In terms of composites versus steel, you are generally looking at:

- More strength/less weight per cross sectional area
- Material is generally more expensive but, at least from a civil standpoint, the savings is generally made up in reduced downtime and cheaper installation costs.
- Composites need to bond to an existing substrate. Steel, on the other hand, can be installed in a void. A compromise is often to apply composites to an under-designed member and install that strengthened member into place.

Again, speaking from the civil/infrastructure world, all the suppliers provide different materials so strengths and systems do vary. If you are just looking to throw something over your patch repair to give it a little non-structural protection, Sika has some lightweight carbon that you can get on the cheap. If you are looking to repair a pressurized pipe or some kind of other structural rehab or NSF 61-approved application, HJ3 has the strongest carbon available. I'd contact the suppliers direct for help on designs, comparisons, and applications.

Kevlar is specific to BASF and, frankly, is more marketing that bite. If you go to any composite design guide (ACI 440 is the standard in concrete design), you'll see aramids (kevlar) always has to take a higher reduction factor than carbon (glass is worse) due to durability restrictions. It may be fine for your application, but a warning on going with a kevlar spec, you're specifying a sole-source supplier and that stuff ain't cheap.
  

That is a very general query! With no background as to the context we can only contribute some basics and see if you can indicate a more specific interest.

Fiber reinforced composites all share an increase in complexity vs. virtually isotropic materials like aluminum and steel. However, they can be oddly simpler, partly because generally they do not suffer so much from corrosion (or rot in the case of wood), which can be a very complicated issue. Because they are generally molded there can often be greater freedom in terms of the shapes made.

For artifical composites that you are probably interested in, like relatively long fiber carbon and aramid fibre reinforced polymers, a major issue is the choice and control of fiber directions. This typically gives an increase in the design costs and usually a shift in the production complexity towards the later stages of the production of a product (generally the rather complicated production of steel or aluminum is very much 'upstream' from the final production, except for cast items). The structural behaviour of fiber composite parts is usually more complicated, with, for instance, higher tensile properties than compressive ones (very much so in the case of Kevlar).

Common fiber/polymer composites all have some increase in weight or cost efficiency vs. common metals, or they would generally not be used (occasionally some other aspect such as esthetics takes precedence). As blaxabbath says, sometimes this can be in the form of better specific strength and stiffness or in something as non-obvious such as installation cost (very significant in terms of, for instance, a road bridge).

Very generally, carbon fiber/polymer composites are usually very roughly equivalent to good aluminum, usually with a benefit in terms of strength and often stiffness. Aramid/polymer composites are a bit less dense and have significantly lower compressive properties but can be very competitive with carbon in tension. Glass/polymer composites are much cheaper than carbon or aramid ones (their specific price can compare with steel), but generally have slightly worse specific properties (except that they are better than an aramid in compression).

All fiber/polymer composites have a low coefficient of thermal expansion compared with aluminum, and they are usually lower than steel. Carbon in particular is basically zero (interestingly both carbon and aramid can have a negative CTE).

The different material and fiber direction options plus the different loading directions make the fiber composite properties so variable it hardly seems worth quoting them, but here are some rough ranges (cost/price is so variable that I cannot even quote possible ranges without a better idea of the end use).

Carbon/polymer:
Density 1.55–1.6 g/cc.
Strength 35–250 ksi.
Stiffness (E) 5–15 Msi (maybe 80+ Msi for exotic fibres).

Aramid/polymer:
Density 1.35–1.4 g/cc.
Strength 15–150 ksi.
Stiffness 3–10 Msi.

Glass/polymer:
Density 1.9–2 g/cc.
Strength 35–150 ksi.
Stiffness 2.5–7.5 Msi.

These would generally be for some sort of continuous fiber material (not injection molded 'short' fibers) and the strength ranges vary a great deal partly because of typical allowances for aerospace damage tolerance reducing the low values.