Lesson 2: Full-availability group.

Full-availability group

The full-availability group is a discrete link model that uses complete sharing policy [COST96]. This system is an example of state-independent system in which the passage between two adjacent states of the process associated with a given-class stream does not depend on the number of busy bandwidth units in the system.

Thus, the probabilities of passing si(n) are equal to one for all states and consequently, Equation (10) is reduced to the Kaufman-Roberts recursion [Kauf81, Robe81]:
          sum M
nP (n) =     aitiP (n - ti) .

Formula (16) determines the occupancy distribution in the full-availability group with different multi-rate traffic streams.

The blocking state in the full-availability group for class i calls occurs only when the group has less than ti free BBU's required for setting up a connection. The blocking probability for the class i stream in this case can be expressed by the formula:
          sum V
b(i) =          P(n).
      n=V -t+1