Real-time task scheduling and network device security for complex embedded systems based on deep learning networks

Abstract

As a hotspot of machine learning research, deep learning is applied in many fields. Embedded systems are becoming more and more complex and networked, so the real-time performance of embedded systems and the security of network embedded devices face severe challenges. Based on this, this paper studies the real-time task scheduling problem for complex embedded systems and the security of embedded network devices. For real-time, this paper proposes a comprehensive task scheduling algorithm. Based on the task classification in the embedded system, different scheduling methods are adopted for different tasks, and the scheduling mode is flexibly changed as the system load changes. A dynamic integrity measurement model is established based on the star trust chain structure, and the hardware implementation mechanism of constructing dynamic trust chain in embedded system is studied. The dynamic reconfigurable hardware design method based on FPGA is applied to the construction of dynamic trust chain, and a verification system is designed to verify the dynamic measurement mechanism. This can solve the security problem of deep network embedded devices to a certain extent.

Introduction

Deep learning has become a hot topic in the field of machine learning research in the past decade. It has made a series of major breakthroughs in practical applications in many fields, such as speech recognition, natural language processing, image and video analysis, computer vision and multimedia. With the development of information technology, especially the popularity of computer network applications, computing platforms have evolved from a loosely coupled infrastructure of point-based isolated computing nodes to massive computing, mobile computing, cloud computing and none with distributed large-scale collaborative features [1,2]. Various organizational forms such as ubiquitous computing, have increased the security requirements for information system architecture, and the requirements for information security have also increased [3,4]. On the other hand, the increased diversity of features and complexity of computing leads to increased vulnerability of information systems and reduced security and reliability [5]. Therefore, in the development trend and security requirements of information systems, it is more and more difficult to realize the overall global security of information systems from top to bottom, and the importance of individual security of computing platforms is particularly prominent [6,7]. Only bottom-up guarantees the security and credibility of individual computing platforms, and the security of large platforms and large systems is guaranteed [8].

For the research of real-time problems in embedded systems, some basic concepts have been formed, such as hard real-time and soft real-time, preemptive scheduling and non-preemptive scheduling, etc. [9]. All related research work is centered around these concepts and some shaped scheduling algorithms have been formed, such as EDF and RM [10,11]. These scheduling algorithms can satisfy some simple embedded applications, but they cannot meet the real-time constraints of the next generation of increasingly complex embedded applications [12]. Research in the field of trusted computing has been ongoing [13]. The world's scientists have gathered various aspects to conduct comprehensive research on the optimization and application of technology, and with the cooperation of industry, university and research, a large number of trusted computing alliances have emerged [14,15]. Follow-up is the industry giant in the world of computer technology, of which Microsoft is the most typical representative, and the world-class institutions of higher learning represented by the Massachusetts Institute of Technology also participated in the alliance [16]. TCG is essentially different from the general alliance organization, because the organization is not a profit structure [17]. It is mainly to strengthen the research on computer security issues, thus promoting the establishment of security standards and making computer security more effective. Control system computers currently used in aerospace and other fields are fault-tolerant computers, and computers used in data storage and transmission are also secure computers based on cryptography [18]. The combination of fault-tolerant technology and security technology can achieve the effect of a trusted computer. At present, the research on reliability technology generally focuses on fault tolerance and testing, especially in the military industry, various products and systems have strict standards for reliability [19]. Some research institutions have invested a lot of energy in reliability technology research, and have experienced many technical problems, and successfully developed different types of fault-tolerant computers, which can be used in the development of high-performance equipment such as aviation and military [20]. In particular, fault-tolerant performance testers have played a role in the practical work of application systems in different fields [21,22]. The security mechanism is an implementation method that clarifies the security of the system. Based on the security policy, the specific implementation method of security protection for the system is studied. Security mechanisms can prevent attacks, detect attacks, or recover from an attack. In order to deal with security threats, an effective security mechanism can be established to ensure the security and reliability of information systems.

This paper first introduces the network layer structure and theoretical characteristics of the convolutional neural network, focuses on the working principle of the convolutional layer and the pooling layer, and does a network structure analysis to understand deep learning and convolutional neural networks from the principle. Secondly, the core technology in the embedded system is elaborated, including the architecture of the embedded operating system. The information security technology is summarized, which is one of the technical foundations for constructing a secure embedded system. This paper proposes an integrated task scheduling algorithm. ITS divides all embedded application tasks into three categories, mission-critical, real-time, and non-real-time. ITS can ensure better real-time performance of the system and achieve higher overall system performance. ITS is aimed at the next generation of intelligent, networked, and universal embedded systems. After further research and improvement, it can be used in all embedded applications. Based on the research of dynamic measurement mechanism in trusted computing, the mechanism of establishing dynamic trust chain on FPGA-based hardware platform and the corresponding measurement model and method are proposed. The trusted root of trusted metric is solved in the design of trusted PC. Combined with the hardware mechanism guarantee provided in the system architecture, a dynamic integrity measurement model is established from the complete trusted measurement mechanism, and the dynamic reconfigurable design method is used to study the dynamic trust chain of the system runtime.

The rest of this article is organized as follows. Section 2 discusses the deep learning and security deep network embedded system, followed by the real-time integrated task scheduling method of the embedded system in Section 3. Section 4 analyzes the embedded system hardware security mechanism of the dynamic trust chain. Section 5 summarizes the full text and points out future research directions.