A method of job scheduling in a computing environment, comprising collecting historical job data related to job submissions received from a plurality of users, the historical job data comprising: resource request data indicative of resources requested by the user for executing a job, and resource usage data indicative of actual resources used to execute the job; calculating a believability score for each of the plurality of users based on a comparison between the job resource request data and the resource usage data related to job submissions received from the user, the believability score indicating a degree to which the user overestimates resources required to execute jobs; receiving a new job submission from a user, specifying a new job that is to be executed and resources requested by the user for executing the new job; and scheduling the new job for execution based on the believability score for the user.

Job scheduling based on historical job data

While workload collocation is a necessity to increase energy efficiency of contemporary multi-core hardware, it also increases the risk of performance anomalies due to workload interference Pinning certain workloads to a subset of CPUs is a simple approach to increasing workload isolation, but its effect depends on workload type and system architecture Apart from common sense guidelines, the effect of pinning has not been extensively studied so far In this paper we study the impact of CPU pinning on performance interference and energy efficiency for pairs of collocated workloads Besides various combinations of workloads, virtualization and resource isolation, we explore the effects of pinning depending on the level of background load The presented results are based on more than 1000 experiments carried out on an Intel-based NUMA system, with all power management features enabled to reflect real-world settings We find that less common CPU pinning configurations improve energy efficiency at partial background loads, indicating that systems hosting collocated workloads could benefit from dynamic CPU pinning based on CPU load and workload type

Analyzing the impact of CPU pinning and partial CPU loads on performance and energy efficiency

We present ease.ml, a declarative machine learning service platform. With ease.ml, a user defines the high-level schema of an ML application and submits the task via a Web interface. The system then deals with the rest, such as model selection and data movement. The ultimate question we hope to understand is that, as a "service provider" that manages a shared cluster of machines running machine learning workloads, what is the resource sharing strategy that maximizes the global satisfaction of all our users?This paper does not completely answer this general question, but focuses on solving the first technical challenge we were facing when trying to build ease.ml. We observe that resource sharing is a critical yet subtle issue in this multi-tenant scenario, as we have to balance between efficiency and fairness. We first formalize the problem that we call multi-tenant model selection, aiming for minimizing the total regret of all users running automatic model selection tasks. We then develop a novel algorithm that combines multi-armed bandits with Bayesian optimization and prove a regret bound under the multi-tenant setting. Finally, we report our evaluation of ease.ml on synthetic data and on two services we are providing to our users, namely, image classification with deep neural networks and binary classification with Azure ML Studio. Our experimental evaluation results show that our proposed solution can be up to 9.8x faster in achieving the same global average accuracy for all users as the two popular heuristics used by our users before ease.ml, and 4.1 x faster than state-of-the-art systems.

/pdf/ease-ml-towards-multi-tenant-resource-sharing-for-machine-3u05j117tm.pdf

Ease.ml: towards multi-tenant resource sharing for machine learning workloads

In accordance with an embodiment, described herein is a system and method for supporting multi-tenancy in an application server, cloud, on-premise, or other environment, which enables categories of components and configurations to be associated with particular application instances or partitions. Resource group templates define, at a domain level, collections of deployable resources that can be referenced from resource groups. Each resource group is a named, fully-qualified collection of deployable resources that can reference a resource group template. A partition provides an administrative and runtime subdivision of the domain, and contains one or more resource groups. Each resource group can reference a resource group template, to bind deployable resources to partition-specific values, for use by the referencing partition. A tenant of the application server or cloud environment can be associated with a partition, or applications deployed therein, for use by that tenant.

System and method for supporting multi-tenancy in an application server, cloud, or other environment

One of the major investment for applying industrial robots in production resides in the software development, which is an interdisciplinary and heterogeneous engineering process. This paper presents a novel model-driven approach that uses AutomationML as modeling framework and ontological reasoning as inference framework for constructing robotic application using Robot Operating System (ROS). We show how different robotic components can be classified and modeled with AutomationML, how these components can be composed together to a production system, and how the AutomationML models can be processed semantically by utilizing Semantic Web technologies and ontological reasoning. By applying model-to-text transformation techniques, executable ROS code can be generated from the models that foster fast prototyping and the reuse of robotic software.

From AutomationML to ROS: A model-driven approach for software engineering of industrial robotics using ontological reasoning

Multi-tenancy is an approach to share one application instance among multiple customers by providing each of them a dedicated view. This approach is commonly used by SaaS providers to reduce the costs for service provisioning. Tenants also expect to be isolated in terms of the performance they observe and the providers inability to offer performance guarantees is a major obstacle for potential cloud customers. To guarantee an isolated performance it is essential to control the resources used by a tenant. This is a challenge, because the layers of the execution environment, responsible for controlling resource usage (e.g., operating system), normally do not have knowledge about entities defined at the application level and thus they cannot distinguish between different tenants. Furthermore, it is hard to predict how tenant requests propagate through the multiple layers of the execution environment down to the physical resource layer. The intended abstraction of the application from the resource controlling layers does not allow to solely solving this problem in the application. In this paper, we propose an approach which applies resource demand estimation techniques in combination with a request based admission control. The resource demand estimation is used to determine resource consumption information for individual requests. The admission control mechanism uses this knowledge to delay requests originating from tenants that exceed their allocated resource share. The proposed method is validated by a widely accepted benchmark showing its applicability in a setup motivated by today's platform environments.

/pdf/resource-usage-control-in-multi-tenant-applications-16nufv1mb6.pdf

Resource usage control in multi-tenant applications

This paper presents a novel ontology-driven software engineering approach for the development of industrial robotics control software. It introduces the ReApp architecture that synthesizes model-driven engineering with semantic technologies to facilitate the development and reuse of ROS-based components and applications. In ReApp, we show how different ontological classification systems for hardware, software, and capabilities help developers in discovering suitable software components for their tasks and in applying them correctly. The proposed model-driven tooling enables developers to work at higher abstraction levels and fosters automatic code generation. It is underpinned by ontologies to minimize discontinuities in the development workflow, with an integrated development environment presenting a seamless interface to the user. First results show the viability and synergy of the selected approach when searching for or developing software with reuse in mind.

A Model-Driven Engineering Approach for ROS using Ontological Semantics

A system includes multiple tenant queues, where each of the queues is associated with a single tenant is configured to queue one or more requests from its respective single tenant. One or more processing nodes have one or more shared resources for processing the requests queued in the multiple tenant queues. A first feedback loop is configured to determine a resource demand for each of the tenants. An admission controller is configured to calculate an actual utilization value of a shared resource for each of the tenants using the knowledge of resource demands for each of the tenants request from the first feedback loop and control processing of the requests from each of the tenant queues based on a reference value for each of the tenants and the actual utilization value of a shared resource for each of the tenants, where the reference value represents an allowed utilization for each of the tenants.

Application level based resource management in multi-tenant applications

to the axiomatic description of components and their features in order to foster the utilization and interoperability between components and applications using ontological domain and task semantics. This work therefore surveys the current state of the art in using semantic technologies and ontology-based knowledge representation frameworks for the description of robotic components and their capabilities. A special emphasis is given to the practical application of such technologies in current robotic projects as well as on initiatives that aim at building industry-wide vocabularies.

A Survey about the Usage of Semantic Technologies for the Description of Robotic Components and Capabilities.

Many companies have a high demand for flexible and economical automation solutions. As the development of such is time-consuming and requires specialized knowledge this usually demands significant investments. The more or less isolated development also prevents sharing and reuse of applications. Due to this, using robot-based automation solutions is often not sensible for medium-sized companies. In order to enable a wider use of robot-based automation solutions in medium-sized companies, we present a common repository of robotics applications to facilitate their reuse. The repository provides a way to find the application one needs, for example based on the requirements of a specific use case and a given situation. The goal of our work is to develop a semantic AppStore with robotic applications (apps) that not only contains a catalog of robotic apps, but also comprises their semantic descriptions and provides semantic search for apps based on their capabilities and properties. This semantic repository of robotic apps is developed in the scope of the ReApp project [3], hence its name ReApp Store. The ReApp Store is implemented based on real-world requirements of the end-user partners of the project consortium with the purpose to offer real-world apps from the domain of robotics.

Nadia Ahmed

Papers

Resource usage control in multi-tenant applications

A Model-Driven Engineering Approach for ROS using Ontological Semantics

Application level based resource management in multi-tenant applications

A Survey about the Usage of Semantic Technologies for the Description of Robotic Components and Capabilities.

ReApp Store - A Semantic AppStore for Applications in the Robotics Domain.