Any network has a structure and needs protocols to achieve its objectives. The researchers of IMDEA Networks Institute have an extensive expertise in architectures and protocols for communication networks, e.g., for network topology design, routing, forwarding, packet classification, in-network storage, congestion control, and media access control. Besides, we have research interests in other networking domains such as social networks, energy networks, and transportation networks.
Search of solutions
Our research takes a multi-disciplinary approach to the design and understanding of network protocols and architectures. We go beyond technological constraints and account also for social and economic factors. For example, our research on Internet routing and forwarding accounts for the multitude of Internet service providers and their individual economic interests. In working on either centralized or decentralized solutions to problems, we assume that perfect information is never available. To deal with such uncertainty as well as selfishness of individual entities, our analysis adopts game-theoretic techniques and online algorithms. Our protocol design assumes that behavior of counterparts is always unpredictable to some extent. Hence, the designed protocols rely on continuous learning and adaptation as the main modes of operation.
Practicality is another distinguishing aspect of our research. Real data serves as a departing point for our analytical efforts as well as a basis for validating our analytical conclusions. For instance, our large-scale simulation studies of Internet routing rely on real Internet typologies. Furthermore, we implement our theoretical ideas and make the prototypes available to the public, either directly or through our commercial partners.
Systems side of networks
An important focus of our work is on the systems side of networks. For example, we explore tradeoffs between simplicity and expressiveness of packet processing engines, new abstractions for heterogeneous control planes, and network virtualization techniques. We also work on networking aspects that pertain to cloud computing.
This research area targets the following objectives:
Novel architectures and protocols for behavioral networking
The Internet is modeled as an association of independent entities
Behavior of counterparts is not taken for granted
Continuous learning and adaptation are main modes of operations
Bridging the gap between network economics and networking
Deployment of innovative designs becomes the primary concern
Economic and political landscapes of the Internet are analyzed with higher fidelity
Economic-political knowledge guides the technical design
Making it easy to develop and deploy reliable, high-performance networked systems
Correct functioning of networks is paramount
SDN (Software Defined Networking) revolutionizes networking, and carries a lot of risk
New abstractions are developed to simplify network management and utilize the underlying network infrastructure more effectively
There are currently no job offers in this section.