So, for many of you who grew up with the Internet, 1994 was the year that it just got too big--in fact it was in December 1994 that the "Internet Liberation Front" hacked into a major ISP's routers, reprogrammed them and issued a manifesto that business should leave the Internet alone. 1994 is also the year in which engineering efforts focused on safely bringing companies onto the Internet with application-level gateways and NAT instead of tackling how to make Internet use easy and safe for everyone.
Don't Ignore History And here we are again, devoting considerable effort to managing all our traffic through a handful of gateways. Today we must deal with fault tolerance on many levels that were unnecessary only a few years ago. Load balancing and general bandwidth performance are becoming more important. We've lost sight of the original Internet hallmark--simplicity in the core--and now pay it lip service, at best.
After all, a network is an Internet, and its core is simple. It's just those fortresses at our borders where we need to put all of this specialized work.
The network is the computer, and we are repeating errors of the past on a grander scale from when the computer was the network. In 1974, we designed IPv4 to replace NCP. It was field-tested for the next eight years until it replaced NCP. IPv6 is now our best chance to repeat the phenomenal success of the original Internet. Engineers have been working on it since 1995. Our job now is to prepare for its deployment.
Larger addresses alone will not relieve the drain on our engineering resources that IPv4 has become. We need to make serious plans for a return to the end-to-end model. Over the next few months in this column, I will address some steps that must be taken to make our hosts independently responsible for their performance--a once-again open Internet. Our goal is to eliminate, or at least greatly reduce, the dependency on failure points such as gateways, NAT and firewalls.
We are approaching 30 years of internetworking (MERIT began in 1969). It's time to stop thinking about how to make better gateways. We need better ways not to gateway at all. Moreover, we have to think ahead so as not to revisit these issues with this sort of architecture in another 10 years.
For more on the history of the Internet, go to www.isoc.org/guest/zakon/ Internet/History/HIT.html.
Robert Moskowitz is a senior technical director at the International Computer Security Association and a member of the Internet Architecture Board. Send your comments on this column to him at rgm@htt-consult.com.
|
REPORTS
ANALYTICS
Follow 
