<s>
Internet	B-Protocol
Mix	I-Protocol
or	O
IMIX	B-Protocol
refers	O
to	O
typical	O
Internet	O
traffic	O
passing	O
some	O
network	B-Device
equipment	I-Device
such	O
as	O
routers	B-Protocol
,	O
switches	O
or	O
firewalls	B-Application
.	O
</s>
<s>
When	O
measuring	O
equipment	O
performance	O
using	O
an	O
IMIX	B-Protocol
of	O
packets	B-Protocol
the	O
performance	O
is	O
assumed	O
to	O
resemble	O
what	O
can	O
be	O
seen	O
in	O
"	O
real-world	O
"	O
conditions	O
.	O
</s>
<s>
Network	B-Device
equipment	I-Device
that	O
performs	O
complex	O
manipulation	O
of	O
packets	B-Protocol
,	O
such	O
as	O
firewalls	B-Application
or	O
VPN	B-Application
routers	B-Protocol
,	O
often	O
have	O
non-wire	O
rate	O
performance	O
and	O
as	O
such	O
it	O
can	O
not	O
deliver	O
wire	O
rate	O
performance	O
on	O
small	O
64	O
byte	O
packets	B-Protocol
,	O
this	O
is	O
easier	O
at	O
1500	O
byte	O
packets	B-Protocol
-	O
but	O
none	O
of	O
these	O
packet	B-Protocol
sizes	O
resemble	O
actual	O
traffic	O
seen	O
on	O
the	O
Internet	O
over	O
some	O
time	O
.	O
</s>
<s>
The	O
IMIX	B-Protocol
traffic	O
profiles	O
are	O
used	O
by	O
network	B-Device
equipment	I-Device
vendors	O
to	O
simulate	O
real-world	O
traffic	O
patterns	O
and	O
packet	B-Protocol
distributions	O
.	O
</s>
<s>
IMIX	B-Protocol
profiles	O
are	O
based	O
on	O
statistical	O
sampling	O
done	O
on	O
Internet	B-Protocol
routers	I-Protocol
,	O
and	O
are	O
published	O
in	O
various	O
levels	O
of	O
granularity	B-Operating_System
,	O
such	O
as	O
"	O
simple	O
"	O
and	O
"	O
complete.	O
"	O
</s>
<s>
Although	O
the	O
actual	O
Internet	O
traffic	O
mix	O
has	O
changed	O
over	O
time	O
,	O
the	O
standardized	O
IMIX	B-Protocol
profiles	O
used	O
for	O
testing	O
have	O
not	O
been	O
updated	O
accordingly	O
because	O
the	O
IMIX	B-Protocol
test	O
results	O
need	O
to	O
be	O
comparable	O
.	O
</s>
<s>
Mix	O
profiles	O
exist	O
for	O
IPv4	B-Protocol
,	O
TCP	B-Protocol
,	O
VPN	B-Application
(	O
IPsec	B-Protocol
)	O
and	O
IPv6	B-Protocol
traffic	O
,	O
distributions	O
are	O
similar	O
but	O
frame	O
sizes	O
vary	O
given	O
the	O
different	O
overhead	O
and	O
upper	O
layer	B-Protocol
limitations	O
on	O
MTU	B-Protocol
.	O
</s>
<s>
Here	O
is	O
an	O
illustration	O
of	O
the	O
Simple	O
IMIX	B-Protocol
,	O
a	O
mix	O
often	O
used	O
by	O
firewall	B-Application
vendors	O
showing	O
IMIX	B-Protocol
throughput	O
performance	O
in	O
their	O
data	O
sheets	O
(	O
along	O
with	O
the	O
"	O
optimal	O
test	O
conditions	O
"	O
throughput	O
performance	O
)	O
:	O
</s>
<s>
Packet	B-Protocol
size	O
(	O
incl	O
.	O
</s>
<s>
Here	O
is	O
an	O
illustration	O
of	O
the	O
"	O
optimal	O
test	O
conditions	O
"	O
IMIX	B-Protocol
used	O
by	O
some	O
firewall	B-Application
vendors	O
showing	O
only	O
theoretical	O
throughput	O
performance	O
in	O
their	O
data	O
sheets	O
:	O
</s>
<s>
Packet	B-Protocol
size	O
(	O
incl	O
.	O
</s>
<s>
Early	O
scholarly	O
studies	O
in	O
2004	O
indicated	O
that	O
TCP	B-Protocol
traffic	O
in	O
particular	O
exhibits	O
a	O
bimodal	O
distribution	O
with	O
spikes	O
around	O
minimum-sized	O
packets	B-Protocol
(	O
less	O
than	O
100	O
bytes	O
)	O
and	O
Ethernet	O
MTU	B-Protocol
(	O
more	O
than	O
1400	O
bytes	O
)	O
.	O
</s>
