The security models used in Grid systems today strongly bear the marks of their diverse origin. Historically retrofitted to the distributed systems they are designed to protect and control, the security model is usually limited in scope and applicability, and its implementation tailored towards a few specific deployment scenarios. A common approach towards even the "basic" elements such as authentication to resources is only now emerging, whereas for more complex issues such as community organization, integration of site access control with operating systems, cross-domain resource provisioning, or overlay community Grids ("late authentication" for pilot job frameworks or community-based virtual machines) there is no single coherent and consistent "security" view. Via this paper we aim to share some observations on current security models and solutions found in Grid architectures and deployments today and identify architectural limitations in solving complex access control and policy enforcement scenarios in distributed resource management. The paper provides a short overview of the OGSA security services and other security solutions used in Grid middleware and operations practice. However, it is becoming clear that further development in Grid requires a fresh look at the concepts, both operationally and security-wise. This paper analyses the security aspects of different types of Grids and a set of use cases that may require extended security functionality, such as dynamic security context management, and management of stateful services. Recent developments in open systems security, and revisiting basic security concepts in networking and computing including the OSI security architecture and the concepts used in the trusted computing base provide interesting examples on how some of the conceptual security problems in Grid can be addressed, and on how the shortcomings of current systems and the frequently proposed "ad-hoc" stop-gaps for what are in fact complex security manageability problems may be avoided. This paper is thus intended to initiate and stimulate the wider discussion on the concepts of Grid security, thereby setting the scene for and providing input to a Grid security taxonomy leading to a more consistent Grid security architecture.

This paper proposes the security infrastructure for user-controlled Virtual Workspace Service (VWSS-UC) that comprises of three layers: trusted computing platform, secure virtualised workspace, and user aplication. The suggestions on the technology selection are provided for the first two layers: industry adopted Trusted Computing (TCG) platform, and Virtual Workspace Service (VWSS) developed in the framework of the Globus Toolkit. Solutions and implementation are proposed and discussed for the application authorisation session security context management. The paper is based on experiences gained from major Grid based projects such as EGEE, Globus Toolkit, and Phosphorus.

Local learning has been proposed as a common framework to predict both application run times and queue wait times based on workload traces [8]. The queue wait time is shown to be more difficult and expensive to predict because its distance calculations typically involve not only job attributes but also resource states. In this paper methods and algorithms are investigated to improve prediction accuracy and prediction performance for queue wait times. Firstly, the so-called "local tuning" is adopted to tune parameters for each training subset divided by a pivot attribute (e.g., group or queue name). Bias-variance analysis of error is conducted on local tuning and its global counterparts - tuning parameters on the whole training set. A method is then developed to select tuning type adaptively based on the generalization error and bias-variance decomposition. Secondly, an efficient search tree structure called "M-Tree" is integrated into our algorithm to speed up k-nearest neighbor search. Experimental studies are conducted to evaluate the proposed methods and algorithms using real-world workload traces, which are collected from the NIKHEF production cluster on the LHC Computing Grid and Blue Horizon in the San Diego Supercomputer Center (SDSC). The results show that adaptive tuning can reduce the average prediction error by 3 to 10 percents compared to global tuning, and that the M-Tree nearest neighbor search is up to 8 times faster than the sequential search.

In this paper we present an initial analysis of job failures in a large-scale data-intensive Grid. Based on three representative periods in production, we characterize the interarrival times and life spans of failed jobs. Different failure types are distinguished and the analysis is carried out further at the Virtual Organization (VO) level. The spatial behavior, namely where job failures occur in the Grid, is also examined. Cross-correlation structures, including how arrivals correlate with life spans of job failures, are analyzed and illustrated. We further investigate statistical models to fit the failure data and propose several failureaware scheduling strategies at the Grid level. Our results show that the overall failure rates in the Grid are quite significant, ranging from 25% to 33% of all submitted jobs. However, only 5% to 8% of the jobs failed after running on a certain Computing Element (CE). The rest of failed jobs are aborted or cancelled without running. A majority of failed jobs come from several large production VOs and a large amount of these failures are centered around several main CEs. The interarrival time processes of failed jobs are shown to be bursty, and the life spans exhibit strong autocorrelations. Based on the failure patterns we argue that it is important for the Grid resource brokers to track historical failure and take it into account in decision making. Some proactive measures and accountability issues are also discussed.

Abstract The D0 experiment has used the European DataGrid (EDG) testbed to reprocess real data obtained from the Tevatron collider at the Fermi National Accelerator Laboratory. Pushing the use of the EDG software beyond feasibility studies has produced a set of recommendations for authors of experiment‐level software, for producers of middleware and for designers of Grid systems. This paper describes the D0 experience with the EDG software and the resulting recommendations. Copyright © 2005 John Wiley & Sons, Ltd.

In a computational Grid which consists of many computer clusters, job start time predictions are useful to guide resource selections and balance the workload distribution. However, the basic Grid middleware available today either has no means of expressing the time that a site will take before starting a job or uses a simple linear scale. In this paper we introduce a system for predicting job start times on clusters. Our predictions are based on statistical analysis of historical job traces and simulation of site schedulers. We have deployed the system on the EDG (European Data-Grid) production cluster at NIKHEF. The experimental results show that acceptable prediction accuracy is achieved to reflect real site states and site-specific scheduling policies. We find that the average error of our job start time predictions is 18.9 percent of the average job queue wait time and this is around 20 times smaller than the average prediction error using the EDG solution.