You can see that the 64 solid rocket tubes are actually parallel to the body, but their nozzles are pointed outward sideways from the center of each tube. 56 of the thrusters are high thrust, while 8 of them are lower thrust for more fine adjustments. The entire MHV rotates at 30 revolutions per second, so continued thrust in a single direction is possible.
The Strix anti-tank mortar round uses a similar divert thruster system, with 12 solid rocket waist thrusters.
Well, for future space combat missiles, why not return to this style of solid rocket guidance? Compared to liquid/gas thrusters, I think solid rocket units could have better acceleration, and the delta-v is not necessarily much worse when you consider the 45 degree penalty of thrusting "between" the four cardinal directions.
Instead of a single layer of somewhat flattened nozzles, I imagine two staggered layers of bell nozzles, one slightly angled forward and the other slightly angled rearward so they all point through the center of mass. And each solid rocket chamber could be partitioned into 4 charge pulses. That way, you could have 64 efficiently large nozzles and 256 charges - some bigger and some smaller.
Compared to a liquid/gas thruster system, this is perhaps cheaper to mass produce and less maintenance. It might also be more compact. But I think the really interesting advantage is that it could be used for a gun launched projectile rather than boosted by a rocket. The projectile might have a stealthy flat face. Note that a gun would typically need to lead the target by a lot in space, so this face will not reflect radar/lidar beams directly back to the target. The face might be pointed, say, 30 degrees away from the target.
This could be an extreme challenge to defend against. It might be impossible to target an incoming guided projectile until it lights up a divert thruster. You might be compelled to perform an occasional evasive maneuver simply to force incoming projectiles to divert.
For guidance, a rear facing receiver could be used for command guidance and/or receive navigation beam data. For example, a beam spiral scanning around the target could implicitly guide multiple projectiles by informing of their relative location compared to the target.
More thoughts on implications for space combat maneuvering and engine/weapon layout to come...
@Richard Healy Solid rocket motors could last a long time in space, but of course you'd generally want to avoid too many hot/cold cycles (for anything, really). There are some hypergolic liquid fuels which offer good performance but which are corrosive to fuel tanks. In that case, you typically want to wait to fuel the missile until shortly before use ... inconvenient, to say the least.
Also, some missiles need coolant filled just before use - also inconvenient. This specific missile used liquid helium tanks on the modified F-15 fighter as well as the missile's booster to ultimately supply the MHV's thermal sensor. As you can see, the MHV's sensor system is pretty large compared to the overall spacecraft size. This wouldn't have to be the case with external beam riding guidance, of course.
@Martijn Vos Yeah, that parallel port really says something about how old that thing was!