Erik Nordström

Haggle: Opportunistic mobile content sharing using search

We present Haggle, a content-sharing system for mobile devices, allowing users to opportunistically share content without the support of infrastructure. Mobile devices share content and interests over direct WiFi or Bluetooth links, and may store-carry-forward content on behalf of others based on interests, bridging otherwise disconnected devices. Unlike traditional Internet-based content sharing systems, Haggle faces disconnections, unpredictable mobility, and time-limited contacts, which pose unique challenges to the systems design and implementation. While similar content-sharing systems typically value every content item the same, Haggle uses a ranked search to judiciously decide which content to exchange, and in which order. The search matches a devices locally stored content against the interests of other users that the device has collected, prioritizing relevant content when contacts are time limited and resources scarce. Thus, search enables dissemination of content in order of how strongly users desire it, offering delay and resource savings by exchanging the content that matters. An optional content delegation mechanism allows Haggle to altruistically disseminate a limited amount of items based on the interests of third-party nodes, increasing the benefit of the network as a whole, and protecting against networks that are partitioned along interests. Ranked searches, combined with delegation, allow Haggle to balance the short-term benefit of exchanging a content item between two nodes against the long-term benefit to the network as a whole. We evaluate Haggle through a real-world experiment with mobile phones, running a picture sharing application, complemented by trace-based emulations. Our results show that a content item whose interest group (the nodes that desire it) has strong interests is delivered with lower delay, and to a higher fraction of members, than an item with a similar sized group with weak interests. Compared to a relevance-agnostic system, Haggle can deliver the most relevant items in one third of the time and at a lower cost.

A formally-verified migration protocol for mobile, multi-homed hosts

Modern consumer devices, like smartphones and tablets, have multiple interfaces (e.g., WiFi and 4G) that attach to new access points as users move. These mobile, multi-homed computers are a poor match with an Internet architecture that binds connections to fixed endpoints with topology-dependent addresses. As a result, hosts typically cannot spread a connection over multiple interfaces or paths, or change locations without breaking existing connections. In this paper, we create an end-to-end connection control protocol (ECCP) that allows hosts to communicate over multiple interfaces with dynamically-changing IP addresses and works with multiple data-delivery protocols (i.e., reliable or unreliable transport). Each ECCP connection consists of one or more flows, each associated with an interface or path. Through end-to-end signaling, a host can move an existing flow from one interface to another, or change its IP address, without any support from the underlying network. We develop formal models to verify that ECCP works correctly in the presence of packet loss, out-of-order delivery, and frequent mobility, and to identify bugs and design limitations in earlier mobility protocols.

Increasing network resilience through edge diversity in NEBULA

The primary focus of the NEBULA Future Internet Architecture is to provide resilient networking for the emerging cloud computing model. One of the attractions of cloud computing is its support for online services and data storage by thin clients such as mobile devices. This paper describes two components of NEBULAs edge network technology, Serval and CRYSTAL. Serval provides a new layer 3.5 service abstraction that naturally supports mobility, multi-homing, and multi-path transport, while CRYSTAL is a new virtualization scheme for software radios that makes it easier to expose greater network diversity at the network edge.