The Directed Acyclic Graph (DAG) task model for real-time scheduling finds its primary practical target in Robot Operating System 2 (ROS 2). However, ROS 2's publish/subscribe API leaves DAG precedence constraints unenforced: a callback may publish mid-execution, and multi-input callbacks let developers choose topic-matching policies. Thus preserving DAG semantics relies on conventions; once violated, the model collapses. We propose the Function-as-Subtask (FasS) API, which expresses each subtask as a function whose arguments/return values are the subtask's incoming/outgoing edges. By minimizing description freedom, DAG semantics is guaranteed at the API rather than by programmer discipline. We implement a DAGnative scheduler using FasS on a Rust-based experimental kernel and evaluate its semantic fidelity, and we outline design guidelines for applying FasS to Linux sched_ext.

As autonomous driving systems mature, the importance of fault tolerance to provide safe operation even in hazardous conditions becomes a key issue in deployment. In an autonomous driving system composed of redundant devices, it is necessary to maintain consistency across the system regarding which devices should currently be handling processing. In this paper, we present a new leader election algorithm for autonomous driving systems that guarantees consistency in the presence of the failure of any single system component. To make the algorithm practical for autonomous driving systems, we design it to both handle network link failures and guarantee liveness in the presence of a single failure. Due to the safety-critical nature of our environment and the known difficulty of testing distributed algorithms, we use formal verification to validate our proposed algorithm. Our experiments show that the algorithm completes with reasonable time in a simulated environment of an autonomous driving system.

Many IoT devices are compromised by exploiting their software vulnerabilities in these devices. A technique during a development phase to fix such vulnerabilities efficiently is leveraging advantages of various platforms (e.g., single-board computers, emulators, and FPGA (Field Programmable Gate Array) boards). As a debugging scenario, developers discover some vulnerability-candidates of software under test on an emulator like QEMU with rich debugging plugins. They then check those candidates in detail on a single-board computer (i.e., a real device) with GDB. To efficiently conduct this kind of debugging, we propose NEMIANA (Noncomplex Execution Migration Integrated Architecture for Nonstop Analyses), an architecture for automatically conducting cross-platform execution migration. It can migrate execution-state of software at any execution points (e.g., 500 steps after the beginning) from a platform (e.g., a single-board computer) to another platform (e.g., a FPGA board). Developers then seamlessly continue debugging at that execution point on the destination platform. A key idea for cross-platform migration is an abstraction model of execution-state (e.g., CPU register/memory values), which can be applied to various platforms (if they have CPUs of the same ISA (Instruction Set of Architecture)). Another key idea is to trace all instructions executed by software. This incurs a system performance overhead; however, it enables developers to choose any execution points, which we hope is very helpful for debugging. In the evaluation, a prototype system of NEMIANA demonstrates three case studies: a typical debugging with GDB, benchmarking, and vulnerability discovery. It shows developers can use NEMIANA for their debugging purposes.

In security operations such as incident response and malware analysis, a large number of logs are gathered from a monitoring network. These logs include security appliance alerts and network communications. Security operators must then find remarkable logs from all these logs. To filter out benign logs from all the logs, security operators commonly use a whitelist, which is a set of benign hosts decided upon by the security operators. When creating a whitelist, security operators typically refer to website-ranking services, the features of a monitoring network topology, their knowledge and expertise about the monitoring network, etc. In contrast, a whitelist should be continuously and manually updated by the security operators; thus, maintaining the whitelist places a higher burden on them. Therefore, in this paper, we propose Byakko, which is a method for automatically generating a whitelist based on statistical features, that is, occurrence distribution and its standard deviation. These features describe the stationarity of communications in a monitoring network. If a host has stationary communications, Byakko decides that the host must be benign. To evaluate Byakko, we conduct a performance evaluation and two case studies, which are for incident responses on an office network and for malware analysis on an analysis network. Our results show that Byakko is a practical method and can be applied to other cases as well.

Third-party Web tracking is a serious privacy issue. Advertisement sites and social networking sites stealthily collect users' Web browsing history for purposes such as targeted advertising or predicting trends. Unfortunately, very few Internet users realize this, and their privacy has been infringed upon because they have no means of recognizing the situation. In this paper we present the design and implementation of a system called MindYourPrivacy that visualizes third-party Web tracking and clarifies the entities threatening users' privacy. The implementation adopts deep packet inspection, DNS-SOA-record-based categorization, and HTTP-referred graphical analysis to visualize collectors of Web browsing histories without device dependency. To demonstrate the effectiveness of our proof-of-concept implementation, we conducted an experiment in an IT technology camp, where 129 attendees discussed IT technologies for four days, The experiment's results revealed that visualizing Web tracking effectively influences users' perception of privacy. Analysis of the user data we collected at the camp also revealed that MCODE clustering and some features derived from graph theory are useful for detecting advertising sites that potentially collect user information by Web tracking for their own purposes.

Users want to enjoy online services without sacrificing their security. Although there is a trade-off between the security of a service and its usability, the level of security required will differ depending on the user and the situation. To optimize the balance between security and usability, it can be customized for each user and each online transaction. Yet in order to do that, both users and service providers need to stipulate their security requirements. We have been working on a framework that provides security requirement classifications in multiple dimensions to help users identify and select their security requirements, and then apply these requirements to different dimensions. This paper shows how we implemented this framework and then evaluated it by conducting a user study along with our implementation. The study verifies that ordinary users without any particular technical knowledge prefer to clarify their security requirements using a taxonomy-based selection scheme (our scheme) as opposed to a free-form input scheme. It also discusses the coverage of pre-defined taxonomies and users' requirements. Through this study, we clarify the future direction of our research.

Distributed-hash-table (DHT) has been proposed to solve the problem of scaling for P2P networks. However, there are some problems of security in P2P networks. One of the problem is that key exchange is not performed preventing against endangering security. This reason may be the fact that P2P networks have no trusted server. Although certification and authentication are able to protect key exchange from spoofing and man-in-the-middle attacks, these cannot be applied because of that. Therefore, an attacker can easily compromise key exchange since P2P networks should accept any node whether it is malicious or not, and every node of P2P networks plays a role of router. Until today, for the above reasons, there is no way to exchange the secret key on P2P networks protecting against encountering security. In this paper, we propose a key exchange method on P2P networks protecting against spoofing and man-in-the-middle attacks. For the purpose of protecting key exchange from encountering security, we present the enhanced routings which are directional and probabilistic routings.