Re: [LOOPS] draft-li-tsvwg-loops-problem-opportunities and use-cases.

[Carsten Bormann <[e-mail]>]

Hi Gorry,

>>>> If LOOPS does little good, due to its cost (additional reverse packets, processing/memory, tunneling if that wasn’t already in use) it is a pessimization; many optimizations have regions of operation in which they don't.
>>> I would like to understand the potential for harm... to.
>> 
>> Pessimization is “harm”.  Not really a lot of harm.
>> 
> On the contrary. Pessimism is important in protecting infrastructure. Please supply data and discuss operational practice.

The tunneling overhead will be on the order of 40 to 60 bytes per packet.
I have made some proposals how to reduce the reverse overhead in early May, we haven’t discussed this to the end, but we are talking about single-digit percentage or less overheads here.  So much for the pessimization (inverse of optimization).

Pessimism is also a necessary component of protocol development.
By choosing reasonable upper bounds for LOOPS behavior (e.g., limiting the retransmission time window to something not much more than a couple RTTs, defining bounds for the FEC/retransmission overhead), we can limit any harm from unforeseen interactions.

> I’ve appreciated the discussion, and I wish this was more clear in the drafts. 

Certainly, more work is needed on the drafts, but that can be done in the WG.

> To me, the scope and intention expressing in  the problem statement is itself problematic to understanding the proposed work, and the impact of this service.

The document that we put most energy into is the charter, and that should be what we discuss today.  The I-Ds are background and illustrations, and we certainly weren't perfect in thinning them in line with the narrowing of the charter to the MVP.  I still think the background/illustration (proof of concept) provided by the I-Ds is useful to evaluate the charter.

Grüße, Carsten