論文

IP層において上位アプリケーションのフレームを考慮するApplication Level Framingは長年の課題である．ルータがIPヘッダ以外の情報をもとにパケットを制御することはスケーラビリティの点で不利である．また同時にルータのパケット中継の作業を複雑にしてしまう恐れがある．本論文で我々は送信ホストはIP Fragmentを積極的に活用することにより，ルータに対してアプリケーションデータの境界線を知らせるべきであると主張する．我々が提案するTransport framE Discard（TED）はルータがパケットを破棄する際に，このIP Fragmentの情報を用いてパケットを破棄する順序を動的に変更させる．TEDを利用することでルータはアプリケーションデータの単位でパケットを破棄することが可能となる．我々はTEDをFIFO，WFQ，RED，Random Drop FIFOに組み合わせて実装し，評価した．その結果，ネットワーク内における全体のパケット欠損率を変化させずに，無駄なパケットの中継を回避し，上位アプリケーションに対して高い通信性能を実現することが実証された．

よりpredictableな計算機環境を構築するために，計算機資源の予約は必須であるが，アプリケーションプログラマからは予約すべき資源量の算定が容易ではないうえに，それが通常は動的に変動してしまうために必ずしも利用は容易ではない．本稿では，従来の資源予約オブジェクトごとにQOSプロファイルハンドラを生成することにより，適切な資源量の見積りを動的に実施し，さらに動的な処理内容やシステム状況の変化への適応性を高めることを目標とした統合的資源管理機構の設計と実装について報告する．本実装は，OS内の低レベル資源予約管理と資源予約調停および翻訳のための資源予約管理サーバ，QOSプロファイルハンドラと動的適応するための非同期イベントメッセージを受信するポートを備えたライブラリにより構成し，GUIインタフェースをサーバに与えることにより，よりユーザに明解で抽象度の高いプログラミング環境を提供することができた．

本論文では，周辺のサービスを利用してユーザの要求を実行する超機能分散環 境を構築する際に重要となる「サービスの代替利用」と「環境変化の検知」を 実現することで，利用可能なサービスが変化してもユーザの要求を継続して実 行できる新たなサービスフレームワークを提案し，それを実現する Adaptive Service Availability Management システム (ASAMA) のプロトタイプを実装 し，評価を行った． ASAMA は，サービスイニシエータ，サービスプロバイダ，ディレクトリサーバ の 3 つのエンティティから構成される．PDA やノート PC などのサービスイ ニシエータは，装備する資源が貧弱なため，周辺のサービスプロバイダからサー ビスを提供してもらい，必要な機能を補完する． ASAMA ではサービスを入出力によって定義し，段階的なサービスの検索を行う ため，代替可能なサービスも発見でき，メディアデータのやりとりでサービス を利用できる．また，環境変化をサービスイニシエータの物理的移動による能 動的環境変化と，サービスプロバイダの起動・終了によって生じる受動的環境 変化に分類し，これらを透過的に検知し，より品質の高いサービスに切り替え られる． 本論文では，本プロトタイプを典型的な超機能分散環境下で実際に動作させ， 目的を実現できるか，実証実験を行い，その利便性を示した．また，各処理に 要する時間や受動的環境変化検知に関するネットワークトラフィックを測定し， ASAMA の基本性能を評価した．

In order to support QoS-aware applications in a mobile Internet environment, it is essential to achieve effective end-to-end QoS control so that these applications can dynamically adapt to various changes in the network resources or environment. In this paper, we describe the design and implementation of our integrated ``End-to-Edge QoS framework'' which consists of mechanism for resource reservation, the QoS translation, and the QoS arbitration.

In this paper, we present a new mobility-support model for applications, Application Module Request Broker (AMRB). We focus on two types of mobility: host mobility and application code mobility. These two types of mobility dynamically change the binding between applications' name and location in the network. AMRB conceals these changes of binding to reduce a complexity in development of applications. In AMRB, we deal with mobile application codes that communicate with each other as Application Modules (Ams). AMRB provides AM's service transparent communication for applications by using a specifier which does not need to include any network location information. Furthermore, applications can use AM's service transparently of migration by exploiting location management mechanism. In this paper, we describe the design and implementation of AMRB and some evaluations. Also, we demonstrate a sample application that AMRB is effective for developing mobile sensor type applications.

Continuous media communication requires timely delivery of data such as digital video and audio packets. Quality of Service (QOS) parameters specify the temporal and spatial characteristic of such continuous media data. To insure timely delivery of continuous media data, the system needs to minimize the communication delay by securing required processor and network resources. We have extended the Capacity-Based Session Reservation Protocol(CBSRP), which was proposed to realizing predictable real-time communications, to support dynamic control of QOS. We have implemented a QOS control scheme by which the network dynamically adjusts the allocations of network bandwidth on a Fiber Distributed Data Interface(FDDI) network. In this paper, we describe the definition of our QOS model, the extension of CBSRP to support dynamic control of QOS, and the implementation and evaluation of extended CBSRP in our distributed systems testbed, Advanced Real-Time System(ARTS).

Modular architectures based on a microkernel are suitable bases for the design and implementation of operating systems. Prototype systems employing microkernel architectures are achieving the levels of functionality and performance expected and required of commercial products. Researchers at Carnegie Mellon University, the Open Software Foundation, and other sites are investigating implementations of a number of operating systems that use the Mach microkernel. The paper describes the Mach microkernel, its use to support implementations of other operating systems, and the status of these efforts.

The real-time object model, an extended object-oriented model for describing real-time systems, is described. The design and implementation of RTC++, a programming language that extends C++ on the basis of the real-time object model, are discussed. The schedulability analysis and specification of rigid timing constraints in systems with active objects are reviewed.

Modular architectures based on a microkernel are suitable bases for the design and implementation of operating systems. Prototype systems employing microkernel architectures are achieving the levels of functionality and performance expected and required of commercial products. Researchers at Carnegie Mellon University, the Open Software Foundation, and other sites are investigating implementations of a number of operating systems (e.g., Unix, MS-DOS) than use the Mach microkernel. The paper describes the Mach microkernel, its use to support implementations of other operating systems, and the status of these efforts.

Workstations for digital audio and video applications must support predictable processing, presentation, and transmission of continuous media data while making efficient use of limited hardware resources. The authors present techniques for managing network resources and processing messages in a predictable fashion. Their solutions use a novel structure for the protocol processing software, and they examine several alternatives and discuss the engineering tradeoffs associated with each. Using simulation, they evaluate their protocol processing structures and show that the more sophisticated structures can improve the response time and predictability of time-constrained messages. The ideas presented in this paper can be applied to resource management in the operating system and to the structuring of application-level servers. The designers of gateway machines (for internetwork routing) which handle both time-constrained traffic and non-time-constrained traffic may also benefit from this study.

Real-Time Mach provides real-time thread and real-time synchronization facilities. A real-time thread can be created for a periodic or aperiodic activity with a timing constraint. Threads can be synchronized among them using a real-time version of the monitor-based synchronization mechanism with a suitable locking protocol. Real-time Mach has several locking policies, such as kernelized monitor, basic priority priority inheritance protocol, priority ceiling protocol, and restartable critical section, for real-time applications. It can also avoid an unbounded priority inversion problem. The authors describe the real-time synchronization facilities in Real-Time Mach and its implementation and performance evaluation. The evaluation results demonstrated that a proper choice of locking policy can avoid unbounded priority inversions and improve the processor schedulability for real-time applications.

In this paper, we present the motivation for using object model for real-time operating systems and we describe the object model used in the ARTS Kernel, a real-time operating system developed in the ART project at Carnegie Mellon University. We discuss our novel object classification and the priority inheritance properties which arise from this taxonomy. We also discuss various methods for implementing critical regions and give some guidelines as to the use of each.

An extension of the uniprocessor Hartstone benchmark for the distributed real-time environment, called the Distributed Hartstone benchmark, is described. The Distributed Hartstone measures system performance in the critical areas of communication latency and bandwidth, protocol preemptability, and priority queueing at the protocol and media access levels. Areas of the system which are particularly important for distributed, real-time computing are described. On the basis of the requirements that specify various areas of the system that a distributed real-time benchmark must stress, a series of task sets in the style of the Hartstone benchmarks are given. The benchmark results from a distributed real-time operating system (ARTS testbed) are given

Distributed real-time systems play a very important role in our modern society. They are used in aircraft control, communication systems, military command and control systems, factory automation, and robotics. However, satisfying the rigid timing requirements of various real-time activities in distributed real-time systems often requires ad hoc methods to tune the system's runtime behavior The objective of Real-Time Mach is to develop a realtime version of the Mach kernel which provides users with a predictable and reliable distributed real-time computing environment. In this paper, we describe a real-time thread model, real-time synchronization, and the ITDS scheduler in Real-Time Mach. We also discuss the implementation issues, a real-time toolset, and the current status of the system.

We propose a new object-oriented programming language called RTC++ for programming real-time applications. RTC++ is an extension of C++ and its features are to specify I) a real-time object which is an active entity, ii) timing constraints in an operation as well as in statements, and iii) a periodic task with rigid timing constraints. In this paper, we first discuss real-time programming issues and what language support should be provided for building real-time applications. Then, the key features of RTC++ are described. Some programming examples are shown to demonstrate RTC++'s expressive power. A comparison to other programming languages are also discussed.

The Mach kernel has been ported to a large collection of uniprocessor and multiprocessor systems and has been used to implement commercial operating system environments such as Unix in a performance competitive way. This paper describes the Mach operating system, the way it was used to implement both 4.3BSD Unix and the Macintosh operating system, and the implications of this technology for real-time and secure programming environments.

In this paper, we address the priority inversion problems in real-time communication and present our solutions developed for the ARTS distributed real-time operating system. We also compare the performance results of our multi-thread based protocol implementation with other implementation.

ARTS is a distributed real-time operating system designed for a real-time systems testbed being developed at Carnegie Mellon University. The objective of the testbed is to develop and verify advanced real-time computing technologies for a distributed environment. The paper gives an overview of the ARTS kernel and ARTS real-time toolset. In particular, it introduces a real-time object model and the integrated time-driven scheduling model. It then describes the basic primitives and major components of the ARTS kernel and the real-time toolset which consists of the schedulability analyzer, Scheduler 1-2-3, and real-time monitor/debugger, ARM.

An integrated real-time tool set is described that consists of a schedulability analyzer and a real-time monitor/debugger for the ARTS kernel, which is being built for a distributed real-time testbed. The schedulability analyzer, called Scheduler 1-2-3, is an X11 window-based interactive tool and can verify whether the given real-time tasks can meet their deadlines under a specific scheduling policy. An advanced real-time monitor, ARM, also runs on the X11 window and visualizes the target system's runtime behavior in real time. ARM can be used with Scheduler 1-2-3 to analyze simulated runs and its monitorability