[governance] Does the Marketplace Want (Need) 4+ Billion New 32-bit Address Spaces ?

[Jim Fleming]

Does the Marketplace Want (Need) 4+ Billion New 32-bit Address Spaces ?

A 32-bit address space can specify over 4 billion end points. If you have a
64 bit address space it does not double the address space (as some clueless
reporters claim) it expands by a factor of 4+ billion. In other words, you
can have 4+ billion times 4+ billion nodes, hosts, end-points, etc.

Adding one more 32-bit address space is easy, NAT does that and people said
it could not be done. You have the inside address space and the outside
address space. Packets are easy to mark as inside or outside. There is
actually an unused (un programmed) bit that can be used for that. One can
mark a packet as using inside addresses or outside addresses. Despite the
lies from THE Big Lie Society about NAT not being useful or workable, the
marketplace seems to like it and has it widely deployed.

Efforts to consider very large address spaces are largely driven from FUD
produced by THE Big Lie Society. The question still remains:  Does the
Marketplace Want (Need) 4+ Billion New 32-bit Address Spaces ?

Now that 24x7 broadband is more widely available and people are paying $30
per month for what they used to pay THE Big Lie Society insiders $3,000 a
month, the picture becomes more clear what people need or want.

Looking at only one market but a large one, the video game market wants
high-speed and LOW LAG. The hop count can kill you in large multi-user video
games. Too many hops produce LAG. You may think a player is in one position
on the the screen but they have moved and there is LAG in finding out that
move. Some games become un-playable. LAG is mostly caused by too many hops.
If you turn up the speed on each link, too much delay in passing the packets
between too many locations can still cause LAG. Some games assume the other
user is directly connected via a wire in the same room.

Despite the problems with LAG and the now and then glitch or outage on
high-speed consumer connections, gamers do manage to push packets to each
other at very high rates. They do this even with some very poorly engineered
protocols and packet formats. Rather than new address space, there is more
and more attention being shifted to better LAN to LAN technology with
Virtual IP and Virtual MAC-level solutions. That does not increase the
address space, but it can create the illusion of more devices, connections,
etc. because the LANs are bridged "virtually" from one location to another
to appear as if it was one location.

In one major corporation's game technology, IP addressing is not really
used, MAC addressing is used. Both the Source and Destination addresses are
set to a token value of 0.0.0.1. It appears as if the IP packet is addressed
to and from the same node. The 48-bit MAC address actually does the
addressing. Some would point out that the 48-bit MAC address is unique (or
supposed to be) and can be used as a unique prefix in an 80 bit field formed
with the 32-bit values. [48+32=80]

In re-looking at the 160 bit storage layout for a packet header, it should
be clear that large addresses will not fit. As it turns out, the features
needed in the gamer market are really more focused on reducing the LAG,
caching the MAC addresses, and bridging the devices on a virtual LAN. Some
of the 160 bits can be better used to deliver those services and that usage
may make more sense than huge addresses that people really do not demand.
What would be the point of having huge addresses if the LAG increases and
the addresses are largely unused ? People have to be careful not to get
caught in the hysteria created by THE Big Lie Society about moving to some
new Internet where they are waiting to tax you for address space that just
slows your system and services down. They just want the tax revenue, they
could not care less whether the large addresses are useful or even routable.

As the LAN to LAN virtual bridge market grows, so will the increase in
traffic CLOSER to where people live and work. When you are closer there are
fewer hops and kids play with their friends they meet in school. Rather than
focusing on large addressing, it may be more prudent to focus on the
multi-homing or packet-exchange with the 4 or 5 dominant local vendors in
both the wireline and wireless markets. When WIMAX arrives, with a 10 mile
radius the size and shape of the digital islands will likely become more
clear.

As people become more interested in the local, high-speed services, their
interest in international low-speed chat may reduce. The interest in photos
and pictures will also likely rise and the interest in text blogs will
likely go down. The assumptions made by early Internet pioneers about the
need for a massive address space may not hold. It may be more important to
make sure that packet nodes are smart and able to compress and coordinate
exchange traffic with modest addressing needs as opposed to interconnecting
billions and billions of dumb-down sessions based on out-dated technology.

People looking at these trade-offs from a "governance" point of view, may
find that they develop a model of a cyberspace to govern that does not match
what the market demands. If people all disconnect, there is no need for
cyberspace governance. Also, they may connect with more simplified
assumptions and not need a next generation load of code bloat to do it. They
write their own code and tune it to their needs.


_______________________________________________
governance mailing list
governance at lists.cpsr.org
https://ssl.cpsr.org/mailman/listinfo/governance