IPNL: A NAT-Extended Internet Architecture

Motivation: This paper attempts to solve the internet address depletion
problem by proposing a NAT extended architecture in order to keep the
benefits of NAT while getting rid of the limitations.

Basically IPNL scheme is an extension to the NAT scheme. The major
attributes of IPNL are as follows.
1. It is a NAT extended architecture.
2. It uses Fully Qualified Domain names as end-to-end host identifier in
packets.
3. It extends the IP address space such that the globally unique IP address
space
forms the high order part of the IPNL address and the private address forms
the
low order address space.

IPNL introduces another layer between IP and TCP/UDP and as a result it can
reuse the IP infrastructure for routing. IPNL addresses are 10 bytes long
consisting of a MRIP, a realm number, and an end host IP. When a connection
is initiated, the client only has the knowledge of the FQDNs of the remote
host and itself as well as its own end host IP. The rest of the information
is obtained through DNS lookup and resolution along the path.

It helps to solve some of the problems like IP address depletion and site
isolation however it is not without limitations. One can spoof FQDN and thus
the server needs to do DNS lookup when responding. It is too expensive for
heavily loaded servers and can result in DoS attacks. Thus, hijack
resistance and neighbor pinging seems to be a problem.

The authors realized these limitations coupled with standardization problems
which might suggest that IPv6 maybe the way to go and using IPNL to draw
upon ideas wherever found appropriate to do so.
Received on Mon Dec 04 2006 - 04:19:58 EST