NNSquad - Network Neutrality Squad

NNSquad Home Page

NNSquad Mailing List Information

 


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[ NNSquad ] Re: "How to Patrol Your ISP"


My comments, from my point of view...

> From: Rahul Tongia <tongia@cmu.edu>
> Sorry for joining (responding) late to the discussions.
> I have been toying with the idea of such measurements for a while, and 
> am in the process of proposing something like this to NSF, in fact.
> I don't think that would conflict with the goals discussed here, and 
> could overlap nicely.
> From what I understand, CAIDA focuses on AS (BGP) level issues (KC or 
> others, please correct me)?
> We're talking about something vastly more distributed.

> IMHO, creation of the tools is one thing, but designing a process and 
> system is a lot more complicated:
> 1) What data will it collect - conversely, what data won't it collect?

Definitely a key matter for discussion.

> 2) Is this meant to be end-user voluntary? Why would they participate? 

Utterly voluntary.  Participation would be a way to help make the
Internet a better place, to put it simply.  I also suspect that
there will be many people who will participate out of a latent (or
not so latent) concern about their own ISPs' activities.

> What information would they be willing to share (or not)?

As per (1), but nobody would need to share data beyond what they wished.

3) Privacy issues

Right.  Absolutely crucial.

4) What would we (anyone) be able to do with such data?

My view is that it will be subject to analysis and public reporting,
as appropriate, absent specific restrictions (and I'd very much
like to avoid such restrictions whenever possible).

5) Who would have access to such data?

As I noted in an earlier message, I feel that raw, unverified mass
data from a large-scale data collection network should be kept private
until its veracity can be characterized.  But the ultimate goal
is public disclosure of findings.

6) International and jurisdiction issues

Yes, though since we are dealing with voluntarily provided data,
I believe (and hope) that these will be relatively minimal.

7) Is this only for retail (consumer) connections?

I'd say no.  Consumer, business, government -- anyone who wants
to participate -- ISPs, too.

> 8) Is the intent to only consider the last (ISP) hop, or also consider 
> end-to-end issues, e.g., bottlenecks or gateway issues?  The latter are 
> esp. key for international or remote users, who might actually do "fine" 
> in the first few hops. There are many subtleties when we consider the 
> flip of last mile as first mile (where content lies, which may be either 
> at the edge or in some centralized server).

As far as I'm concerned, end-to-end issues are absolutely critical.

--Lauren--
NNSquad Moderator