1: \begin{abstract}

2: The PageRank algorithm enables to rank the nodes of a network

3: through a specific eigenvector of the Google matrix, using  

4: a damping parameter $\alpha \in ]0,1[$.  

5: Using  extensive

6: numerical simulations of large web networks, with a special accent on

7: British University networks, we determine

8: numerically and analytically the universal features of PageRank

9: vector at its emergence when $\alpha \rightarrow 1$. The whole network

10: can be divided into a core part and a group of invariant subspaces.

11: For $ \alpha \rightarrow 1$ the PageRank converges to a 

12: universal power law distribution on the invariant subspaces

13: whose size distribution also follows a universal power law.

14: The convergence of PageRank at  $ \alpha \rightarrow 1$

15: is controlled by  eigenvalues of the core part of the Google matrix

16: which are extremely close to unity leading to large relaxation 

17: times as for example in  spin glasses.

18: \end{abstract}

19: