This paper is a sizing exercise, made before the thrusters in question had been flown on any spacecraft. The moment of inertia can be backed out from the numbers in the article. Given the bang-bang control scheme used in the article, the relation between angular acceleration $\dot \omega$, firing time $t$, and angular change $\theta$ is
$$\frac12\dot\omega\left(\frac t2\right)^2 = \frac12\theta
\quad\Rightarrow\quad\dot\omega = \frac{4\theta}{t^2}$$
The torque is related to angular acceleration via
$$I\dot\omega = \tau
\quad\Rightarrow\quad I = \frac{\tau}{\dot\omega}
= \frac{\tau t^2}{4\theta}$$
With this, I get $998\,\text{kg}\cdot\text{m}^2$ for two 100 μN thrusters firing for 22 minutes, $1031\,\text{kg}\cdot\text{m}^2$ for two 500 μN thrusters firing for 10 minutes, and $1010\,\text{kg}\cdot\text{m}^2$ for two 1 mN thrusters firing for 7 minutes. Call it $1000\,\text{kg}\cdot\text{m}^2$ as a nice round number. (This is a sizing exercise, after all.)
What kind of spacecraft is this? It's not a one meter radius spherical cowcraft. Such a vehicle would have too low a moment of inertia, at most $467\,\text{kg}\cdot\text{m}^2$ if all of the mass was on the surface of the sphere. It's not a cylindrical spacecraft, either, with all the mass on the surface of the cylinder. (Once again, too low of a moment of inertia.)
It appears to be a cubical spacecraft with edges that are two meters long and with all of the mass on four faces of the cube; the top and bottom faces have negligible mass. The axis of rotation is about the axis passing through the middle of the top and bottom faces. The thrusters are at the center of opposite side faces. This gives a moment of inertia of $933\,\text{kg}\cdot\text{m}^2$ about the rotation axis of interest. Since this is a sizing exercise, it makes sense to round this up to $1000\,\text{kg}\cdot\text{m}^2$.
