[CII] Advocatus Diaboli

[tvest at eyeconomics.com]

On Dec 2, 2009, at 9:31 AM, bmanning at vacation.karoshi.com wrote:

> On Tue, Dec 01, 2009 at 11:37:16AM -0500, tvest at eyeconomics.com wrote:
>>
>> On Dec 1, 2009, at 11:03 AM, bmanning at vacation.karoshi.com wrote:
>>
>>> On Tue, Dec 01, 2009 at 07:20:55AM -0800, Barry Raveendran Greene
>>> wrote:
>>>>
>>>>
>>>>> 	i challenge your lema that there exists a global Internet that
>>>>> can
>>>>> 	-unilaterally- fail, taking out all communications over IP.
>>>>
>>>> Change the crisis to "major failure of the interconnection
>>>> dependencies of
>>>> the Internet."
>>>>
>>>> It is hard to "take out the Internet." It is feasible to have the
>>>> interconnection dependencies massively disrupted. This disruption
>>>> would
>>>> clear the path to continue with FX's thought experiment.
>>>>
>>>>
>>>
>>> 	ok... willing suspension of disbelief.... for now.
>>> 	I'll note - in passing - that if 99.98%  of the global,
>>> 	interconnection dependencies of the Internet on a global scale,
>>> 	fail - and in the remaining 0.02% of remaining connectivity,
>>> 	I can reach / communicate with everyone I need to - then the
>>> 	Internet is not broken - FOR ME.
>>
>> Hi Bill,
>>
>> Could you clarify your "can reach / communicate with everyone I need
>> to" condition a bit?
>
> 	seems pretty clear to me... :)
>
>>
>> Would it suffice to be "not broken" if you could each / communicate
>> with everyone/everything you've ever needed to up to the moment of
>> that 99.8% failure? Or does your assertion imply that the remaining
>> 0.2% would have to encompass everyone you've needed to reach /
>> communicate with in the past *plus* everyone you personally will need
>> to reach / communicate communicate in the future?
>
> 	two questions there Tom...
>
> 	) I never see/know of the 99.8% failures

Sorry, decimal placement error -- should have been 99.98% / 0.02% --  
i.e., was just trying to parrot your statement.
I wasn't trying to make any kind of point about the plausibility of  
the scenario itself, just about what would be required to satisfy the  
"not broken FOR ME" condition.


> 	) the future may require work.

Let's hope so!
But more seriously, it seems to me that the "net present confidence"  
that one can have about the Internet remaining not-broken-for-me  
following a significant change-event is highly sensitive to (at least)  
three variables:

1. the absolute number and splay of people and things you've  
demonstrably needed to reach/communicate with in the past;
2. the duration of the change-event, i.e., until ex ante conditions  
are restored or everyone accepts the ex post conditions as the new  
status quo;
3. the diff of operating conditions before and after the change-event.

I chimed in because it seemed to me that your original formulation,

not broken for me = I can personally reach/communicate with everyone  
that I need to reach/communicate with

made light of the very real, non-theoretical requirement that this  
condition would need to be sustained *over time* under conditions that  
might be quite unlike those that exist today.

In my last message I emphasized the idea of a long duration event, or  
one that might conceivably represent a permanent shift to a new, less  
attractive status quo. However, if you think about it all of the  
problems that one might associate with that kind of situation differ  
only by degrees from the kind of short-term challenges that (I think)  
the creators of this list originally had in mind. In either case, the  
change in operating conditions that (definitionally) marks such events  
is likely to dramatically affect condition (1, above) -- and in ways  
that     would be difficult if not impossible to anticipate fully in  
advance.

That's why I suggested that e2e is directly relevant to the challenge  
of sustaining CI in times of traumatic change. If the goal of  
emergency planning is to increase net present confidence that such  
moments will less than maximally debilitating, e.g., by taking steps  
to minimize the absolute scope/scale/intensity/duration of the  
associated trauma, then I would hope that the importance of e2e  
continues to receive all due consideration from everyone who's  
involved in emergency planning at any/every level.


>> IMO this distinction is the point at which e2e becomes an unavoidable
>> component of the CI debate.
>>
>>> 	Lets face it - at any given point in time, some parts of the  
>>> Internet
>>> 	are not functionally working/connected  to other parts of the
>>> Internet.
>>> 	Its -always- partially broken.
>>>
>>> 	The critical (imho) components of this best-effort service are:
>>>
>>> 	:: triage - who gets cut off and why
>>> 	:: restoration - who gets added first and why
>>
>> The term "restoration" makes sense given this particular thought
>> experiment, but the corresponding definition you provide sounds like
>> it might cover more than just those being "restored" -- was that
>> intentional?
>
> 	yes

Ok

>> In case it's not obvious, such distinctions could have profound real-
>> world consequences. Suppose, for example, the emergency that we face
>> is not sudden or event-driven ala "outage," but rather cumulative and
>> ecological in origin? What if the only prospects for restoration are
>> equally slow and uncertain? Should triage and restoration rules be
>> different than in cases where a quick fix is more-or-less assured?
>
> 	I suspect so.

Ok. I'd personally be very interested in discussing *how* such triage  
and restoration rules should differ in cases where a quick fix is  
known to be impossible.

If you (or any other list members) have any thoughts on that count  
that you'd be willing to share, I'd love to hear them...


>> Would your answer to the previous question about who and what you
>> personally need to reach / communicate with be different if you knew
>> that your own restoration might be years in coming, or never come at
>> all?
>
> 	yes - socially there is a heirarchy.  family comes first.


I personally understand and share your sentiment here, as I suspect  
most *individuals* do.

That said, I don't think that I'd feel the same way if this were the  
formal emergency management policy position of every (or any) official  
steward of a major piece of global critical infrastructure. "Every man  
for himself" doesn't scale well, especially if the vast majority of  
people rely on a tiny minority of people to maintain the  
infrastructure elements that are "critical" to everyone... the  
elements that make e2e possible, for example...

TV