Re Why Russia is Building Its Own Internet

["Dave Farber" <dave () farber net>]

---------- Forwarded message ---------
From: Karl Auerbach <karl () cavebear com>
Date: Thu, Jan 25, 2018 at 3:28 PM
Subject: Re: [IP] Why Russia is Building Its Own Internet
To: <dave () farber net>, ip <ip () listbox com>
CC: <dewayne () warpspeed com>, Steve Goldstein <steve.goldstein () comcast net>


On 1/25/18 8:01 AM, Dave Farber wrote:

[Note:  This item comes from friend Steve Goldstein.  DLH]

Why Russia is Building Its Own Internet
By Tracy Staedter
Jan 17 2018
<
https://spectrum.ieee.org/tech-talk/telecom/internet/could-russia-really-build-its-own-alternate-internet
>

It is not just countries that are slowly evolving to have their own
internets, it is also companies.

Remember how the net was at first just a network - the old ARPAnet kind of
thing - and then we learned to create a "catnet" or "internet" in which we
joined those separate networks into the larger whole that we call "the
internet".

Well, that crank can be turned again so that we have a network of
internets.  And it's happening in various ways and for various - and in
many ways rather compelling - reasons.

Back in 2016 I wrote a note about how I see this happening -
https://www.cavebear.com/cavebear-blog/internet_quo_vadis/

The thesis of that note is that the end-to-end principle at the IP packet
layer is essentially dead, but that that principle has strength today at
the application layer - which means that any internet-of-internet that
gives users the end-to-end experience they want from their favorite apps is
a working internet no matter how inelegant the underlying plumbing may be
to the eyes of old internet techies like many of us here.

We've seen this user interest in application services rather than plumbing
happen already: Both telephone and television traffic, forms once tightly
tied to particular specific underlying physical infrastructures, are now
delivered by a variety of means.  And for the most part, users do not care.

Many today think that this carving up of the internet would occur along
lines scribed into the domain name system or by creating an new DNS root.
Others tend to think of super carrier grade Network Address Translators
(GCNATs), which is a technology that does not adequately scale.  Almost
everyone tends to think in terms of one global public IP address space -
and that is where my thesis diverges.

In the network of internets the address at the application level - which is
the level that users care about - is what it has already become, a world of
application-specific handles ranging from things like URL/URI constructs to
Facebook names and Twitter handles to Google logins.  That opens the door
to a future network of internets in which IP address become mere local
tokens just as MAC addresses really only have meaning within the context of
a single Ethernet-like domain.

My thesis is that today's internet can be evolve into a something that
looks like isolated islands that are connected by well-guarded bridges.
Each island could have its own complete - and I mean 100% complete - IPv4
or IPv6 address space.  Those bridges would be application-specific
application level gateways, with a foot in the IP address space of the two
(or more) islands that it bridges between.

There are lots of compelling reasons for this evolutionary path.  Most have
to do with the channelization of inter-island/inter-internet traffic onto
the narrow bridges.  This is not a new idea - medieval cities often
constructed walls around the city not just for security against attackers
but also to channel commerce and people through well guarded (and taxed)
gates/portals.

On a security basis it means that we don't have IP packet layer pathways
around the bridges.  And the channeling of traffic to the bridges means
that it can be more readily observed, blocked, or modified by
national/corporate security officers.

And those who are concerned with copyright and trademark and the like -
they can have their enforcement engines inserted into the bridge.

We've already taken steps in this direction.  The IPv6 internet is largely
one that is logically distinct from the IPv4 internet even if both are
overlaid onto the same wires - just as IPv4 and Novell IPX co-existed on
the same physical plant in an earlier era.

And although China and Facebook and Verizon and Netflix have not yet carved
themselves off into separate IP based network, the number of technical ties
that keep them from doing that is diminishing.  Eventually the reasons
driving separation will exceed the reasons for remaining and an internet
island will calve off from today's internet.  A bridge will be
constructed.  And most users will not even notice the difference.

By-the-way, DNS does play a technical role in this island-and-bridge
system.  Basically no change is needed to DNS technology or
implementations.  However, each island would have its own DNS root and the
data records under that root would be such that for local resources DNS
would provide much as it does today.  But for resources that exist
elsewhere the DNS resource records would point to the appropriate bridge
(and thus the appropriate application level gateway) to use.  This is not
much different in concept that the idea of an IP packet layer default route
to which packets going off the local network are to be sent.


        --karl--




This message was sent to the list address and trashed, but can be found
online.
<https://www.listbox.com/login/messages/view/20180125152827:4A92DDA0-020E-11E8-BDEE-FA82CB4DEBE6/>



-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/18849915-ae8fa580
Modify Your Subscription: https://www.listbox.com/member/?member_id=18849915&id_secret=18849915-aa268125
Unsubscribe Now: 
https://www.listbox.com/unsubscribe/?member_id=18849915&id_secret=18849915-32545cb4&post_id=20180125163032:F6F94928-0216-11E8-AD09-886FAA5C378C
Powered by Listbox: http://www.listbox.com