Вы находитесь на странице: 1из 13

AVPF vs AVP RTCP Intevals

AVP and AVPF Interop

Whe AVP and AVPF is interoperating one must
choose certain parameters correctly.
In trying to determine the most suitable based
on the timing rules a AVPF timing issue was
found that is problematic for interop.
Based on simulations of the algorithm
There might be bugs in the simulation

AVPF vs AVP Problems
Simulated distributions of
RTCP transmission
Tn is AVPF sending rules
without suppression or
reconsideration, i.e.
AVPF(TN no supp) is with
AVP is RFC 3550 including
When Td calculation
becomes close to, but
below Trr-int AVPF
suppression pushes
transmissions interval
distribution upwards and
uneven in distribution

200 bytes
RR: 500 bps
RS: 1000 bps
Senders: 1
Members: 2
Tmin:5 (AVP
Td << Trr-int
200 bytes
RR: 10000 bps
RS: 1000 bps
Senders: 1
Members: 2
Tmin:5 (AVP
Here the RTCP bandwidth
portion the calculation uses
has been increased from
500 to 10k bps, i.e. 20
Reduces Td
Therefore both Tn and
AVPF without suppression
are all ending up in the
0.25-0.5 s range
As can be seen the AVPF
distribution are mostly
uniform in the range 0.5-
1.5*Trr-int, with a little bit
of tail
However, still longer tail
than AVP
200 bytes
RR: 500 bps
RS: 1000 bps
Senders: 1
Members: 4
Tmin:5 (AVP
Same RR as the problem
slide 500 bps, but now
with 4 members, i.e. 3
receivers sharing RR
As can be seen here only a
little bit of suppression
happens at the 4-7.5 s
range, i.e. 0.5-1.5*Trr-int.
That suppression do show
up as bit of long tail above
12.5 seconds
Question: What gives best interop?
So how does an AVPF user set his parameters to best interoperate from a
perspective of avoiding accidental timeout:
The Regular RTCP transmission distribution is one factor which is depending
Trr-int that sets suppression
The Td deciding values:
RR and RS
Average RTCP packet size
The other is when AVPF times out AVP participants
5*Trr-Int given that Trr-int != 0
Thus we must combine Trr-int large enough so that an AVPF particpant do
not time out AVP participants
At the same time AVPF regular RTCP transmission interval should not be
so long that AVP times out
How many consecutive packet losses are needed before timeout happens?
Td parameterization
When using AVPF one should try to avoid setting
RR and RS so that Td in the given session are close
to Trr-int.
The tail in the AVPF RTCP interval distribution is at its
extreme at 1.5*Trr-int + 1.5*Td/1.21828
Worst case Td = Trr-int: 2.73*Trr-int
Td should preferably be less than 1/4
of Trr-Int
That gives us max tail length of:
1.5*Trr-int + 1.5*Trr-int/(4*1.21828) 1.81 Trr-int
AVPs tail ends at 1.5*Tmin/1.21828 = 1.231*Tmin
Given that Td is less than Tmin
Otherwise replace Tmin with Td
Equalizing the tails
If we would like to have equally values for the
extreme randomization intervals given that Td
is less than Tmin and Trr-int:
Trr-int = 1.231/1.81*Tmin
1.81 (following the Td < 1/4
Trr-int recommendation)
Tmin = 5 s => Trr-int = 3.40 s
This ensures that no AVPF session participant
is more likely to time out than an AVP one.
However, the reverse is not true.
Timing out AVP participants
An AVPF participant will timeout an
participant at 5*Trr-int,
Equivalent of AVP participants using a Trr-int as a
factor for number of intervals before timing out.
This assumes Tmin=5s and Td<Tmin

Finding the optimal both ways
So what is the point of least decrease in timeout robustness
Given that Tmin = 5 s and Td < Tmin
AVPF intervals before AVP timeout(target):
AVPF intervals before AVP timeout(1/4 * Trr-int):
AVPF Intervals before AVP timeout (1/10 * Trr-int):
AVP intervals before AVPF timeout (target):
AVP intervals before AVPF timeout (worst case):

Number of Reporting Intervals
given Trr-int
Based on these findings it appears that the
following recommendations for AVPF vs AVP
interop should be given:
RR and RS sufficiently large that Td becomes less
than of Trr-int
Trr-int should be chosen as 4.0
Future Work?
Do the Suppression algorithms bad behavior
needs to be addressed?