, available fuel, any possible damage, elapsed mission time (important because crew may get tired etc), missile status, how many left etc. Gun mounts would publish similar reports - though augumented with any possible sensor data, as they'd most likely have their own sensors for local engagements.
Radar sub units would publish all that data plus precise times of their EM pulses. IR sensors would publish images from their IIR camera with precise timecodes.
Now, while every unit that has its own sensor would be able to act on its own, it would first use the processed data returned to it, rather than act on own data. It would go like this: every unit with a sensor (be it radar, tv, IR or even humint report) would send its more or less raw data gathered from its sensors to various command centres. There would, naturally, have to be many of those, for redundancy reasons. There all the data would be compiled and processed - creating an unified real time image of the defended airspace. That image would then be sent to every sub unit within the network.
In practice, that would mean that even if one sensor is getting some weak on-and-off signal and isnt sure what it is - data sent from the command centre would tell with higher precision what that signal really is. (overlap in sensor coverage is a must, though)
Sensors would come in two flavours - radar sensors and IIR sensors with great optical magnification. IIRs would work like this: There would be a very wide angle lens receptor, scanning for heat signals. Another IIR sensor would then be free to jump from one possible signal source to another, zooming in and determining what it could be. With enough separate IIR sensors in a relatively small space (10-20 km?) it would be possible to triangulate the range of every possible target.
Radars would come in various flavours of course. Accent would be on mobility, though, so very large arrays are a no no. That sort of excludes any very low frequency radar but if possible, low frequency radars would also be used, as low as one could go. Key would be this: Each radar would have its own EM emission detector, and any EM waves that hit it would be sorted out and data would be sent through the network to the command centres. Also, every radar would emit a signal with embedded code. Even the same class of radars, working in same pattern, same wavelengths, pulses, etc - each unit would have its own signature code within the signal.
Data within each of those signals, compared to the data given by the command, which knows where each radar is and when it emitted each pulse, would create a network where each radar not only sends its own signal but receives bounced off signals from all the neighbouring operating radars within range.
Basically, as objects in the sky would fly over the network, or alongside it or whatever, command centres would 'see' them not only from the angle of each individual radar, but also from the angles of all the radars within range simultaneously. If we for example have 4 radars: A, B, C and D covering an area, a plane going through that area would 'send' bounced off radar signals from A, B, C and D radars to every radar. If not networked, those 4 radars would each get its own signal and thats it. This way, each radar would, beside its own signal, get signals from remaining 3 radars. And best of all, those bounced off signals from other radars may get to be even more worthy than his own - as they'd be scattered away signals - which is how a good portion of low RCS works - by scattering most of the radar signal in other directions, away from original radar who emitted those signals.
All of the sensor units would have to be as mobile as possible. That means no trailers - that means no radars that need to be assembled, that means no complicated truck stabilization rods. Basically, each vehicle would have the radar mounted on a self stabilizing gymbal. Only action needed would be folding the array up from the transport position, and even that only for biggest arrays. Goal would be this: to stop the vehicle and start surveying within 30 seconds even for the biggest radars. Same time for packing up and going away.
To lower costs in case of a hit by a weapon homing in on a signal - the radar would basically be just the truck, power generator and array with signal modulator. Any other component for actually analyzing data and sending it further through network would be based in another, nearby vehicle, same from HARMs, etc.
Larger arrays that are by definition less mobile - as they have to be on relatively even ground, would work from predetermined areas. Not spots but large areas, im talking tens of kilometers, where they would move around in random fashion between relatively short emission bursts. Several radars would therefore be leapfrogging each other, making sure there are emissions going on at all times.
Alongside IIR sensors mounted on various SPAAGs and short range SAMs, there would also be jeeps/trucks with dedicated IIR sensors (of greater capability, if feasible). They would have several IIR sensors combined, covering a very wide angle of sky. Basically a half of a sphere. More importantly, they would at all times be looking straight up. Any object in the sky passing above them, if it emits enough heat, might be spotted, if not identified.
While IIR sensors are no good against planes at great ranges of 50 or even 100 km, especially from the front hemisphere, situation changes when youre looking from below, or even slightly from the rear, with the exhaust unobstructed, and you're some 15-25 km away. (as planes do have altitude limits). When it comes to very important spots to protect, one could surround them with these IIR sensors, forming several circles. With 5 km or so between each sensor, with mabye 2 or 3 circles - chances of detecting any kind of aircraft, be it a plane or a missile increase greatly. Not only detecting, but also identifying and locating, as all the sensors share data from various directions and compile it to get a real time image of a moving target.
Now to the guns and missiles. All the missiles would be guided the same way - by a standard datalink interface - where the missile would be getting data where to go, and such data could/would be sent from any sub-unit within range. Remember, those sub-units wouldnt be creating that data, command centres would do that - they would merely be repeating it. When the missile gets to a certain distance to the target area, it would actively search for the target - be it with its own radar or its own IIR sensor. Yes, those two guidance mechanisms would be the only ones employed. Where possible, both would be used on the same missile (SM-2IIIb has that) or/and there would be interchangeable seekers for the same missile so one of each kind (at leasT) would be fired towards the target. Naturally, for very small missiles only IIR, IR or even laser guidance would be used.
Actual missile battery, unless we're talking about short ranged systems - would be blind. As a matter of fact - the concept of sam batteries would get trashed. Instead, there would just be individual sam launching vehicles running around. Sure, they could be grouped in the sense that more of them would be close to each other, but each could operate separetely. They wouldnt have dedicated radars and command posts as there wouldnt be need for those. With every radar and every command center being networked, they would get data from the network. In theory, that means that if we had a 1000 km range missile, and we detect a target 900 km away from the missile through some random sensors, that missile could be fired and fly all the way to the target, as itd receive continuous course updates as it passed over other units under it who would all send the course corrections.
And it doesnt need to be just land launchers. Nature of the missile launcher would be of no importance. It could be from a ship, from a truck or from a plane. Missile would fly, get course updates from all the neighbouring units (be they land based, seaborne or airborne) till it gets to go active.
Very long range, long range and medium range missiles would all share the same final stage - to increase commonality and decrease costs. Difference would be in the length/width of the booster. Since i envision a ramjet final stage, there would have to be a starting rocket booster even for the shortest missile in that family. Still, the missile would be large enough to house a decently sized radar array and a decently sized IIR sensor cooled to low enough temperature to work to the best of its capabilities. Like i said, if possible - both would be present on the missile.
I am guessing such a family would operationally be used anything from 40 or so km to several hundred km of range. (minimal engagement ranges would naturally be less than 40 km) Then there would be another class of missile, rocket powered, covering the 15-40 km gap, something in the amraam class size wise, with interchangable seekers, like MICA has.
Thirdly there would be TOR sized missile, for 2-15 km engagement envelope. Those would be mounted on self sufficient vehicles, that would have own sensors and command. While they would always first look for the networked data, theyd have the ability to operate on their own, if needed. These would come on two kinds of vehicles - very mobile tracked ones for shooting on the go, fighting alongside advancing forces, and somewheat less mobile trucks for defense of various fixed sites or accompanying large radars etc.
Last in the chain would be two systems on two kinds of vehicles (again very mobile tracked ones and bit less mobile truck based ones) for a total of 4 variants. They would have their own sensors but would be networked. Also, i guess its important to say that in addition to the grand network, all the units (this includes large radars, large SAMs etc) would also be able to engage in shorter range inner networks, linking all the systems within certain range. These last two systems would be gun systems - for short range engagement but more importantly for defense against munitions themselves. Naturally they'd be able to shoot down cruise missiles but they would be constructed so they can engage targets like jdams, harms, SDBs etc.
While perhaps just one system would be enough, im not sure which would be better so im including both, in hope each has its own strong points. One would be a 76mm rapid fire very high velocity gun which would shoot guided shells. basically unpowered missiles with fuzes, guided by the laser from the mother vehicle. If some other sort of guidance is cheaper/better, that could be used too, but i went with laser guidance as it seems reliable enough while cheap enough to be used in large quantities.
Other system would be much more down to earth 35 or 40 mm gun (twin guns?) - revolver based for added rate of fire, but not gatling as it decreases reaction time and is very heavy at these calibers. Naturally with some sort of AHEAD system.
Oh and every soldier with a manpad would also have at least a receiver unit, getting filtered real time data from the network so it has better situational awareness. When possible, naturally. If all satellites are down, all comm planes are down, all comm relays of any kind are down - then thered be no network data sent.
Sure, all this is not achievable today, but some day soon it may be. Parts of it already are feasable today, parts already are in use today. It would be very expensive, very high on manpower but also extremely robust and very versatile.
