Software Deployment for Distributed Embedded Real-Time Systems of Automotive Applications

Administration and deployment of software systems become increasingly complex. This complexity results from the need of uniform access to applications from heterogeneous terminals through different communication infrastructures. Moreover, applications consists in complex architectures of lot of small components connected together. A first step to simplify deployment is to have an unified and abstract model for representing deployment dependencies and managing them properly. Therefore, we propose here a deployment model and a conceptual foundation for component installation. Installation dependencies and installation rules are expressed in a logic language.

LHC experiments are depending on a rich palette of software components to build their specific applications. These underlying software components include the ROOT analysis framework, the Geant4 simulation toolkit, monte carlo generators, grid middle-ware, graphics libraries, scripting languages, databases, tools, etc. which are provided centrally in up to date versions on multiple platforms (Linux, Mac, Windows). Until recently this set of packages has been tested and released in a tree like structure as a consistent set of versions across operating systems, architectures and compilers for LHC experiments only. Because of the tree like deployment these releases were only usable in connection with a configuration management tool which provided the proper build and run-time environments and was hindering other parties outside LHC from easily using this palette of packages.In a new approach the releases will be grouped in "flat structure" such that interested parties can start using it without configuration management, retaining all the above mentioned advantages. In addition to an increased usability the software shall also be distributed via system provided package deployment systems (rpm, apt, etc.). The approach of software deployment is following the ideas of providing a wide range of HEP specific software packages and tools in a coherent, up to date and modular way on multiple platforms. The target audience for such software deployments are individual developers or smaller development groups / experiments who don't have the resources to maintain this kind of infrastructure. This new software deployment strategy has already been successfully implemented for groups at CERN.

Как правило, сертифицированные средства обеспечения сетевой безопасности представляют собой узкоспециализированные комплексы с неизменной программно-аппаратной платформой. Главным недостатком такой архитектуры является невозможность прозрачного масштабирования устройств с ростом вычислительных возможностей сети. Кроме того, усложняется разработка и поддержка такого комплекса, так как аппаратная платформа быстро устаревает, что приводит к необходимости её замены, а, значит, и к доработке программных компонентов для поддержки нового оборудования. В работе описывается масштабируемая архитектура комплексов обеспечения сетевой безопасности, которая позволяет производителям упростить процесс обновления и разработки средств защиты информации. Главным свойством новой архитектуры является ориентированность на предоставление совокупности специализированных микро-сервисов. Она строится на принципах виртуализации сетевых функций и использует понятие унифицированной доверенной программно-аппаратной платформы. Каждая сетевая функция запускается на некоторой программно-аппаратной платформе, работающей под управлением операционной системы с гипервизором. Ясно, что в случае сертификации по требованиям безопасности конечных продуктов, необходимо обеспечить доверие к аппаратной платформе, операционной системе и гипервизору. Однако архитектура требует унификации программно-аппаратной платформы для всех сетевых функций. Это упрощает разработчикам поддержку конечных продуктов. Благодаря единой доверенной платформе архитектура позволяет упростить процедуры сертификации по требованиям безопасности информации в процессе поддержки и развития конечного продукта. Балансировка нагрузки и согласованность архитектуры обеспечивается средствами, реализующими технологию децентрализованных распределённых реестров (блокчейн). As a rule, certified means of ensuring network security are highly specialized complexes with a constant hardware and software platform. The main disadvantage of this architecture is the impossibility of transparent scaling of devices when the computing power of the network increases. In addition, the development and support of such a complex is complicated, as the hardware platform quickly becomes obsolete, which results in the necessity for replacing it, which means that the software components for the support of new equipment need to be improved. The paper describes the scalable architecture of network security systems, allowing manufacturers to simplify the process of updating and developing information security tools. The main feature of the new architecture is the focus on the provision of a set of specialized micro-services. It is based on the principles of virtualization of network functions and it uses the concept of a unified trusted software and hardware platform. Each network function runs on a hardware and software platform running by the hypervisor operating system. It is clear that in the case of certification for the security requirements of the final products, it is necessary to ensure confidence in the hardware platform, operating system and hypervisor. However, the architecture requires unification of the software and hardware platform for all network functions. This makes it easier for developers to support end-products. Thanks to a single trusted platform, the architecture allows simplifying certification procedures for information security requirements in the process of supporting and developing the final product. Load balancing and architecture consistency are provided by means of implementing the technology of decentralized distributed registries (blockchain).

SummaryThe execution of agent‐based simulations (ABSs) on hardware accelerator devices such as graphics processing units (GPUs) has been shown to offer great performance potentials. However, in heterogeneous hardware environments, it can become increasingly difficult to find viable partitions of the simulation and provide implementations for different hardware devices. To automate this process, we present OpenABLext, an extension to OpenABL, a model specification language for ABSs. By providing a device‐aware OpenCL backend, OpenABLext enables the co‐execution of ABS on heterogeneous hardware platforms consisting of central processing units, GPUs, and field programmable gate arrays (FPGAs). We present a novel online dispatching method that efficiently profiles partitions of the simulation during run‐time to optimize the hardware assignment while using the profiling results to advance the simulation itself. In addition, OpenABLext features automated conflict resolution based on user‐specified rules, supports graph‐based simulation spaces, and utilizes an efficient neighbor search algorithm. We show the improved performance of OpenABLext and demonstrate the potential of FPGAs in the context of ABS. We illustrate how co‐execution can be used to further lower execution times. OpenABLext can be seen as an enabler to tap the computing power of heterogeneous hardware platforms for ABS.

Large-scale reuse of software architectures and components is an important element for improving software productivity and reducing the time needed to design, develop, and deploy software systems. For maximum benefit, the application of reuse technologies must be coupled with effective process changes. In this paper, we focus on AT&T's approach to software reuse, particularly the use of the AT&T BaseWorX™ applications platform to achieve reuse. We also discuss the process changes associated with software reuse, and the results that have been achieved from applying reuse to a wide range of software products.

Heterogeneous multi-core platforms that contain different types of cores, organized as clusters, are emerging, e.g., ARM’s big.LITTLE architecture. These platforms often need to deal with multiple applications, having different performance requirements, executing concurrently. This leads to the generation of varying and mixed workloads (e.g., compute and memory intensive) due to resource sharing. Run-time management is required for adapting to such performance requirements and workload variabilities and to achieve energy efficiency. Moreover, the management becomes challenging when the applications are multi-threaded and the heterogeneity needs to be exploited. The existing run-time management approaches do not efficiently exploit cores situated in different clusters simultaneously (referred to as inter-cluster exploitation) and DVFS potential of cores, which is the aim of this paper. Such exploitation might help to satisfy the performance requirement while achieving energy savings at the same time. Therefore, in this paper, we propose a run-time management approach that first selects thread-to-core mapping based on the performance requirements and resource availability. Then, it applies online adaptation by adjusting the voltage-frequency ( V-f ) levels to achieve energy optimization, without trading-off application performance. For thread-to-core mapping, offline profiled results are used, which contain performance and energy characteristics of applications when executed on the heterogeneous platform by using different types of cores in various possible combinations. For an application, thread-to-core mapping process defines the number of used cores and their type, which are situated in different clusters. The online adaptation process classifies the inherent workload characteristics of concurrently executing applications, incurring a lower overhead than existing learning-based approaches as demonstrated in this paper. The classification of workload is performed using the metric Memory Reads Per Instruction (MRPI). The adaptation process pro-actively selects an appropriate V-f pair for a predicted workload. Subsequently, it monitors the workload prediction error and performance loss, quantified by instructions per second (IPS), and adjusts the chosen V-f to compensate. We validate the proposed run-time management approach on a hardware platform, the Odroid-XU3, with various combinations of multi-threaded applications from PARSEC and SPLASH benchmarks. Results show an average improvement in energy efficiency up to 33 percent compared to existing approaches while meeting the performance requirements.

The Linux operating system has become the platform of choice in the HEP community. However, the migration process from another operating system to Linux can be a tremendous effort for developers and system administrators. The ultimate goal of such a transition is to maximize agreement between the nal results of identical calculations on the different platforms. Apart from the ne tuning of the existing software, the following issues need to be resolved: choice of Linux distribution, development tools (compiler, debugger, prolers etc.), compatibility with third-party software, and deployment strategy. It would be ideal to develop, run and test software using ofce desktops, local farm systems, or personal laptops regardless of the Linux distribution choosen. To accomplish this task, you need to have a e xible package management system that is able to install/upgrade/verify/uninstall the necessary software components without particular knowledge of remote system conguration and user privileges. We discuss how Linux became the third ofcial computing platform of the CLEO collaboration, outlining the details of the transition from Tru64 and Solaris operating systems to Linux, and the software model and deployment strategy employed.

The next generation of software systems will be highly distributed, component-based, service-oriented, will need to operate in unattended mode and possibly in hostile environments, will be composed of a large number of "replaceable" components discoverable at run-time, and will have to run on a multitude of unknown and heterogeneous hardware and network platforms. This paper focuses on service oriented-architectures in which each component provides a set of interrelated services to other components. These components are QoS-aware (i.e., aware of Quality of Service requirements) and are capable of engaging in QoS negotiations with other components of a distributed application. The main contributions of this paper are: i) the description of an architecture for QoS-aware software components that are able to negotiate QoS requirements with other components, ii) the specification of the protocols used for QoS negotiation and admission control at the QoS-aware components, iii) a report on the implementation of a QoS-aware component, and iv) the experimental validation of the ideas presented in the paper.

A recent development of heterogeneous platforms (i.e. those containing different types of computing units such as multicore CPUs, GPUs, and FPGAs) has enabled significant improvements in performance processing large amount of data in realtime. This possibility however is still not fully utilized due to a lack of methods for optimal configuration of software; the allocation of different software components to different computing unit types is crucial for getting the maximal utilization of the platform, but for more complex systems it is difficult to find ad-hoc a good enough or the best configuration. In this paper we present an approach to find a feasible and locally optimal solution for allocating software components to processing units in a heterogeneous platform.

We introduce an initial implementation of a real-time component-based software framework for autonomous mobile robots. We argue that real-world autonomous mobile robots shall be controlled by self-contained software systems able to meet hard timing constraints. The proposed solution empowers specialized roboticists to contribute software components that can be integrated into complete real-time systems. The framework also facilitates robotic software components reuse and portability across hardware platforms. Based on an indoor navigation case study we evaluate the advantages and the limitations of the framework in terms of ease of use, modularity and real-time capabilities.

Multi-tenant inference, as a prevalent inference paradigm nowadays, requires deploying multiple deep learning models on the hardware platform to concurrently process inference tasks. Modern platforms are typically equipped with various heterogeneous processors, such as CPU-GPU platform. To reduce resource contention and improve Quality of Service (QoS) in the multi-tenant scenario, existing work has studied cross-processor inference at the model- and layer-level. However, coarse-grained scheduling cannot flexibly account for subtle resource fluctuations, which may lead to task blockages and incur significant processor switching overheads. Such work usually requires extensive modification and retraining of the models. Therefore, we propose a finer-grained operator-level cross-processor scheduling framework COS, which can more precisely divide the computational workloads and switching overheads for the tenants, without modifying or retraining. We introduce a novel intermediate representation to abstract and simplify the scheduling problem, and propose an efficient two-phase search algorithm. COS is automated and easy-to-scale, through experiments on various heterogeneous hardware platforms and models, we demonstrate that COS is more flexible and effective than layer-level scheduling, and achieves higher throughput than single-processor processing in the multi-tenant scenario. Furthermore, COS is an offline optimization method, and its overhead is highly acceptable.

The implementation and continuous support of a software product at a customer with evolving requirements is a complex task for a product software vendor. There are many customers for the vendor to serve, all of whom might require their own version or variant of the application. Furthermore, the software application itself will consist of many (software) components that depend on each other to function correctly. On top of that, these components will evolve over time to meet the changing needs of customers. To alleviate this problem we propose to alleviate the software release and deployment effort and reduce risks associated with it. This will be achieved by explicitly managing typical knowledge about the software product, such as configuration and dependency information, thereby allowing software vendors to improve the customer configuration updating process. The proposed solution of knowledge management at both the customer and vendor site, is validated through industrial case studies.

A new information integration platform for computer integrated manufacturing system (CIMS) is presented, which is based on agent and CORBA. CORBA enhances the system integration because it is an industry-standard for interoperable, distributed objects across heterogeneous hardware and software platform. Agent technology is used to improve intelligence of integration system. In order to implement the information integration platform, we use network integration server to integrate network, design a generic database agent to integrate database, adopt multi-agent based architecture to integrate application, and utilize wrapper as CORBA object to integrate legacy code.

A survey on routing protocols supported by the Contiki Internet of things operating system

The low robustness of homogeneous controllers deployment in software defined network can reduce the cost of attacks on the network, and the network is more susceptible to virus infection and transmission, which causes network paralysis and seriously jeopardizes network security and availability. In this paper, we consider introducing diversity into controller deployment problem, and propose to use the existing controller diversity to enhance the robustness of the control plane and establish a highly robust controller deployment model. Then, based on the application of robustness and diversity in the network, the robustness measurement method in the network is given. Finally, the controller deployment method is designed to meet the requirements of network delay, performance and security cost. The experimental results show that the proposed method can significantly improve the robustness of the network at a certain security cost.