hi Sebastian,
I had read that research report and use as my reference, but I see that research by using simulator only and my research in the real condition. Anyway I also use wireshark to analysis IBR-DTN throughput but the result was not real time condition of course it is different with iperf. I don't know what the iperf can use to analysis IBR-DTN also???? So far I don't have any reference the dtnperf can use on IBR-DTN and I know this software only for DTN2. If you have any reference please share to me.
Best regards
On Fri, May 1, 2015 at 5:00 PM, ibr-dtn-request@ibr.cs.tu-bs.de wrote:
Send Ibr-dtn mailing list submissions to ibr-dtn@ibr.cs.tu-bs.de
To subscribe or unsubscribe via the World Wide Web, visit https://mail.ibr.cs.tu-bs.de/listinfo/ibr-dtn or, via email, send a message with subject or body 'help' to ibr-dtn-request@ibr.cs.tu-bs.de
You can reach the person managing the list at ibr-dtn-owner@ibr.cs.tu-bs.de
When replying, please edit your Subject line so it is more specific than "Re: Contents of Ibr-dtn digest..."
Today's Topics:
- Re: throughput IBR-DTN (Sebastian Schildt)
- Re: Bundle Queueing (Johannes Morgenroth)
Message: 1 Date: Thu, 30 Apr 2015 18:39:09 +0200 From: Sebastian Schildt schildt@ibr.cs.tu-bs.de To: ibr-dtn ibr-dtn@ibr.cs.tu-bs.de Subject: Re: [ibr-dtn] throughput IBR-DTN Message-ID: DC1EE5A8-5993-4341-A1DD-73D541EF28C9@ibr.cs.tu-bs.de Content-Type: text/plain; charset=utf-8
If you want the goodput, that is you want to find out how many user data is transmitted, you can just create some test-data, and send it. For example create a 100MB file and send it 100 times using dtnsend/dtnrecv and then using netcat for TCP/IP and measure the time needed.
However, interpretation of results is not that easy, as for example small files lead to different performance: In the Bundle Protocol, because there is some additional processing overhead per Bundle and also TCP itself will perform differently due to slow-start. For DTN Storage configuration can have a large impact too.
Just look at the (probably confusing) graphs here: http://jenkins.ibr.cs.tu-bs.de/dtnperf/ Form top to bottom every graph uses larger bundle sizes. Inside each graph configuration of storage or security is varied. The point is, on the same machine for very small bundles you get tenths of kilobits or 900 Mbit for large bundles when operating in-memory….
Some information what can be measured, and what things influence performance can be seen in this older technical report: https://www.ibr.cs.tu-bs.de/papers/poettner-tr201108.pdf However the IBR-DTN/DTN-2 and ION versions used in that paper are outdated now, so the actual performance numbers might be different
Sebastian
On 28 Apr 2015, at 02:20, abu fahruddin abu.fahruddin@gmail.com wrote:
Hi all,
I just want to compare throughput IBT-DTN and TCP IP. Could you help me the correct tools. Currently I use wireshark but I
still confiuse to analysis wireshark for live striming data.
-- !! This message is brought to you via the `ibr-dtn' mailing list. !! Please do not reply to this message to unsubscribe. To unsubscribe or
adjust
!! your settings, send a mail message to <
ibr-dtn-request@ibr.cs.tu-bs.de>
!! or look at https://mail.ibr.cs.tu-bs.de/listinfo/ibr-dtn.
Message: 2 Date: Fri, 01 May 2015 09:33:57 +0200 From: Johannes Morgenroth morgenroth@ibr.cs.tu-bs.de To: ibr-dtn@ibr.cs.tu-bs.de Subject: Re: [ibr-dtn] Bundle Queueing Message-ID: 55432C65.6020809@ibr.cs.tu-bs.de Content-Type: text/plain; charset=utf-8
Hello Alex,
there is no 'specific' reason for that. It is more a historical engineering decision. The memory-based storage was the first which got implemented. In the later development, the disk-based variant was implemented to support devices with small amounts of memory. To achieve a high performance, the disk operations are processed within a separate thread. In the memory-based implementation this is not necessary, because there are no long blocking operations.
Kind regards, Johannes
Am 20.04.2015 um 10:24 schrieb Alex van der Linden:
Currently I'm looking into the bundle queuing of IBR-DTN. In the code of IBR-DTN_1.0.1 I've found different queue implementations for HDD (SimpleBundleStorage) and memory (MemoryBundleStorage). i.e. In the SimpleBundleStorage _priority_index queue is encapsulated by the class MetaStorage and in the MemoryBundleStorage it is not. Is there a specific reason for the difference in the context of memory management?
Subject: Digest Footer
-- !! This message is brought to you via the `ibr-dtn' mailing list. !! Please do not reply to this message to unsubscribe. To unsubscribe or adjust !! your settings, send a mail message to ibr-dtn-request@ibr.cs.tu-bs.de !! or look at https://mail.ibr.cs.tu-bs.de/listinfo/ibr-dtn.
End of Ibr-dtn Digest, Vol 58, Issue 1