UC Berkeley School of Information

The network research community has been challenging itself to conceive of what our global network of 15 years from now should be. While the Internet of today is a great success, it also has limitations that suggest it may not be well-positioned to meet all the requirements we will face as we move into the future. As we consider design options for a future network, we have come to realize that some of the original design decisions, while effective, were not the only way to go. In fact, we could have taken different forks in the road as we designed, and ended up in very different places, perhaps with a network that is equally effective at supporting a wide range of applications, but different in other ways: more secure, easier to install and manage, better aligned to motivate investment, and so on. In this talk, I will identify a few of these alternate designs, and describe how the network they would induce would differ from what we see today. In doing so, I will try to illustrate both the nature of the design