There are turbos with two 'resistance systems'. The Cateye turbo had both a fan (like a Concept II rowing machine) which gives resistance according to the speed your back wheel is turning at, thus is s very good simulation of the wind resistance of normal cycling. But it also has a electro-magnetic system which can also be switched in and gives a steeped resistance which is equated to increasing the gradient of the road. My specific model gave gradients of 1, 2, 3, 5, 7 and 10 degrees which felt pretty much appropriate.
The gel roller is, as someone else noted, something to reduce the noise of a hard tyre on a metal roller.
A set of rollers should have a flywheel, which causes the drive wheel to remain spinning when you stop pedaling. It is weighted so the drive wheel's speed reduces similar to when the vehicle is rolling along a flat road on a calm day. Inertia simulation.
A set of rollers should have a device ( in Fiona's case, a fan ) which increases the load on the rollers as the drive wheels speed increases. This is called "Road Load Simulation" or RLS in the car testing world. It is a MUST.
A good set of rollers should have some extra braking mechanism to simulate a gradient. This can be a second air fan which is adjustable, swishing water in a barrel, magnets, an Eddy current asborption dynamometer or a DC or AC Dynamometer.
A really good set of rollers will use an AC dynamometer to simulate inertia, RLS and gradient which follows a pre-programmed distance/grade schedule.
But alas, these cost thousands of pounds.
