Performance Analysis for Large IaaS Clouds - Large Scale and Big Data: Processing and Management - page 567

loop i,1,Lh

$(i)_1_$(j) $(i-1)_1_$(j+1) beta_h

end

end
loop i,1,Lh

$(i)_1_$(m-1) $(i)_0_$(m) beta_h

end
loop i,1,m

0_0_$(i) 0_0_$(i-1) i*mu

end
loop j,1,m-1

loop i,0,Lh

$(i)_1_$(j) $(i)_1_$(j-1) j*mu

end

end
loop i,1,Lh

$(i)_0_$(m) $(i-1)_1_$(m-1) m*mu

end
end
end
* hot_full(i) computes value of 1-Ph with `i' hot PMs
func hot_full(i)
if (i==0)

1

else
^(i*ln(sum(p, 0, m-1, prob(hot,$(Lh)_1_$(p);i)) +
prob(hot,$(Lh)_0_$(m);i)))
end
end
func compute_Ph(i)
if(i == 0)

0

else

if(lambda_h(i) > epsilon_zero)

(1-hot_full(i))

else

1

end

end
end
bind P_h compute_Ph(n_h)
expr P_h
end

Observe the use of functions to compute the values of different input parameter
and output measures. Input parameter effective job arrival rate is computed using
the function lambda_ h(.) . Steady-state probability that a PM cannot accept a job
for provisioning is given by the function hot_ full(.) , while P h is computed using the
function compute_ Ph(.) . Further, the value of P block is obtained as an output from
the RPDE submodel, but we use a dummy value to describe the working of the code.