Wireshark-users: Re: [Wireshark-users] Bandwidth-Delay Product
From: Martin Visser <martinvisser99@xxxxxxxxx>
Date: Sun, 6 Sep 2009 11:23:20 +1000

As you are probably aware BDP is a measure of the capacity of network pipe - or simply put how much data does it take to fill it. For most purposes using the forward delay is sufficient - this is because in most purposes the protocol you are using would normally be tuned to allow you put your data packets end-to-end without any gaps. For a reliable protocol like TCP this implies that you have a large enough window size to allow enough unacknowledged packets to be sent, without having to wait for the ACK from the other end. This would also apply to say using UDP to transport a multimedia stream over RTP - you just need to make sure you have enough of receive side playout buffer.

Now the important thing with BDP is that until the receiver acknowledges that data, the sender has to maintain a copy of it in its send buffers in case it has to resend a lost or corrupted packet. In this case you are interested in the buffer being as big as the BDP (where BDP is the full round-trip-time).

So I think if you are thinking about link capacity or send buffer sizes will depend on how you think about BDP.

Hope that helps (and hopefully someone else might chime in if I have it totally wrong).

Regards, Martin


On Fri, Sep 4, 2009 at 4:20 AM, Gordon Widera <gwidera@xxxxxxxxx> wrote:

The Bandwidth-Delay Product, round tripe time verses one way latency:

BDP(bits) = available_bandwidth (bits/sec) x round_trip_time (sec)

BDP(bits) = available_bandwidth (bits/sec) x one_way_latency (sec)

In my searches, I've seen it calculated both ways. Could someone explain why one would use round trip verses one way and vise versa?



Sent via:    Wireshark-users mailing list <wireshark-users@xxxxxxxxxxxxx>
Archives:    http://www.wireshark.org/lists/wireshark-users
Unsubscribe: https://wireshark.org/mailman/options/wireshark-users