The highly successful architecture and protocols of today’s Internet may operate poorly in environments characterized by very long delay paths and frequent network partitions. These problems are exacerbated by end nodes with limited power or memory resources. Often deployed in mobile and extreme environments lacking continuous connectivity, many such networks have their own specialized protocols, and do not utilize IP. To achieve interoperability between them, we propose a network architecture and application interface structured around optionally-reliable asynchronous message forwarding, with limited expectations of end-to-end connectivity and node resources. The architecture operates as an overlay above the transport layers of the networks it interconnects, and provides key services such as in-network data storage and retransmission, interoperable naming, authenticated forwarding and a coarse-grained class of service.

Abstract—The use of peer-to-peer (P2P) applications is growing dramatically, particularly for sharing large video/audio files and software. In this paper, we analyze P2P traffic by measuring flowlevel information collected at multiple border routers across a large ISP network, and report our investigation of three popular P2P systems—FastTrack, Gnutella, and Direct-Connect. We characterize the P2P trafffic observed at a single ISP and its impact on the underlying network. We observe very skewed distribution in the traffic across the network at different levels of spatial aggregation (IP, prefix, AS). All three P2P systems exhibit significant dynamics at short time scale and particularly at the IP address level. Still, the fraction of P2P traffic contributed by each prefix is more stable than the corresponding distribution of either Web traffic or overall traffic. The high volume and good stability properties of P2P traffic suggests that the P2P workload is a good candidate for being managed via application-specific layer-3 traffic engineering in an ISP’s network. Index Terms—File sharing, peer-to-peer, P2P, traffic characterization, traffic measurement.

The primary use of the Internet is content distribution --- the delivery of web pages, audio, and video to client applications --- yet the Internet was never architected for scalable content delivery. The result has been a proliferation of proprietary protocols and ad hoc mechanisms to meet growing content demand. In this paper, we describe a content routing design based on name-based routing as part of an explicit Internet content layer. We claim that this content routing is a natural extension of current Internet directory and routing systems, allows efficient content location, and can be implemented to scale with the Internet.

It is accepted wisdom that the current Internet architecture conflates network locations and host identities, but there is no agreement on how a future architecture should distinguish the two. One could sidestep this quandary by routing directly on host identities themselves, and eliminating the need for network-layer protocols to include any mention of network location. The key to achieving this is the ability to route on flat labels. In this paper we take an initial stab at this challenge, proposing and analyzing our ROFL routing algorithm. While its scaling and efficiency properties are far from ideal, our results suggest that the idea of routing on flat labels cannot be immediately dismissed.

In today’s Internet, users can choose their local Internet service providers (ISPs), but once their packets have entered the network, they have little control over the overall routes their packets take. Giving a user the ability to choose between provider-level routes has the potential of fostering ISP competition to offer enhanced service and improving end-to-end performance and reliability. This paper presents the design and evaluation of a new Internet routing architecture (NIRA) that gives a user the ability to choose the sequence of providers his packets take. NIRA addresses a broad range of issues, including practical provider compensation, scalable route discovery, efficient route representation, fast route fail-over, and security. NIRA supports user choice without running a global link-state routing protocol. It breaks an end-to-end route into a sender part and a receiver part and uses address assignment to represent each part. A user can specify a route with only a source and a destination address, and switch routes by switching addresses. We evaluate NIRA using a combination of network measurement, simulation, and analysis. Our evaluation shows that NIRA supports user choice with low overhead.

The Internet architecture can be characterized as having a rather coarse grained and imperative style of network packet handling: confronted with an IP packet and its source and destination addresses, the infrastructure almost blindly and unalterably executes hundreds of resolution, routing and forwarding decisions. There are numerous attempts that try to "extend" the Internet in order to either reduce the immediate impact an arbitrary packet can have (e.g., NAT) , or to insert diversions from the normal processing paths in order to better use the existing resources (e.g., content delivery). In this paper we argue that we need a more fine grained control, in the hands of end nodes, over how packets are handled. The basic abstraction presented here is that of networking pointers, which we show to relate to low level concepts like ARP caches, but also high level routing decisions for terminal mobility, content delivery networks, or peer-to-peer overlay forming. We report on first implementation experiences of an "underlay" networking approach which uses pointer tricks underneath IP in order to provide new network layer services.

We propose 4+4, a simple address extension architecture for Internet that provides an evolutionary approach to extending the existing IPv4 address space in comparison to more complex and disruptive approaches best exemplified by IPv6 deployment. The 4+4 architecture leverages the existence of Network Address Translators (NATs) and private address realms, and importantly, enables the return to end-to-end address transparency as the incremental deployment of 4+4 progresses. During the transition to 4+4, only NATs and end-hosts need to be updated and not the network routers. The 4+4 architecture retains the existing semantics of Internet names and addresses, and only proposes simple changes to the network layer that focus entirely on address extension. Encapsulation is used as the main tool to maintain backward compatibility. We present the design, implementation, and evaluation of the 4+4 architecture and discuss our implementation experiences and results from local and wide-area Internet experimentation. The 4+4 source code is freely available from the Web (comet.columbia.edu/ipv44) for experimentation.

This paper describes two systems to recover the Internet connectivity impaired by private networks and firewalls. These devices cause asymmetry in the Internet, making peer-to-peer computing difficult or even impossible. The Condor system is one of those that are severely impaired by the asymmetry. Compared to normal peer-to-peer computing applications, Condor has stricter requirements, which are representative to any grid computing. To make Condor seamlessly work across private networks and over firewalls, we designed and implemented Dynamic Port Forwarding (DPF) and Generic Connection Brokering (GCB). Both DPF and GCB satisfy the representative requirements. Furthermore DPF supports dedicated large clusters very well because it is simple, efficient, and highly scalable. On the other hand, GCB perfectly supports non-dedicated or personal clusters because it is independent to private network or firewall technologies and does not require any administrative power to deploy it. In this paper, we describe the implementations of DPF and GCB and analyze them with respect to performance, deployability, security, and scalability.

In this paper, we argue that cache consistency mechanisms designed for stand-alone proxies do not scale to the large number of proxies in a content distribution network and are not flexible enough to allow consistency guarantees to be tailored to object needs. To meet the twin challenges of scalability and flexibility, we introduce the notion of cooperative consistency along with a mechanism, called cooperative leases, to achieve it. By supporting #-consistency semantics and by using a single lease for multiple proxies, cooperative leases allows the notion of leases to be applied in a flexible, scalable manner to CDNs. Further, the approach employs application-level multicast to propagate server notifications to proxies in a scalable manner. We implement our approach in the Apache web server and the Squid proxy cache and demonstrate its efficacy using a detailed experimental evaluation. Our results show a factor of 2.5 reduction in server message overhead and a 20% reduction in server state space overhead when compared to original leases albeit at an increased inter-proxy communication overhead.

Important research efforts are being conducted in the area of search, lookup, and routing, and are even increasing in the quest for P2P middleware that is both scalable and decentralized. To structure and classify current as well as to facilitate and give direction to future research, this methodology proposes a top-down two-dimensional design space. This design space has been developed for exhaustiveness so as to cover all possible design options, existing or yet to be conceived. A comprehensive survey of P2P search systems serves as a reference for the reader while at the same time validating the framework. An identification of areas in the design space not being covered by current systems leads to the design of a novel peer-to-peer-based keyword routing scheme. Finally, an evaluation of possible design options along the most important requirements will help guide system designers.