Re: [Banana] Introduction of a BANANA-related research project

Hi Lihao,

the link is a very high-level description of the project, and as the 
project has just started there are more open questions than solutions at 
the moment. The cited statement is somehow misleading as you of course 
do not get the low latency for the whole high data rate. E.g. let's 
assume Video over IP with a required data rate of 10 MBit/s and a delay 
< 100ms. You can get this by a DSL (5 MBit/s, ~30ms) and LTE (5 Mbit/s, 
~50ms) bonding. You cannot get this by a terrestrial (1 MBit/s, ~30ms) 
and satellite (10 Mbit/s, ~300ms) bonding.

Yes, we aim for a per-packet based traffic distribution (e.g. due to 
IPsec encrypted packets, where we don't have access to TCP 
headers/flows). Yes, when using a reordering buffer the delay of the 
satellite link (in the order of several hundred milliseconds) will be 
the resulting overall delay of the bonded links, making the low delay of 
a DSL link useless.

I've attached some slides which I'd like to present at the next BANANA 
BoF. Simply speaking, we are very interested in the BANANA approach, and 
we would like to contribute to it in the context of satellite communication.

Best regards,
Jörg

On 06.07.2017 04:04, chenlihao wrote:
> Hi Jörg
> 
> In the link you've given, it says "An illustrative example is the combination of a rural DSL connection with low data rate/low latency and a satellite connection with high data rate but high latency, which results in a user's internet access with high data rate and low latency providing a better Quality of Experience (QoE)."
> 
> My question is: Does the solution use a per-packet basis traffic load balancing? If it does, I will be curious about how the combination of these two links could result in low latency, as in my opinion, it is very difficult.
> 
> Thanks,
> Lihao
> 
> 
>> -----Original Message-----
>> From: Banana [mailto:banana-bounces@ietf.org] On Behalf Of Joerg
>> Deutschmann
>> Sent: Wednesday, July 05, 2017 6:06 PM
>> To: banana@ietf.org
>> Subject: [Banana] Introduction of a BANANA-related research project
>>
>> Dear Banana users and developers,
>>
>> recently a project which aims to combine satellite and terrestrial internet
>> access has been started. A (high-level) description of the project called
>> "Transparent Multichannel IPv6 (TMC-IPv6)" is available
>> here: http://www7.cs.fau.de/en/research/transparent-multichannel-ipv6
>>
>> I came across BANANA and I'm very interested in this WG/BOF. Probably
>> satellite links are also of interest for BANANA solutions.
>>
>> I would like to present the TMC-IPv6 project at the upcoming IETF 99 meeting
>> in Prague. A time slot of at most 5 minutes would be sufficient, I can send slides
>> prior to the meeting. It would be great to get into contact with you (and to
>> attend an IETF meeting, of course).
>>
>> Best regards,
>> Jörg
>>
>> --
>> Computer Science, Chair for Computer Networks and Communication Systems
>> Universität Erlangen-Nürnberg Martensstr. 3, D-91058 Erlangen, Germany
>> e-mail: joerg.deutschmann@fau.de
>> phone:  +49-9131-8527914
>>
>> _______________________________________________
>> Banana mailing list
>> Banana@ietf.org
>> https://www.ietf.org/mailman/listinfo/banana

Re: [Banana] Introduction of a BANANA-related research project

Attachment: ietf99_tmc-ipv6_draft.pdf