DuDE-Cloud: A resilient high performance cloud Danielis, Peter; Skodzik, Jan; Altmann, Vlado ...
2015 IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA),
09/2015
Conference Proceeding
In recent years, cloud computing, which provides users with network resources such as memory and computing power depending on the users' needs, has gained enormously in importance. Providers such as ...Amazon allow the users to access their cloud storage and computing resources by different interfaces. In this regard, guaranteeing compatibility is a severe issue as many different proprietary interfaces exist for accessing data in different clouds. The solution to this problem is the RESTful Cloud Data Management Interface (CDMI) standard, which has passed the ISO audit as first standard and is therefore evolving into the most common standard for accessing clouds. This paper investigates the combination of CDMI with a P2P-based storage and computing back-end in order to realize a self-organizing cloud for distributed data storage and processing called DuDE-Cloud. The basis is the self-organizing hash table (DHT)-based P2P storage and computing back-end called DuDE, which utilizes the DHT protocol Kad. DuDE is able to bundle existing storage and computing resources of different devices dynamically and has already proven its advantageous performance over centralized solutions. In a test scenario, resources of DuDE-Cloud were successfully accessed through CDMI using a GUI front-end thereby proving the proper functionality. As a result, a resilient high-performance distributed cloud solution is available that is compatible with other clouds.
An optimized WS-eventing for large-scale networks Skodzik, Jan; Altmann, Vlado; Danielis, Peter ...
Proceedings of the 2014 IEEE Emerging Technology and Factory Automation (ETFA),
2014-Sept.
Conference Proceeding
Web Services are becoming more and more relevant also in the domain of embedded devices as they are becoming an important aspect of the Internet of Things. Embedded devices, especially in the field ...of automation, require real-time behavior. The Devices Profile for Web Services defines WS-Standards for embedded systems. The WS-Eventing is one of this standards within DPWS and enables the distributions of events. However, WS-Eventing has a scalability problem as in ad-hoc networks the notification is transmitted sequential by an event source. Under these circumstances, it is not suitable for large-scale networks with a high amount of devices. Therefore, a new approach is presented to solve this issue by acquiring helping devices. The communication is only based on unicast IP communication and thus is easy to integrate into automation infrastructures and could be established over several subnets. This approach has been investigated with an experimental setup and shows the high scalability compared to the original notification mechanism. Additionally, the real-time capability is given due to the chosen platform and real-time operating system. Thus, the new approach enables the idea of Internet of Things in automation scenarios.
Real-time systems play a major role in the realm of industrial automation. It is predicted for the number of smart interconnected devices that participate in such systems to grow significantly in the ...future. This development is also referred to as Industrial Internet of Things (IIoT) or Industry 4.0. The high number of devices results in highly distributed applications. Therefore, it is no longer sufficient for each device to be real-time capable. In fact, the influence of the communication on the overall timing behavior of the applications grows. Taking this into account, a variety of real-time capable Ethernet approaches called Industrial Ethernet (IE) emerged. However, the established IE solutions rely on proprietary hardware and/or non standard conform protocol adaptations. This leads to very expensive hardware and incompatibilities with other IE solutions or common Ethernet, and thus degrades the interoperability. HaRTKad describes a purely software-based P2P approach that allows deterministic communication over common Ethernet. Although it solely relies on well-known standards and allows real-time communication, it suffers from a multitude of problems. In this paper, the low network utilization, the handling of hash collisions and the traffic prioritization are revealed as the most significant limitations of HaRTKad and possible solutions to these problems are presented.
Emulation of SDN-supported automation networks Danielis, Peter; Altmann, Vlado; Skodzik, Jan ...
2015 IEEE 20th Conference on Emerging Technologies & Factory Automation (ETFA),
09/2015
Conference Proceeding
Software-defined networking (SDN) is a principle for the flexible configuration of networks, which recently has aroused an increasing interest of researchers and companies. Today, the operation and ...configuration of networks as well as their adaptation to changing requirements represent a major challenge if legacy network management protocols are used due to their inability to provide network-wide configuration. Compatibility issues with legacy protocols or even proprietary protocols have further contributed to hardly reconfigurable and inefficient networks and thus prepare the way for the new concept called SDN. SDN implementations are mainly in the data center today. Implementations will find their way into broader networking applications over the next few years. However, even automation networks with highest QoS requirements can benefit from SDN support by achieving significantly more devices that meet these requirements. Therefore, this paper addresses the emulation of SDN-supported automation networks to examine possible design options. Limitations and possibilities when using the popular and widely accepted SDN emulator Mininet are analyzed.
The P2P-based system for the distributed storing of session data of Internet service providers' access nodes called P2P-based storage platform is presented. Session data is continuously changing due ...to customers connected to access nodes, i.e., it is highly volatile. However, it has to be stored persistently as it is required for data forwarding, traffic filtering, and deriving statistics. Failing access nodes must be able to restore their data after reentering the network. Today, the session data is stored in the access nodes' flash memory, which is limited in life time and size and intended for other purposes. Contrary, the P2P-based storage platform allows to store session data in the access nodes' available RAMs connected by the distributed hash table-based P2P network Kad. A Kad network consisting of 8,000 access nodes for distributed storage of session data is setup for simulation. The simulation results show the traffic overhead to be minimal, linear scalability, and prove the high availability of the stored session data.
Device and service discovery is an essential part of automation networks in order to ensure the interoperability of devices and Plug&Play functionalities. Embedded Web services such as Devices ...Profile for Web Services are an upcoming technology for building and factory automation. For ad hoc discovery, the WS-Discovery standard is used. This approach avoids single points of failures. However, it shows low scalability due to static response time values. The rising number of devices leads to an increasing network load and especially to an increasing device processing load as a result of the discovery process. In this work, we evaluate the issues of ad hoc service discovery and suggest an optimized algorithm for dynamic data rate reduction in large scale networks. Moreover, the suggested improvements allow reducing peak data rates by more than 50% for constant timing parameters. Using the suggested approach, the ad hoc discovery can be applied in any network size. The desired peak data rates or response times of the system can be significantly reduced and dynamically adjusted according to the desired values. This approach guarantees a scalable ad hoc discovery mechanism and is backward compatible with legacy equipment.
The rising number of IoT systems deployed in production increases the risk of becoming a victim of costly attacks. The complex architecture of such systems increases their attack surface making the ...systems more vulnerable by attacks that can even paralyze production. For 2020, it is estimated that more than 30 billion IoT devices are in use worldwide; of which more than 5.8 billion in the industrial sector. It is therefore imperative today to have a threat modeling tool that examines all system components and assigns them a risk of being exposed. This paper proposes a novel ISO-compliant test procedure for the technical risk analysis of IoT systems since state-of-the-art methods are either not ISO-compliant or are not suitable for technical risk analyses. It is based on the generic threat model Microsoft STRIDE, which has been developed to consider and evaluate all system components. The test procedure is able to identify vulnerabilities and assigns a risk to them. It complies with ISO/IEC 27001 and implies ISO/IEC 27005 and ISO 31000. The threats are stored in a threat database and linked to countermeasures. The designed tool implementing the novel test procedure automatically proposes countermeasures for certain attacks to the user. The tool uses a learning database in which new insights can be entered during each risk analysis so that the tool is improved with each use. As proof of concept, an IoT system is analyzed for its risks and the high achievable accuracy of the proposed risk assessment is shown.
In this paper, a highly integrable Field Programmable Gate Array-based hardware design of multilayer perceptron as a realization of an artificial neural network is presented. Such a hardware solution ...ensures a deterministic behavior required for any hard real-time compositions. The integration into existing systems is achieved by the application of UDP/IP. %A developed protocol enables the hardware solution to act as a stand-alone device with no need for an additional host PC. Additionally, the presented design is highly flexible due to a parameterizable multilayer perceptron approach. However, most reconfigurations usually require a hard coded reimplementation, resynthesis, and the download of a new bit file to the target platform, which also requires an additional host PC. Contrary with the presented solution, it is possible to configure the multilayer perceptron's parameters during runtime via a software interface. This approach allows the multilayer perceptron to be adapted to nearly any application. The developed design combines the flexibility of a software solution to generate and comfortably reconfigure the multilayer perceptron as well as the high performance of a hardware solution. %The investigation of hardware utilization and performance of a running prototype stützt%Finally, the hardware utilization and performance are investigated. As proof of concept, a running prototype has been realized, which shows the design to be highly flexible and with good performance while the hardware resource consumption is kept minimal.
Clock synchronization protocols such as the precision time protocol (PTP), which are used to synchronize components of distributed systems, are fundamental to enable timed and coordinated activities, ...e.g., in real-time applications within the industrial Internet of things (IIoT). In theory, PTP is able to achieve a precision on the order of nanoseconds. However, its practical accuracy remains limited by packet delay variations. In this paper, we hence present a novel approach to increase the synchronization precision of PTP. Our approach (PTP-LP) relies on PTP to obtain precise hardware timestamps taken during multiple synchronization periods. These timestamps establish the constraints for a Linear Programming (LP) solver that is used to estimate the clock differences between devices. Moreover, we propose the heuristic PTP-H that achieves comparable accuracy but is less computationally complex. We evaluate PTP-LP and PTP-H in comparison with two state-of-the-art approaches under various conditions in terms of clock stabilities and packet delay distributions. PTP-LP and PTP-H are fully compatible with existing standards and show to be in particular robust to varying packet delays. Especially, PTP-LP outperforms previous approaches in presence of a stable hardware clock and unknown non-negligible network delay, which are both realistic working conditions.
Future applications of the Industrial Internet of Things will increasingly depend on the timely exchange of information and data. Time-Sensitive Networking, which is currently being standardized, ...therefore extends switched Ethernet networks by realtime communication capabilities offering deterministic message delays. In its time-triggered communication variant, a network path needs to be planned for each realtime data flow and corresponding transmission time slots have to be reserved on the network links along that path. In this paper, we present the first joint mathematical model for path routing and time slot scheduling that combines multicast support with an individual transmission scheduling on each link along a flow's path. We provide a formalization as an Integer Linear Programming (ILP) problem and discuss several optimizations to significantly reduce the ILP solver's runtime without affecting the solution quality. In a thorough evaluation, we show the applicability of our approach and analyze the effects and trade-offs of different objective functions.
