In many fields of applied physics, the phenomenology of the space-time phen
omena to be understood tin general for prediction purposes) may be describe
d in the following most simple way: events with random common positive ampl
itude occur randomly in time according to a continuous time random walk (CT
RW) model; the prerequisite is therefore a statistical model for both the a
mplitude and inter-arrival times between events, here assumed mutually inde
pendent. Special attention is paid here to CTRW for which both amplitude an
d holding lime have infinite mean value (the extreme and rare hypothesis).
Such processes and their limiting version arise in particular as inverses o
f processes with stationary independent increments of special interest (chi
efly related to the Levy stable subordinator).
Among other related models, we investigate here some properties of this CTR
W in situations where the occurrence of events is modelled by a discrete in
verse-linnik process which shares the rare event hypothesis; this class der
ives (statistically) its importance from its close relationship to many oth
er meaningful processes such as the Levy, gamma and Mittag-Leffler ones. Ph
ysically, Linnik and inverse-linnik processes appear as a recurrent paradig
m in relaxation theory of condensed matter. The limit laws for cumulative L
innik sequences and their time to failure are finally discussed.