[outages] Communication Infrastructure [was: Ok, ok; Jeezus, people... it was a joke.]

[Stephen Wilcox]

On 8 February 2013 03:48, Patrick W. Gilmore <patrick at ianai.net> wrote:

> On Feb 07, 2013, at 21:01 , Stephen Wilcox <steve.wilcox at ixreach.com>
> wrote:
> > On 8 February 2013 01:47, Corey Quinn <corey at sequestered.net> wrote:
> >> On Feb 7, 2013, at 5:41 PM, Stephen Wilcox <steve.wilcox at ixreach.com>
> wrote:
>
> >> No, it qualifies given that almost every consumer (and a fair few
> >> business) facing service out there lets you use either Twitter or
> Facebook
> >> to authenticate; it's in the same bucket as 8.8.8.8 going down.
>  Widespread
> >> service failures for end users result if it happens, and is the kind of
> >> thing that's useful to understand in the context of "why is the helpdesk
> >> phone exploding."
> >
> > The problem here is that 8.8.8.8 as well as many large web sites are
> > distributed. Its legitimately down for you and not for 95% of the rest of
> > its users which may in turn be only a small % of the wider
> telecom/internet
> > user base.
>
> And a cut between L3 & Sprint, or NYC & LON, which you seem to think
> qualify, is still "local" and "may in turn be only a small % of the wider
> telecom/internet user base".
>
> Aren't definitions fun?
>
>
> > Its also clear from the recent posts that the details of how operators
> like
> > Facebook and Google architect their applications is not well understood
> by
> > its users.
>
> Oh, please, the details of how Operators like L3 & COLT operate their
> fiber infrastructure is not understood by L3 & COLT!
>
>
> > Hence we have issues reported which result in large threads which end up
> > being of little use to anyone, even the users of those apps.
> >
> > Signal to Noise = poor
>
> I think the S:N ratio is orthogonal to whether it is FB, DNS, or fiber.
>
>
> > I refer again to the list's already written mandate:  "failures of major
> > communications infrastructure components having significant
> > traffic-carrying capacity"
>
> Assume Akamai has an outage (using an intentionally impossible scenario
> just to prove a point :).  Would that be "communications infrastructure"?
>  Akamai has essentially zero fiber.  How about "significant
> traffic-carrying capacity"?  Akamai has no network.  Is it useful to the
> outages community?
>
> I would argue that "communications infrastructure" can easily encompass
> more than routers, DWDM, and fiber.  I would also argue that things which
> direct enormous amounts of traffic (FB, Netflix, Google, iPlayer, etc.)
> count, even though they do not actually _carry_ any traffic.
>
>
> > This is definitely not encompassing Layer-7. I understand its important
> to
> > a set of users, thats why I suggest there ought to be a place where those
> > users can discuss those issues.
>
> This list isn't for discussion, we have -discuss for that.  (And I wonder
> if this thread should be there?)
>
> This list is for notification.  And I would like to be notified if 8.8.8.8
> or Netflix is down.
>

And I would like to be notified if my local Papa Johns has any specials on
combo pizza and chicken strips...


> The list owners (or a vote of the members? I don't know how this works)
> can decide otherwise.  But your arguments are unpersuasive.
>

I'm just quoting what the mailman page says the list is for.... if that
description is wrong, then I'm on the wrong list.

Steve



>
> OTOH: I don't like seeing long threads of useless info (e.g. "Is FB down?
>  I can't load it." with no other info) any more than the next guy.  But I
> repeat, that's orthogonal to the thing which is down.  We see just as many
> clueless "fiber cut" notices as "google is down" notices, if not more.
>
> --
> TTFN,
> patrick
>
>
> _______________________________________________
> Outages mailing list
> Outages at outages.org
> https://puck.nether.net/mailman/listinfo/outages
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://puck.nether.net/pipermail/outages/attachments/20130208/d3c7e0ce/attachment.htm>