Clovis Chapman

Predictive Resource Scheduling in Computational Grids

The integration of clusters of computers into computational grids has recently gained the attention of many computational scientists. While considerable progress has been made in building middleware and workflow tools that facilitate the sharing of compute resources, little attention has been paid to grid scheduling and load balancing techniques to reduce job waiting time. Based on a detailed analysis of usage characteristics of an existing grid that involves a large CPU cluster, we observe that grid scheduling decisions can be significantly improved if the characteristics of current usage patterns are understood and extrapolated into the future. The paper describes an architecture and an implementation for a predictive grid scheduling framework which relies on Kalman filter theory to predict future CPU resource utilisation. By way of replicated experiments we demonstrate that the prediction achieves a precision within 15-20% of the utilisation later observed and can significantly improve scheduling quality, compared to approaches that only take into account current load indicators.

Design for Future Internet Service Infrastructures

This paper presents current research in the design and integration of advance systems, service and management technologies into a new generation of Service Infrastructure for Future Internet of Services, which includes Service Clouds Computing. These developments are part of the FP7 RESERVOIR project and represent a creative mixture of service and network virtualisation, service computing, network and service management techniques.

The eMinerals collaboratory: tools and experience

Collaboratories provide an environment where researchers at distant locations work together at tackling important scientific and industrial problems. In this paper we outline the tools and principles used to form the eMinerals collaboratory, and discuss the experience, from within, of working towards establishing the eMinerals project team as a functioning virtual organisation. Much of the emphasis of this paper is on experience with the IT tools. We introduce a new application sharing tool.

Workflow issues in atomistic simulations

This article describes the techniques and mechanisms that have been used to tackle workflow problems encountered in the eMinerals project. We examine how established tools and technologies can be brought together to specify and deploy a computational process, consisting of a set of jobs and tasks, on our production level mini-grid infrastructure, with respect to a specific problem—the distribution of calculations required to determine, in a systematic way, the mechanisms by which pollutant molecules such as DDT, dioxins and biphenyls, become bound to soil minerals. We also briefly discuss the use of data standards such as the Chemical Mark-up Language (CML) and web-service based grid standards as a means to facilitate workflow specification.

Towards a federated cloud ecosystem (invited industrial talk)

Cloud computing has today become a widespread practice for the provisioning of IT services. Cloud infrastructures provide the means to lease computational resources on demand, typically on a pay per use or subscription model and without the need for significant capital investment into hardware. With enterprises seeking to migrate their services to the cloud to save on deployment costs, cater for rapid growth or generally relieve themselves from the responsibility of maintaining their own computing infrastructures, a diverse range of services is required to help fulfil business processes. In this talk, we discuss some of the challenges involved in deploying and managing an ecosystem of loosely coupled cloud services that may be accessed through and integrate with a wide range of devices and third party applications. In particular, we focus on how projects such as OpenStack are accelerating the evolution towards a federated cloud service ecosystem. We also examine how the portfolio of existing and emerging standards such as OAuth and the Simple Cloud Identity Management framework can be exploited to seamlessly incorporate cloud services into business processes and solve the problem of identity and access management when dealing with applications exploiting services across organisational boundaries.

Computational Grids for Mid-Sized Collaborative Projects: The eMinerals Experience

Grid computing has the potential to revolutionise how small groups of simulation scientists work together to tackle new science problems. In this paper we report how the eMinerals project has developed a small scale integrated compute and data grid infrastructure - the eMinerals minigrid - and developed generic job submission tools that exploit this infrastructure and which enable the science users to also access other grid systems.

Who Gets the Lion's Share in the Sharing Economy: A Case Study of Social Inequality in AirBnB.

Sharing economy platforms have rapidly disrupted and transformed many traditional markets. Companies such as AirBnB, in the housing market, and Uber, in the ride-sharing space, have thrived by creating opportunities for so-called “micro-entrepreneurs”, allowing them to leverage existing personal assets, such as a spare room or car, to generate additional income. While often heralded as an opportunity to reduce income inequality, opening opportunities through technology to a much larger segment of the population, there is however a latent concern that these platforms are in practice not as inclusive as advertised. In this paper we study the AirBnB listings in Chicago and examine a number of different dimensions regarding the hosts, their property and the environment within which they operate. Specifically we examine who the hosts are by detecting hosts’ ethnicity, gender and age using images posted publicly on the site. Leveraging this information and socio-economic metrics from the Census, we examine the properties different hosts offer and what is received in return. Finally we study how these hosts present their properties by measuring the aesthetic score of the main listing photographs using a deep learning algorithm. Our results suggest an ethnical discrepancy that affects minorities from lower socio-economic backgrounds, even when taking into account location and other attributes such as price of AirBnB listings. The findings also suggest that a wider range of factors, such as poorer pictures of listings, maybe affecting the inclusion and that could be corrected with internal policies and assistance of the platform owners.

Erratum: The eMinerals collaboratory: tools and experience

M.T. DOVE†, M. CALLEJA, R. BRUIN, J. WAKELIN, M.G. TUCKER, G.J. LEWIS, S. MEHMOOD HASAN, V.N. ALEXANDROV, M. KEEGAN, S. BALLARD, R.P. TYER, I. TODOROV, P.B. WILSON, M. ALFREDSSON, G.D. PRICE, C. CHAPMAN, W. EMMERICH, S.A. WELLS, A. MARMIER, S.C. PARKER and Z. DU Several of the figures in this paper did not reproduce correctly. For higher quality figures, please see the online version of this paper at http://www.tandf.co.uk/journals/ titles/08927022.asp