Creating Tomorrow's Internet.
TCP (The Transport Control Protocol) in particular has come in for significant
criticism, and a growing body of experts believe it will need to be replaced.
Indeed, if it were easy to replace a fundamental Internet Protocol, this may
have been done some time ago. It's the complexity of the change management
problem that has delayed action rather than lack of recognised need for change.
Particular issues with TCP/IP include:
The study concludes that TCP - if not TCP/IP - needs to be replaced, probably
within a five to ten year time frame. The major issue to overcome is the
migration issue which is discussed below.
Each host on the Internet can be located via an IP number. The Domain Name
System (DNS) maps the numbers to names of hosts or websites (eg www.google.com,
www.hotmail.com ). Thus, when a user enters a name, the Internet knows which
number to send the query to by looking up the DNS database.
It should be noted that the other widespread user of distributed network
infrastructure, the telephone system, operates quite differently. It has no
domain name equivalent with trade mark implications in normal uses - to contact
a telephone address, you simply enter the number.
The DNS was introduced in 1984, several years before commercial traffic was able
to be part of the Internet. At the same time, a public database called Whois was
introduced, essentially to allow technical managers of hosts to contact their
peers. This is the Internet equivalent of a telephone directory, but also serves
a number of related purposes.
One issue with DNS is that it has not been possible to use native languages in
email addresses, domain names, and the WHOIS database. This poses significant
barriers to adoption for non-English speaking people.
The main problems here are that
Internationalised domain names (IDNs) have become a fundamental part of and
an iconic symbol for the digital divide issue. ICANN has been criticised at its
regular Public Forums for not giving the matter sufficient attention, failing to
make significant progress, and being negative in its analysis of this issue. The
Internet Analysis Report - 2004 examines this issue in detail.
Other issues with DNS include:
These are again problems that need to be addressed in a five year timeframe at the outside - some of them would be best handled more quickly if possible.
To all intents and purposes, email is already broke, and must be
fixed. The Internet's first and greatest killer application is now problematic.
In a survey examining email usage in 2003, the Pew Internet Project found that
25% of email users stated that the ever increasing volume of spam has reduced their overall use of email
70% of email users claimed spam had affected the quality of their on line experience
30% of users expressed fears that filtering approaches would cause loss of wanted mail
76% of users are bothered by offensive or obscene content in spam email
80% of users are bothered by deceptive or dishonest content in spam email.
Costs associated with spam have been estimated by various
research firms at between $10 billion (European Union, 2004) and $87 billion
(Nucleus Research, 2003) per annum. Spam volume is now estimated to exceed
legitimate email volume; in May 2004, 76 percent of inbound e-mails scanned by
email security provider MessageLabs Ltd were spam, up from 67 percent a month
earlier.
ICANN claims spam issues as out of scope. "… issues of concern to Internet
users, such as the rules for financial transactions, Internet content control,
unsolicited commercial email (spam), and data protection are outside the range
of ICANN's mission of technical coordination" (ICANN website). IETF has been
very slow at doing anything in this field, preferring to leave investigation of
the issues to a separate Internet Research Task Force (IRTF) group.
As a result, there is a general belief that nothing technical can be done to
prevent spam. However, our analysis suggests that the existing protocols are
significant contributors to the problem, and protocol reform could see spam
volume drop by at least 80%.
SMTP, the basic email standard, is the online equivalent of borders without
checkpoints and passports, or bank vaults without doors and locks. Some of the
SMTP security weaknesses are:
It allows anyone to connect with anyone without any system to say who they are
It is simple to forge messages and pretend to be someone you are not with no checking whatsoever
Not being one to one like telephone calls, it is easy to mass market to millions of email addresses at very low cost to the email sender.
These issues have been known for some time. Various attempts to
provide improved protocols have been undertaken, but essentially have resulted
in a mass of conflicting systems and standards. As a result, change is becoming
more complex to initiate.
Email upgrades are complicated by
Old systems which are never upgraded
Incorrect applications of email systems
The variety of applying protocols (eg http for webmail, smtp, nntp, pop etc)
The ubiquitous nature of email
IETF difficulties in handling big problems.
The Internet Analysis Report - 2004
analyses recent IETF work in this area and concludes that both governance issues
and protocol reform need to be addressed to provide a more comprehensive
solution.