Norman Bobroff

Dynamic Placement of Virtual Machines for Managing SLA Violations

A dynamic server migration and consolidation algorithm is introduced. The algorithm is shown to provide substantial improvement over static server consolidation in reducing the amount of required capacity and the rate of service level agreement violations. Benefits accrue for workloads that are variable and can be forecast over intervals shorter than the time scale of demand variability. The management algorithm reduces the amount of physical capacity required to support a specified rate of SLA violations for a given workload by as much as 50% as compared to static consolidation approach. Another result is that the rate of SLA violations at fixed capacity may be reduced by up to 20%. The results are based on hundreds of production workload traces across a variety of operating systems, applications, and industries.

Position measurement with a resolution and noise‐limited instrument

A common problem in experimental data analysis is to locate the position of a signal to an accuracy which is substantially less than the actual signal width. By applying maximum likelihood estimation to this problem, this paper derives theoretical limits on the ability to locate signal position. The limiting error in position measurement is shown to be a simple function of the instrument resolution, the density of sample points, and the signal‐to‐noise ratio of the data. An interesting conclusion is that position information on a much finer scale than the minimum instrument sampling interval is contained in data of modest signal‐to‐noise ratio. The common procedure of excluding the portion of the data lying below an amplitude threshold to guard against background fluctuations is incorporated in the maximum likelihood analysis. It is shown that selection of the optimum amplitude threshold level depends on the type of noise present in the data, and can be an important factor in position accuracy. The analyt...

Cloud federation in a layered service model

We show how a layered Cloud service model of software (SaaS), platform (PaaS), and infrastructure (IaaS) leverages multiple independent Clouds by creating a federation among the providers. The layered architecture leads naturally to a design in which inter-Cloud federation takes place at each service layer, mediated by a broker specific to the concerns of the parties at that layer. Federation increases consumer value for and facilitates providing IT services as a commodity. This business model for the Cloud is consistent with broker mediated supply and service delivery chains in other commodity sectors such as finance and manufacturing. Concreteness is added to the federated Cloud model by considering how it works in delivering the Weather Research and Forecasting service (WRF) as SaaS using PaaS and IaaS support. WRF is used to illustrate the concepts of delegation and federation, the translation of service requirements between service layers, and inter-Cloud broker functions needed to achieve federation.

Residual errors in laser interferometry from air turbulence and nonlinearity

Laser displacement interferometry is used extensively in precision equipment for semiconductor manufacture. In these applications it is often necessary to introduce a high velocity airflow to the measurement environment to minimize the density of airborne particulate contaminants. The performance of the heterodyne interferometer is degraded by the resulting fluctuations in the index of refraction along the beam path. The magnitude, correlation length, and probability distribution of the optical path length (OPL) fluctuations are measured for several airflow conditions. The data are interpreted in terms of the path length errors for some common interferometric configurations. The OPL fluctuations are generally less significant than the systematic sources of measurement error. A more fundamental limit on the accuracy of the heterodyne Michelson interferometer is the periodic nonlinearity caused by leakage of the frequency components in the beamsplitter. The effect is discussed in detail. A direct observation of the nonlinearity is reported. The magnitude of the effect is about lambda/64 for the beam splitters used in this experiment. A simple technique which indicates the presence and magnitude of the nonlinearity is described.

Optical sources of non-linearity in heterodyne interferometers

Abstract Non-linearity of the two-frequency Michelson interferometer is studied by measurement of the modulation in phase and amplitude of the interference signal at the receiver. An important conclusion is that frequency mixing is not a major cause of non-linearity unless the mixing is asymmetric between the two arms of the interferometer. One source of asymmetry is non-orthogonality of the frequency states produced by the laser. Differential transmission between the reference and measurement paths can cause polarization states that are initially orthogonal to become non-orthogonal at the interferometer output. The influence of beam-splitter leakage, misorientation of the polarization axes of the source, and imperfect waveplates are also considered. It is not widely known that typical retroreflectors cause about a 7° rotation of the polarization state. However, polarization mixing induced by the retroreflectors is small in most, but not all, common configurations of the interferometer.

Enabling Interoperability among Meta-Schedulers

Grid computing supports shared access to computing resources from cooperating organizations or institutes in the form of virtual organizations. Resource brokering middleware, commonly known as a meta-scheduler or a resource broker, matches jobs to distributed resources. Recent advances in meta- scheduling capabilities are extended to enable resource matching across multiple virtual organizations. Several architectures have been proposed for interoperating meta-scheduling systems. This paper presents a hybrid approach, combining hierarchical and peer-to-peer architectures for flexibility and extensibility of these systems. A set of protocols are introduced to allow different meta-scheduler instances to communicate over Web Services. Interoperability between three heterogeneous and distributed organizations (namely, BSC, FIU, and IBM), each using different meta-scheduling technologies, is demonstrated under these protocols and resource models.

BPEL4Job: A Fault-Handling Design for Job Flow Management

Workflow technology is an emerging paradigm for systematic modeling and orchestration of job flow for enterprise and scientific applications. This paper introduces BPEL4Job, a BPEL-based design for fault handling of job flow in a distributed computing environment. The features of the proposed design include: a two-stage approach for job flow modeling that separates base flow structure from fault-handling policy, a generic job proxy that isolates the interaction complexity between the flow engine and the job scheduler, and a method for migrating flow instances between different flow engines for fault handling in a distributed system. An implementation of the design based on a set of industrial products from IBM is presented and validated using a Montage application.

Enabling Interoperability among Grid Meta-Schedulers

The goal of Grid computing is to integrate the usage of computer resources from cooperating partners in the form of Virtual Organizations (VO). One of its key functions is to match jobs to execution resources efficiently. For interoperability between VOs, this matching operation occurs in resource brokering middleware, commonly referred to as the meta-scheduler or meta-broker. In this paper, we present an approach to a meta-scheduler architecture, combining hierarchical and peer-to-peer models for flexibility and extensibility. Interoperability is further promoted through the introduction of a set of protocols, allowing meta-schedulers to maintain sessions and exchange job and resource state using Web Services. Our architecture also incorporates a resource model that enables an efficient resource matching across multiple Virtual Organizations, especially where the compute resources and state are dynamic. Experiments demonstrate these new functional features across three distributed organizations (BSC, FIU, and IBM), that internally use different job scheduling technologies, computing infrastructure and security mechanisms. Performance evaluations through actual system measurements and simulations provide the insights on the architecture’s effectiveness and scalability.

Alignment errors from resist coating topography

The registration difficulty associated with a resist‐coated alignment mark depends on the type of optics used to illuminate and view the wafer alignment aid. The interaction of the light with the mark geometry is studied by calculating the outgoing wave front that results when an arbitrary collection of correlated or uncorrelated incident wave fronts is reflected from the nonplanar target structure. This solution is combined with a model of the partially coherent imaging system to predict the alignment signal. Bright‐field and dark‐field systems typical of those used in optical lithography are studied. Dark‐field viewing optics are generally less sensitive to offsets caused by resist coating asymmetries. The statistical alignment precision is related to the shape of the alignment signal and the signal‐to‐noise ratio. It is shown that the increased contrast in dark‐field systems can provide an alignment precision comparable to that of bright‐field systems despite a substantially lower photon collection eff...

Critical alignments in plane mirror interferometry

Abstract Each axis of a plane mirror interferometer has three associated measurement vectors: the interferometer beam, mirror normal, and axis of measurement. The transducer axis of a plane mirror interferometer target is the mirror normal. This contrasts with a retroflector for which the transducer axis is the interferometer beam. The classical cosine error in x-y plane mirror interferometry is removed by the orienting the mirror normals in the travel plane. The axes do not have to be orthogonal to eliminate cosine error.