FIA: Collaborative Research: MobilityFirst: A Robust and Trustworthy Architecture for the Future Mobile Internet

Principal Investigator: Michael Reiter
Funding Agency: National Science Foundation
Agency Number: CNS-1040626

Abstract
The "MobilityFirst" architecture introduced in this proposal is based on the recognition that Internet usage is changing very rapidly from fixed hosts to mobile devices such as cell phones, portable computers, machines and sensors. It is anticipated that by 2015, mobile/wireless devices will vastly outnumber fixed hosts (~10B mobiles vs. ~1B hosts), motivating us to design a new Internet architecture that directly addresses this very predictable yet fundamental change in network usage. Our vision is that of a trustworthy and robust Internet architecture which supports mobile and embedded devices as "first-class" users, thus enabling a variety of new applications efficiently, securely, and at scale. The primary design goals of the proposed architecture are identified as: dynamic end point and network mobility as the norm; robust transport in the presence of disconnection or varying path quality; strong security for network names and addresses; intentional data receipt specified by receivers; addressability of content or context; multicast, multi-path and multi-homing modes as basic services; location as a fundamental network attribute; viability of the economic framework; manageability; some degree of evolvability and backward compatibility; and the ability to conform to different kinds of regulation. Consideration of the above mentioned design goals led us to propose a specific MobilityFirst network architecture with the following key components: (1) separation of naming and addressing, implemented via a fast dynamic name resolution service for mapping mobile endpoint names or content/context identifiers to network addresses; (2) self-certifying public key addresses to support strong authentication and security; (3) generalized delay tolerant intranet routing with built-in router storage; (4) flat-label internetwork routing with public key addresses; (5) hop-by-hop (rather than end-to-end) transport protocols; (6) a cleanly separated network management plane to provide enhanced visibility; (7) optional privacy features for preserving user and location data; and (8) an integrated computing and storage layer at routers to support programmability and virtualization of network services. We believe that an architecture built with these components can achieve most of the desired functional requirements of the future Internet with order-of-magnitude gains in network capacity, performance and usability when compared with IP-based solutions for representative mobile and wireless usage scenarios. In addition, the same architecture also works very well for the special case of fixed hosts because mobility and disconnection imply extra-strength network capabilities for security, privacy, availability and robustness. The proposed project is a collaborative effort involving Rutgers, UMass, MIT, Duke, U Michigan, UNC, U Nebraska with input and feedback from several industrial research partners. The team includes a small core of highly experienced network/security architects along with a number of capable young investigators (including several NSF CAREER award winners) who can be expected to bring a fresh perspective to the Internet design problem. The team has both the breadth and depth of expertise (as well as relevant prior research experience) on protocol architecture, network security, network analysis/evaluation, network economics, wireless/mobile systems and large-scale prototyping as needed to carry out the project successfully. The project's work plan has been organized into eight work packages of which five are focused tasks (WP1-naming/addressing/routing, WP2-security/privacy, WP3-network management, WP4-pervasive/wireless usage scenarios; WP5-network economics) to be done by specialized expertise groups and three are cross-cutting tasks (WP6-network architecture; WP7-evaluation/validation; WP8-system prototyping and deployment) across multiple areas. The project's major milestones are: Year 1 - an architecture white paper, protocol design documents, preliminary evaluation results and a small-scale proof-of-concept laboratory prototype; Year 2 - detailed validations of each key protocol component, overall network evaluation results, and a multi-site proof-of-concept network prototype (using available GENI infrastructure); Year 3 - updated protocol design based on feedback evaluation results, and medium-scale network protocol deployment on multiple campus testbeds. The project will conclude with a comprehensive validation and evaluation of the usability and performance of the MobilityFirst architecture using both controlled experiments and application trials with real-world end-users. Broader impacts: This project is expected to have a significant influence on the design and practice of next-generation Internet protocols. The mobility centric approach adopted in this proposal will help to bridge the gap between cellular network protocols and the Internet protocol, enabling a unified future Internet which effectively supports both mobile and fixed devices leading to significant new economic
benefits and application functionality. The mobile Internet will support new location- and context-aware applications which will improve safety, convenience and productivity for everyone who uses a cellular or mobile data device (over 4 billion worldwide in 2010). The project will also have significant educational and workforce training benefits at undergraduate and graduate levels, including new graduate seminar courses on Internet architecture, undergraduate network prototyping projects and annual research workshops involving students and faculty from CS and ECE departments nationwide.