Chad Yoshikawa

WebOS: operating system services for wide area applications

Demonstrates the power of providing a common set of operating system services to wide-area applications, including mechanisms for naming, persistent storage, remote process execution, resource management, authentication and security. On a single machine, application developers can rely on the local operating system to provide these abstractions. In the wide area, however, application developers are forced to build these abstractions themselves or to do without. This ad-hoc approach often results in individual programmers implementing non-optimal solutions, wasting both programmer effort and system resources. To address these problems, we are building a system, WebOS, that provides the basic operating systems services needed to build applications that are geographically distributed, highly available, incrementally scalable and dynamically reconfigurable. Experience with a number of applications developed under WebOS indicates that it simplifies system development and improves resource utilization. In particular, we use WebOS to implement Rent-A-Server to provide dynamic replication of overloaded Web services across the wide area in response to client demands.

Assessing fast network interfaces

Assessing the performance of emerging high-speed networks is difficult. Our communication microbenchmark evaluates design changes in a communications system. Our microbenchmark generates a graphical signature from which we extract communication performance parameters for the hardware-software tandem. We use the LogP conceptual model for communication systems. (LogP stands for the parameters latency, overhead, gap, and processors). We study three important platforms that represent diverse points in the network interface design space: the Intel Paragon, Meiko CS-2, and a cluster of Sparcstation-20s connected by Myrinet switches using LANai SBus cards. Our study views each machine as a gray box supporting Active Messages and conforming to the LogP framework. We devise a simple set of communication microbenchmarks and measure the performance on each platform to obtain the LogP parameters.

Benchmarking message passing performance using MPI

This paper describes a set of microbenchmarks for measuring the performance characteristics of MPI implementations. We explain the rationale behind the benchmarks and present the benchmark results on the Intel Paragon, IBM SP2, and SGI Power Challenge. Our measurements reveal how the hardware architecture and the underlying MPI implementation affect the message passing performance on the three platforms.

Distributed hash queues: architecture and design

We introduce a new distributed data structure, the Distributed-Hash Queue, which enables communication between Network-Address Translated (NATed) peers in a P2P network. DHQs are an extension of distributed hash tables (DHTs) which allow for push and pop operators vs. the traditional DHT put and get operators. We describe the architecture in detail and show how it can be used to build a delay-tolerant network for use in P2P applications such as delayed-messaging. We have developed an initial prototype implementation of the DHQ which runs on PlanetLab using the Pastry key-based routing protocol.

The lonely NATed node

In this paper we take the position that current research in the area of distributed systems has all but forgotten about one of the largest collective Internet resources - the NATed node. These are hosts that are behind Network Address Translation (NAT) gateways and are hidden by the fact that they have private IP addresses. We argue that Distributed-Hash Tables [20], P2P systems [6], and Grid Computing [10] could greatly benefit by tapping into this forgotten pool of resources. Also, we give an outline of a service, the Distributed-Hash Queue (DHQ), that can enable these NATed resources to be exploited.

Why locally-fair maximal flows in client-server networks perform well

Maximal flows reach at least a 1/2 approximation of the maximum flow in client-server networks. By adding 1 additional time round to any distributed maximal flow algorithm we show how this 1/2-approximation can be improved on bounded-degree networks. We call these modified maximal flows ‘locally fair’ since there is a measure of fairness prescribed to each client and server in the network. Let N=(U,V,E,b) represent a client-server network with clients U, servers V, network links E, and node capacities b, where we assume that each capacity is at least one unit. Let d(u) denote the b-weighted degree of any node u∈U∪V, Δ=maxu2009{d(u)|u∈U} and δ=minu2009{d(v)|v∈V}. We show that a locally-fair maximal flow f achieves an approximation to the maximum flow of

Why Locally-Fair Maximal Flows in Client-Server Networks Perform Well

min{1,frac{varDelta^{2}-delta}{2varDelta^{2}-deltavarDelta-varDelta}

Using smart clients to build scalable services

}, and this result is sharp for any given integers δ and Δ. This results are of practical importance since local-fairness loosely models the steady-state behavior of TCP/IP and these types of degree-bounds often occur naturally (or are easy to enforce) in real client-server systems.

Parallel computing on the berkeley now

Maximal flows reach at least a 1/2 approximation of the maximum flow in client-server networks. By adding only 1 additional time round to any distributed maximal flow algorithm we show how this 1/2-approximation can be improved on bounded-degree networks. We call these modified maximal flows `locally fair since there is a measure of fairness prescribed to each client and server in the network. Let N = (U ,V ,E ,b ) represent a client-server network with clients U , servers V , network links E , and node capacities b , where we assume that each capacity is at least one unit. Let d (u ) denote the b -weighted degree of any node u *** U *** V , Δ = max {d (u ) | u *** U } and *** = min { d (v ) | v *** V }. We show that a locally-fair maximal flow f achieves an approximation to the maximum flow of

LogP Performance Assessment of Fast Network Interfaces

min { 1, frac{Delta^2 - delta} {2Delta^2 - deltaDelta - Delta}