|
There are just a few things I'd like to
mention to take this thread out of the list.
1. On the number of packets: ULE directly gives the
packet lenght so it's easy to pack/depack the packets
2. There are some MPE implementations that limit
the the number of encapsulated packets to 2 which comes directly from the use of
the PUSI (please don't deluge/flame me with "but our company has another
one")
3. yes the timer is
a big issue and waiting for packing may end up on other impacts at the network
layer (time outs, QoS etc.) this has been known for a while in any encapsulation
and we did simulation codes in ns2 to vary the timer
4. the goal of the study was not to show MPE
(with section packing) was bad and but to show ULE could be used on
the return channel, develop ns2 code for ULE and MPE encapsulation and
test timers etc.
I would like to end the thread here.
Marie-Jose
----- Original Message -----
Sent: Friday, April 16, 2004 5:12
AM
Subject: RE: Adaptation field use in ULE
/ MPG2-TS specification
There is no limitations agreed, But when there is a
time slot to push off your data into the traffic bust, will you be waiting for
more sections to get packed( say you have only one icmp message
right now with 64 bytes) or you just push the section so far you have in
hand.
Otherwise, you have some timer for which you wait for
sections and push the traffic into bust.
Though it is more of implemenation, i believe there
is no standard, how to do section packing as such.
Best
Regards,
William.
NOTE: all sections are for a
particular/single PID.
There is no limitations on the number of sectinos in each MPG for MPE
either. An implementation may of course limit this.
--Tor
"Marie-Jose Montpetit"
<mariejose.montpetit@verizon.net>
Sent by: owner-ip-dvb@erg.abdn.ac.uk
15.04.2004 22:25
|
Please respond
to
ip-dvb@erg.abdn.ac.uk |
|
|
To
| <ip-dvb@erg.abdn.ac.uk>
|
|
cc
|
|
|
Subject
| Re: Adaptation field
use in ULE / MPG2-TS
specification |
|
We used section packing for MPE since it would have not been
fair... However
ULE had not limitations in the number of segments we
could pack.
Marie-Jose
----- Original Message -----
From:
"Patrick Cipiere" <Patrick.Cipiere@udcast.com>
To:
<ip-dvb@erg.abdn.ac.uk>
Sent: Thursday, April 15, 2004 3:24
PM
Subject: Re: Adaptation field use in ULE / MPG2-TS
specification
> Marie-Jose said
> ----------
> So
I do believe we could use ULE in RCS. Actually in a recent study
>
sponsored by ESA we showed that for traffic with a lot of ACK bursts
>
(http: for example) ULE was very efficient because of it's packing
>
capability.
> ----------
>
> Section packing is not
specific to ULE.
> It is a feature shared by the mpeg2 layer: thanks
to the PUSI bit,
> and the section layer: thanks to the embedded
length included in the
> section header.
> It is up to the
encapsulator do decide doing or not section packing.
> Section packing
can be done with any kind of section, and works fine
> with
MPE.
>
> Patrick.
> --
> UDcast: Full IP over
Broadcast Media
>
> Phone: (+33) (0)4 93 00 16 99
>
Mobile: (+33) (0)6 14 21 55 98
> Fax: (+33) (0)4 93 00 16
61
http://www.UDcast.com
>
>
|