John Kubiatowicz

OceanStore: an architecture for global-scale persistent storage

OceanStore is a utility infrastructure designed to span the globe and provide continuous access to persistent information. Since this infrastructure is comprised of untrusted servers, data is protected through redundancy and cryptographic techniques. To improve performance, data is allowed to be cached anywhere, anytime. Additionally, monitoring of usage patterns allows adaptation to regional outages and denial of service attacks; monitoring also enhances performance through pro-active movement of data. A prototype implementation is currently under development.

Tapestry: a resilient global-scale overlay for service deployment

We present Tapestry, a peer-to-peer overlay routing infrastructure offering efficient, scalable, location-independent routing of messages directly to nearby copies of an object or service using only localized resources. Tapestry supports a generic decentralized object location and routing applications programming interface using a self-repairing, soft-state-based routing layer. The paper presents the Tapestry architecture, algorithms, and implementation. It explores the behavior of a Tapestry deployment on PlanetLab, a global testbed of approximately 100 machines. Experimental results show that Tapestry exhibits stable behavior and performance as an overlay, despite the instability of the underlying network layers. Several widely distributed applications have been implemented on Tapestry, illustrating its utility as a deployment infrastructure.

Bayeux: an architecture for scalable and fault-tolerant wide-area data dissemination

The demand for streaming multimedia applications is growing at an incr edible rate. In this paper, we propose Bayeux, an efficient application-level multicast system that scales to arbitrarily large receiver groups while tolerating failures in routers and network links. Bayeux also includes specific mechanisms for load-balancing across replicate root nodes and more efficient bandwidth consumption. Our simulation results indicate that Bayeux maintains these properties while keeping transmission overhead low. To achieve these properties, Bayeux leverages the architecture of Tapestry, a fault-tolerant, wide-area overlay routing and location network.

Erasure Coding Vs. Replication: A Quantitative Comparison

Peer-to-peer systems are positioned to take advantage of gains in network bandwidth, storage capacity, and computational resources to provide long-term durable storage infrastructures. In this paper, we quantitatively compare building a distributed storage infrastructure that is self-repairing and resilient to faults using either a replicated system or an erasure-resilient system. We show that systems employing erasure codes have mean time to failures many orders of magnitude higher than replicated systems with similar storage and bandwidth requirements. More importantly, erasure-resilient systems use an order of magnitude less bandwidth and storage to provide similar system durability as replicated systems.

Towards a Common API for Structured Peer-to-Peer Overlays

In this paper, we describe an ongoing effort to define common APIs for structured peer-to-peer overlays and the key abstractions that can be built on them. In doing so, we hope to facilitate independent innovation in overlay protocols, services, and applications, to allow direct experimental comparisons, and to encourage application development by third parties. We provide a snapshot of our efforts and discuss open problems in an effort to solicit feedback from the research community.

A view of the parallel computing landscape

Writing programs that scale with increasing numbers of cores should be as easy as writing programs for sequential computers.

APRIL: a processor architecture for multiprocessing

Processors in large-scale multiprocessors must be able to tolerate large communication latencies and synchronization delays. This paper describes the architecture of a rapid-context-switching processor called APRIL with support for fine-grain threads and synchronization. APRIL achieves high single-thread performance and supports virtual dynamic threads. A commercial RISC-based implementation of APRIL and a run-time software system that can switch contexts in about 10 cycles is described. Measurements taken for several parallel applications on an APRIL simulator show that the overhead for supporting parallel tasks based on futures is reduced by a factor of two over a corresponding implementation on the Encore Multimax. The scalability of a multiprocessor based on APRIL is explored using a performance model. We show that the SPARC-based implementation of APRIL can achieve close to 80% processor utilization with as few as three resident threads per processor in a large-scale cache-based machine with an average base network latency of 55 cycles.

Probabilistic location and routing

We propose probabilistic location to enhance the performance of existing peer-to-peer location mechanisms in the case where a replica for the queried data item exists close to the query source. We introduce the attenuated Bloom filter, a lossy distributed index data structure. We describe how to use these data structures for document location and how to maintain them despite document motion. We include a detailed performance study which indicates that our algorithm performs as desired, both finding closer replicas and finding them faster than deterministic algorithms alone.

Maintenance-free global data storage

Explores mechanisms for storage-level management in OceanStore, a global-scale distributed storage utility infrastructure, designed to scale to billions of users and exabytes of data. OceanStore automatically recovers from server and network failures, incorporates new resources and adjusts to usage patterns. It provides its storage platform through adaptation, fault tolerance and repair. The only role of human administrators in the system is to physically attach or remove server hardware. Of course, an open question is how to scale a research prototype in such a way to demonstrate the basic thesis of this article - that OceanStore is self-maintaining. The allure of connecting millions or billions of components together is the hope that aggregate systems can provide scalability and predictable behavior under a wide variety of failures. The OceanStore architecture is a step towards this goal.

Brocade: Landmark Routing on Overlay Networks

Recent work such as Tapestry, Pastry, Chord and CAN provide efficient location utilities in the form of overlay infrastructures. These systems treat nodes as if they possessed uniform resources, such as network bandwidth and connectivity. In this paper, we propose a systemic design for a secondaryoverlay of super-nodes which can be used to deliver messages directly to the destinations local network, thus improving route efficiency. We demonstrate the potential performance benefits by proposing a name mapping scheme for a Tapestry-Tapestry secondary overlay, and show preliminary simulation results demonstrating significant routing performance improvement.