Information Technology Reference
In-Depth Information
F
IG
. 10. Cumulative errors vs. cum. bytes transferred for SMU/SEAS.
T
ABLE
IV
D
AILY
E
RROR
R
AT E
(
OR
F
AILURE
R
AT E
)
FOR
SMU/SEAS
Error rate
min
max
mean
std.dev.
rse
2
.
35
×
10
−
6
5
.
30
×
10
−
6
3
.
83
×
10
−
6
9
.
33
×
10
−
7
errors/bytes
0.244
errors/hits
0.0287
0.0466
0.0379
0.00480
0.126
errors/sessions
0.269
0.595
0.463
0.0834
0.180
errors/users
0.304
0.656
0.5103
0.0859
0.168
errors/day
501
1582
1101
312
0.283
than daily error count, which indicates that they provide more consistent and stable
reliability estimates than daily error count alone.
Since individual web failures are directly associated with individual hits, we can
use the Nelson model
[36]
to evaluate the overall web software reliability using fail-
ures and hits for the complete 26 days. This gives us the site software reliability of
R =
0
.
962, or 96.2% of the individual web accesses will be successful. This model
also give us an MTBF
=
26.6 hits, or averaging one error for every 26.6 hits.
