The architecture of the Internet is based on a number of principles, including the self-describing datagram packet, the end-to-end arguments, diversity in technology and global addressing. As the Internet has moved from a research curiosity to a recognized component of mainstream society, new requirements have emerged that suggest new design principles, and perhaps suggest that we revisit some old ones. This paper explores one important reality that surrounds the Internet today: different stakeholders that are part of the Internet milieu have interests that may be adverse to each other, and these parties each vie to favor their particular interests. We call this process "the tussle". Our position is that accommodating this tussle is crucial to the evolution of the network's technical architecture. We discuss some examples of tussle, and offer some technical design principles that take it into account.

Connectivity is central to pervasive computing environments. We seek to catalyze a world of rich and diverse connectivity through technologies that drastically simplify the task of providing, choosing, and using wireless network services; creating a new and more competitive environment for these capabilities. A critical requirement is that users actually benefit from this rich environment, rather than simply being overloaded with choices. We address this with an intelligent software agent that transparently and continually chooses from among available network services based on its user's individual needs and preferences, while requiring only minimal guidance and user interaction. We present an overview and model of the network service selection problem. We then describe an adaptive user agent that learns its user's network service preferences from a very minimal, intuitive set of inputs, and autonomously and continually selects the service that best meets the user's needs. Results from preliminary user experiments are presented that demonstrate the effectiveness of our agent.

We propose a new objective for network research: to build a fundamentally different sort of network that can assemble itself given high level instructions, reassemble itself as requirements change, automatically discover when something goes wrong, and automatically fix a detected problem or explain why it cannot do so.We further argue that to achieve this goal, it is not sufficient to improve incrementally on the techniques and algorithms we know today. Instead, we propose a new construct, the Knowledge Plane, a pervasive system within the network that builds and maintains high-level models of what the network is supposed to do, in order to provide services and advice to other elements of the network. The knowledge plane is novel in its reliance on the tools of AI and cognitive systems. We argue that cognitive techniques, rather than traditional algorithmic approaches, are best suited to meeting the uncertainties and complexity of our objective.

The architecture of the Internet is based on a number of principles, including the self-describing datagram packet, the end to end arguments, diversity in technology and global addressing. As the Internet has moved from a research curiosity to a recognized component of mainstream society, new requirements have emerged that suggest new design principles, and perhaps suggest that we revisit some old ones. This paper explores one important reality that surrounds the Internet today: different stakeholders that are part of the Internet milieu have interests that may be adverse to each other, and these parties each vie to favor their particular interests. We call this process "the tussle". Our position is that accommodating this tussle is crucial to the evolution of the network's technical architecture. We discuss some examples of tussle, and offer some technical design principles that take it into account.

This paper proposes GIA, a scalable architecture for global IP-anycast. Existing designs for providing IP-anycast must either globally distribute routes to individual anycast groups, or confine each anycast group to a pre-configured topological region. The first approach does not scale because of excessive growth in the routing tables, whereas the second one severely limits the utility of the service. Our design scales by dividing inter-domain anycast routing into two components. The first component builds inexpensive default anycast routes that consume no bandwidth or storage space. The second component, controlled by the edge domains, generates enhanced anycast routes that are customized according to the beneficiary domain's interests. We evaluate the performance of our design using simulation, and prove its practicality by implementing it in the Multi-threaded Routing Toolkit.

This paper describes a new at-most-once message passing protocol that provides guaranteed detection of duplicate messages even when the receiver has no state stored for the sender. It also discusses how to use at-most-once messages to implement higher-level primitives such as at-once-remote procedure calls and sequenced bytestream protocols. Our performance measurements indicate that at-most-once RPCs can provide at the same cost as less desirable forms of RPCs that do not guarantee at-most-once execution. Our method is based on the assumption that clocks throughout the system are loosely synchronized. Modern clock synchronization protocols provide good bounds on clock skew with high probability; our method depends on the bound for performance but not for correctness.

This paper describes a new message passing protocol that provides guaranteed detection of duplicate messages even when the receiver has no state stored for the sender. It also discusses how to use these messages to implement higher-level primitives such as at-most-once remote procedure calls and sequenced bytestream protocols, and describes an implementation of at-most-once RPCs using our method. Our performance measurements indicate that at-most-once RPCs can be provided at the same cost as less desirable RPCs that do not guarantee at-most-once execution. Our method is based on the assumption that clocks throughout the system are loosely synchronized. Modern clock synchronization protocols provide good bounds on clock skew with high probability; our method depends on the bound for performance but not for correctness.