Our analysis yields many interesting results concerning cycle
stealing.
While cycle stealing obviously benefits the beneficiary jobs and hurts
the donor jobs, we find that when , cycle stealing is
profitable overall even under significant switching times, as it may
ensure stability of the beneficiary queue. For
, we
define load-regions under which cycle stealing pays. We find that in
general the switching time is only prohibitive when it is large
compared with
. Under zero switching times, cycle stealing
always pays.
Two counterintuitive results are that when , the
mean response time of the beneficiary jobs is surprisingly insensitive to the
switching time, and also insensitive to the variability of the donor
job size distribution. Even when the variability of the donor job
sizes is very high, and donor help thus is very bursty, the
beneficiary jobs still enjoy significant benefits.
Our analysis also allows us to investigate characteristics of
the beneficiary and donor side thresholds, and
,
both with respect to their impact on stability and their impact on
mean response time. With respect to beneficiary stability, we find
that
has no effect, while increasing
increases
the stability region. Donor stability is not affected by either
threshold. With respect to overall mean response time, we find that mean
response time is far more sensitive to changes in
than to
changes in
.
We find the optimal value of
tends
to be well above 1. The reason is that increasing
does not
appreciably diminish beneficiary gain, but it does alleviate donor
pain. We find that the optimal setting of
is an increasing
function of
,
, and switching times. By contrast, we
find that the optimal value of
is often close to 1,
provided
. Increasing
significantly hurts the
donor, although it may provide significant help to the beneficiary if
is high. We find that the optimal
is not a
monotonic function of
, but is an increasing function of
and switching times.