Network Science and Engineering:
Fundamental Challenges
- Understand emergent behaviors, local–global interactions, system failures and/or degradations
- Develop models that accurately predict and control network behaviors
- Develop architectures for self-evolving, robust, manageable future networks
- Develop design principles for seamless mobility support
- Leverage optical and wireless substrates for reliability and performance
- Understand the fundamental potential and limitations of technology
- Design secure, survivable, persistent systems, especially when under attack
- Understand technical, economic and legal design trade-offs, enable privacy protection
- Explore AI-inspired and game-theoretic paradigms for resource and performance optimization 
Science
Technology
Society
Enable new applications and new economies, while ensuring security and privacy
Security, privacy, economics, AI, social science researchers
Network science and engineering researchers
Understand the complexity of large-scale networks
Distributed systems and substrate researchers
Develop new architectures, exploiting new substrates
[Jeannette Wing, CMU and NSF]
The field of computing has science drivers, technology drivers, and societal drivers.  Network science and engineering is no different.  For all three drivers, there are fundamental challenges that face us if we are to make progress in answering the question stated earlier.  Let’s use these three interacting sets of drivers to frame a research agenda for network science and engineering.   I’ll start with the science drivers.

The scientific challenge is to understand the complexity of networked systems.  We do not have theories of our networks such that we understand their emergent properties.  We do not have formal models of our networks such that we can assert any guarantees of reliability or survivability, especially in the presence of disruptive events or malicious attacks.  We do not have models of our networks such that we can accurately predict their performance; for example Poisson models and heavy-tail distribution models are unrealistic or overly simplistic.

The technology drivers come from new communication substrates such as wireless and optical.  These new substrates change the physical characteristics of the network, suggesting new network architectures, where the goal is not just simply optimizing the transmission of a packet from one host to another.  For example, because of increased bandwidth and large distances between nodes we already view the network not just as a communication medium but as a storage medium.  Technology drivers also come from new devices. The ubiquity of mobile computing and communication devices, for example, is forcing us to rethink our networks in truly fundamental ways, to include at the very least the human dimension of what networks are.

Moreover, new theories and new architectures enable new societal uses of the network—creating new opportunities such as virtual worlds and new challenges such as ensuring the security and privacy of users and their data.

For your interest, I’ve included in small font, a sampling of fundamental challenges for each of these three drivers.

The research communities active in addressing these intellectual challenges are correspondingly listed on the right.  Indeed one of the opportunities in developing a research agenda for network science and engineering, is ambitiously to ignite a diverse range of communities to work together and look at networked systems from a holistic viewpoint—it’s not enough to invent a new theory, a new architecture, or a new application without understanding the implications it has for all aspects of the networked system.

I would like to emphasize the expanse of these intellectual challenges.  I am talking not just about networks as the invisible infrastructure that we all take for granted today, but about networks of people and organizations that creatively use the invisible infrastructure in unforeseen ways.

To calibrate the rapid growth in societal impact of the Internet, let me give you some concrete examples: Mosaic, the first browser, was created in 1993 and Google was founded in 1998, less than ten years ago.  Google is known for search but it makes its money on ads;  who would have guessed?  (As as aside: remember that Mosaic and Google, like other browsers and search engines, were born out of NSF/CISE-funded research projects in universities.) Within just the past four years, we have seen the creation and influence of social networks embraced by the younger generation: SecondLife, FaceBook, and YouTube.  We are also seeing old institutions remake themselves, including the print media like newspapers and the music recording industry.  More sobering, in April of this year we witnessed the first cyber attack on a nation—Estonia—by a well-organized group in Russia.  And just three weeks ago, the US company Seagate reported that hard drives sold in Taiwan contained Trojan Horses planted by subcontractors in China so that any information stored on those tainted hard drives was automatically uploaded to websites in Beijing.  We did not anticipate any of these phenomena—good and bad—nor did we anticipate how they have completely changed the way in which people and organizations interact with each other.

It is the expanse of all these intellectual challenges that necessitates including not just researchers in computer and information science and engineering, but researchers in economics and the social sciences.