I would first look at my Stack Exchange post How did the universe shift from "dark matter dominated" to "dark energy dominated"? on a related topic. I derive the accelerated expansion of the universe using Newtonian mechanics. The surprising thing is that a constant positive energy or mass equivalent in the vacuum can result in a repulsive force. This basic result carries over with general relativity.
The $120$ orders of magnitude problem has to do with the nature of this vacuum and finding the actual energy density. The cosmological constant $\Lambda~\simeq~10^{-52}cm^{-2}$ is extremely small, which is one reason it took so long to discover the accelerated expansion of the universe. It required observations of type I supernova sufficiently far removed to detect this variation. This problem is a part of the difficulty with renormalization of gravitation.
The difficulty with the renormalization of gravity can be seen in the following argument. The main difficulty lies with loops. We consider the following loop
In this diagram we have vertices $V~\sim~p^2$ and internal lines $\sim~1/p^2$ and the loop with $O~\sim~\int d^4p$. The degree of divergence (the exponent on divergence) gives the $D~=~2V~-~2L~+~4O$ The Euler characteristic for this graph is $1~=~V~-~L~+~O$ or that $2(V~-~L)~=~2~-~2O$ so that $D~=~2(1~+~O)$. This means the divergence increases without limit as the order of the diagram increases.
We might though suppose that the order of these loop diagrams cuts off at one. We then have the integration from $k~=~0$to $k~=~1/\ell_{p}$, for $\ell_p~=~\sqrt{g\hbar/c^3}$ $=~1.6\times 10^{-35}m$ the Planck length. The divergence is then order $4$ or $\simeq~10^{140}m^{-4}$. The expected cosmological constant would then be $\Lambda_p~\simeq~10^{70}m^{-2}$ this is $122$ orders of magnitude larger than the measured cosmological constant.