<html>
<body>
At 14:07 08/08/2013, Carolina wrote:<br>
<blockquote type=cite class=cite cite="">One of the business associations
made a very compelling presentation on the negative effects of Internet
Balkanization, presenting its points on economic terms and how any
regulation regarding local servers and local data storage would increase
the country cost and make a country even less competitive. It followed to
present comparative country costs regarding storage and data centers.
Pretty compelling for the politicians in the
audience.</blockquote><br><br>
Carolina,<br><br>
There is a general business fear I explained in another mail. This fear
is the same as in the economy. Its root is that the cyberspace clearly
shows that the actual reality is different from the business models that
they learned and that they use and teach. Therefore, some want to
convince themselves and everyone else that the reality can be forced to
follow their model and its architectures, and that at the end of the day
reality will be curbed to their will (along a normal probaility
distribution curb).<br><br>
To do that, they are trying to restrict the architectonic debate within a
closure between two centralized architectures: <br>
- "RA" (reference architecture):
<a href="http://en.wikipedia.org/wiki/Reference_architecture">
http://en.wikipedia.org/wiki/Reference_architecture</a><br>
- "SOA (called SOA – service oriented architecture):
http://en.wikipedia.org/wiki/Service-oriented_architecture.<br><br>
The closure is the cloud. Whatever the solution is to be it MUST fit into
the cloud, as the Internet MUST fit into the DNS root. Otherwise, the
digital dominance would be thereby diluted (what is qualified as
"balkanization" [try to find somewhere an architectural
definition of “internet balkanization”: this is a joke because a network
is made of layers and multiple usages]). This supposed
"balkanization" is documented to be expensive. This is correct,
but not because other use models would be more costly, but rather because
(however cheaper, more secure and surer) they may be costly to implement
within the constraints of the imposed architectures. What has a cost is
an imposed rigidity to flexible and diversified uses.<br><br>
The erroneous trick is the same as with subprime loans. Subprime loans
were OK, if the risk had been distributed to everyone. However, subprime
loans were covered by an insurance policy and mainly large banks bought
subprime loans as a secure investment. It is in this way that they became
too big to fail. The cloud and services are the same. SOA is the good
solution if it is distributed (i.e. memory is at each other person's
place – what one calls the “haze”). It is not when the cloud is owned by
a few biggies. What happens then when a biggie fails or leaks? One can
lend money to banks, but one cannot restore lost confidentiality to
clouds. Owning others' memory becomes the proprietary danger. This is why
we need a "free memory control" (like in FLOSS). This is
something that other network architectures can permit. This is what
intelligent use leads to. Because possibilities follows a power law
distribution cur).<br><br>
However, in that intelligent, efficient and resilent case, please
remember that reality will forget about network control, vendor lock-in,
and easy network tapping.<br>
<a name="_GoBack"></a>These are the stakes.<br>
jfc<br>
</body>
</html>