Burkhard Englert

Virtualization Technology and its Impact on Computer Hardware Architecture

The computer hardware of todays world was designed and architected for the purpose of running a single operating system and its applications and therefore most of the hardware of such computer would be left underutilized. Virtualization enables the option of running multiple virtual computers on a single physical system. This paper studies the various methods, through which virtualization is supported in computer systems. The impact of this solution on computer hardware architecture is also investigated.

On the Efficiency of Atomic Multi-reader, Multi-writer Distributed Memory

This paper considers quorum-replicated, multi-writer, multi-reader (MWMR) implementations of survivable atomic registers in a distributed message-passing system with processors prone to failures. Previous implementations in such settings invariably required two rounds of communication between readers/writers and replica owners. Hence the question arises whether it is possible to have single round read and/or write operations in this setting. We thus devise an algorithm, called Sfw , that exploits a new technique called server side ordering ( SSO ), which ---unlike previous approaches--- places partial responsibility for the ordering of write operations on the replica owners (the servers). With SSO, fast write operations are introduced for the very first time in the MWMR setting. We prove that our algorithm preserves atomicity in all permissible executions. While algorithm SFW shows that in principle fast writes are possible, we also show that under certain conditions the MWMR model imposes inherent limitations on any quorum-based fast write implementation of a safe read/write register and potentially even restricts the number of writer participants in the system. In this case our algorithm achieves near optimal efficiency.

Megaphone: Fault Tolerant, Scalable, and Trustworthy P2P Microblogging

Micro-blogging, or the posting of weblogs entries that have a small number of characters (160 characters or less), has recently become more mainstream. Services that implement micro-blogging such as Twitter are usually based on the client- server model. This limits their scalability and fault tolerance. In this paper, we present a new secure microblogging system that is based on a peer-to-peer network. The network is arranged based on user certificates and is scalable, does not have a single point of failure, and does not depend on a single vendor’s proprietary service. The paper outlines the protocol specifics and provides implementation details for a secure, scaleable microblogging system.

Multivalued and Deterministic Peer-to-Peer Polling in Social Networks with Reputation Conscious Participants

We study the polling problem in a social network and present a private and precise n-ary voting protocol. We assume that members are concerned about their reputation in the sense that they do not want their votes to be disclosed and their potentially malicious actions to become public. Our protocol neither requires a central authority, strong cryptography, nor a randomization mechanism, using a simple secret sharing scheme and verification procedures to accurately determine the result of each poll. Our scalable protocol expands on the binary polling protocol by Guerraoui, Huguenin, Kermarrec and Monod and enables deterministic n-ary voting without significantly increasing the complexity of the protocol while preserving optimal voter privacy, allowing nodes to abstain from a poll--assuming honest nodes--and tabulating exact poll results. Misbehaving nodes are exposed with a non-zero probability and the probability of dishonest members violating voter privacy is balanced by their impact on the accuracy of the result. Limiting the number of dishonest nodes (B) such that they do not constitute a majority in any two consecutive groups of nodes and using a privacy parameter k, the impact of colluding malicious nodes is bounded by (2nk-k+1)B for any one element of the global tally and significantly less for all successive elements. Further, our algorithm improves on prior algorithms and ensures that voter privacy is optimally protected and that in the worst case, a nodes vote is only compromised with probability (Bi/nk+1)k+1.

Can memory be used adaptively by uniform algorithms

We introduce a novel term, memory-adaptive, that intuitively captures what it means for a distributed protocol to most efficiently make use of its shared memory. We also prove three results that relate to our memory-adaptive model. In our store/release protocols processors are required to store a value in shared MWMR memory so that it cannot be overwritten until it has been released by the processor. We show that there do not exist uniformly wait-free store/release protocols using only the basic operations read and write that are memory-adaptive to point contention. We further show that there exists a uniformly wait-free store/release protocol using only the basic operations read and write that is memory-adaptive to total contention. We finally show that there exists a uniformly wait-free store/release protocol using only the basic operations read, write, and write-plus that is memory-adaptive to interval contention and time-adaptive to total contention.

Optimal Sorting Algorithms for a Simplified 2D Array with Reconfigurable Pipelined Bus System

In recent years, many researchers have investigated optical interconnections as parallel computing. Optical interconnections are attractive due to their high bandwidth and concurrent access to the bus in a pipelined fashion. The Linear Array with Reconfigurable Pipelined Bus System (LARPBS) model is a powerful optical bus system that combines both the advantages of optical buses and reconfiguration. To increase the scalability of the LARPBS model, we propose a two-dimensional extension: a simplified two-dimensional Array with Reconfigurable Pipelined Bus System (2D ARPBS). While achieving better scalability, we show the effectiveness of this newly proposed model by designing two novel optimal sorting algorithms on this model. The first sorting algorithm is an extension of Leightons seven-phase columnsort algorithm that eliminates the restriction of sorting only an r × s array, where r ¿ s2 , and sorts an n × n array in O(log n) time. The second one is an optimal multiway mergesort algorithm that uses a novel processor efficient two-way mergesort algorithm and a novel multiway merge scheme to sort n2 items in O(log n) time. Using an optimal sorting algorithm Pipelined Mergesort designed for the LARPBS model as a building block, we extend our research on parallel sorting on the LARPBS to a more scalable 2D ARPBS model and achieve optimality in both sorting algorithms.

Implementing overlapping domain decomposition methods on a virtual parallel machine

To solve many partial differential equations of different types domain decomposition techniques were developed. Such algorithms are generally very well suited for implementation on a virtual parallel machine, simulated on a distributed system. While such algorithms are readily available and well established in the literature, authors do usually not concern themselves with questions of the practical implementability of their algorithms. In particular issues such as finding the optimal size of overlap in domain decompositions, finding the most effective number of subdomains or deciding whether to use exact or inexact subdomain solvers are beyond the scope of these results. In this paper we will address these questions. We first develop suitable domain decomposition algorithms for our virtual parallel machine. Through numerical experiments using our algorithms we then show that smaller linear systems work well even without any overlap while larger systems require that at least 10% of the subdomain size overlap to have convergency. The data also indicates that between 20% to 35% of the subdomain is the optimal overlap size. We next increase the number of subdomains and analyze its effect on the parallel solver. Our data shows that for a sufficiently large linear system computational speed of convergence improves significantly as the number of subdomains increases. We finally compare the effectiveness of exact and inexact domain solvers and show that the appropriate choice of the number of iterations in the worker algorithm, is much more efficient in the inexact solver than in the exact solver.

On Using Bittorrent for File Sharing in Mobile Ad-Hoc Networks

Many challenges exist in the design of a good distributed file sharing system. Transporting huge files over the network can create latency and sluggish performance. Scalability of a distributed file sharing system is another challenge. There is also no universal software that can transfer files between various platforms and devices and functions in all types of networks. In this paper we focus on the problem of distributed file sharing in the context of Mobile Ad Hoc Networks MANETs. In such networks nodes may join or leave at any given point in time so that traditional file sharing algorithms are not viable anymore. We present the architecture and implementation of an efficient distributed file sharing system for MANETs. Our system employs the Bittorrent protocol and a new tracker replicating procedure that enables file sharing even with nodes joining or leaving. By removing the tracker as a single point of failure our system outperforms other existing file sharing applications.

Cost-Based Optimization of Shipment Scheduling

Abstract Major suppliers and importers ship their products from many parts of the world, Products are typically packaged and loaded into containers at their points of origin, transported to a local port and shipped to a destination port where the containers will be unloaded and moved to their designated warehouses. We construct a nonlinear optimization model that minimizes the end-to-end cost of this transportation operation. A major shipper often has contracts with several carriers, each maintains a different voyage schedule and a different cost structure for shipping the containers. Containers arriving at their destination terminal are usually kept in the terminal a number of days before they are moved away either by trucks or by rail. Terminals typically allow free container dwell time up to a limit, beyond which demurrage charge will be imposed. The limit for free storage differs from one terminal to another and may even be different from one customer to another. How fast can a shipper move containers out of the terminal depends on warehouse capacity and truck availability that the shipper has contracted for the transport service. Based on these constraints, we want to find an optimal way to ship the containers from the point of origins to the warehouse destination. The optimality is in terms of end to end cost.

On the Use of XML for Port Communications

Abstract In this paper we analyze and compare EDI and XML based communication between shipping lines, terminal operators, government agencies, trucking companies, rail operators and other agents at the example of the Los Angeles and Long Beach ports. We investigate the suitability of (1) XML/EDI, (2) EDIINT (Web EDI), (3) Collaboration EDI and (4) Web Services for use as communication platforms at the ports. Currently at the Los Angeles and Long Beach ports information is mostly shared in a bilateral manner. As a result we describe and recommend an approach where terminals share data with shipping lines/carriers, rail operators, trucking companies, truckers and government agencies using XML based Web Services. This approach allows agents to either make information accessible through a web browser or through direct computer system to computer system communication. In the latter approach the terminals computer system will function as an automated information clearing house that can provide up-to-date, real-time information to all interested and authorized parties. Web Services furthermore allow securing and authenticating this information, protecting its dissemination. We believe that this approach can lead to increased throughput at the ports and an increase in capacity.