New Internet Architecture

New Internet architectures
Shpëtim Latifi
Purdue University
shlatifi@iupui.edu

Abstract – The original Internet architecture, developed
since over 40 years ago has been evolving ever since,
and as a result, some issues and problems that could
have not been foreseen or predicted a few decades ago
are now appearing; some principles like end-to-end
connectivity have been violated, and inthe meantime
new issues are emerging
- identity and
authentication, security, mobility, energy efficiency, etc.
While recently there are calls for changes in the
current Internet architecture, we stress the hidden
risks and additional issues related to these ideas,
including but not limited to network complexity,
concerns about privacy, lack of guarantees for security,
anddiscouragement for Internet access to ordinary
people, and discouragement for developing new
applications for the developers. This paper/study
instead, maintains that the approach of patching the
current architecture - firewalls, proxies, and other
middle boxes are helpful and do not increase the
complexity, thus do not endanger the evolvability and
simplicity of the network. We believe that by betterand
more global laws the security issues will be minimized this is what even the most advanced ideas for new
architectures call for, since there are no guarantees for
absolute security. The integration with cell phone
networks and services can solve majority of existing
issues with mobility, and energy efficiency. Also models
for naming, multicast and anycast are also proposed,
and as theneed for mobility is growing, together with
the need for better identification of data and services
instead of machines, these models will emerge slowly.
This industry is in the beginning of its era, and there
are some great solutions on issues related to mobile
support to smartphone devices, and we believe this
field is moving in the right direction, thus rejecting the
need for majorchanges in the current architecture.
1. Introduction

The reasons for having the current Internet architecture as
such (and the way how it was developed) are very strong.

The historical perspective of over four decades of
evolution of this architecture is tremendously rich in
research and innovations. While most of the proposed
technologies over the years have failed and very few have
beenadopted, one thing is true – that is the Internet
architecture is prone to change very slowly. Some of the
arguments for this claim will be stressed shortly. While
contrasting arguments for a need for quick and
comprehensive change in the architecture (i.e. start from
scratch, revolution) vs. arguments for slow change
(evolution), one needs to stress the basic principles on
which thecurrent Internet was developed, and well as
stress the major reasons why the current reality requires a
different view, and poses the needs for a possibly new
architecture.
This paper barely proposes an innovative idea on the future
internet architecture, and merely tries to summarize the
differences between the requirements for the Internet of 40
years ago and the nowadays Internet; theexperience during
this time that had led to the need for making a fundamental
change in the current architecture, as well as discuss some
of the already proposed solutions to some of the issues.
While there are sound reasons for a call about a new
architecture that would address current issues with the
Internet, two things must be considered in order to be able
to say that the current Internetgeneration should be
replaced by a new architecture: 1. how strong are the
reasons for changing it?; this view takes into consideration
the costs and benefits for replacing a whole architecture
with a new one, and as long as patches can be provided for
issues that do not concern a vast majority of Internet users
anyway, chances are that patches are more feasible, and 2.
if the idea for a...