システム・ソフトウェア
System Software / Operating Systems

Outline of This Lecture

• Review of Students' Programming Backgrounds
• Course Outline
• Prerequisites
• Grading Criteria and Requirements
• Contact Information, Office Hours and Class Notes
• Operating System Principles
• Operating System History
• Computing Today
• Homework

Review of Students' Programming Backgrounds

Take a sheet of paper. On it, put your name and student ID number. Then answer the following questions:

1. Do you currently have a working C compiler on your system?
   現在、C言語のコンパイラーを持っていますか。
2. List the operating systems and versions you have used:
   今までの使ったことがあるオペレーティングシステムをリストアップしてください。
   Microsoft Windows, MacOS, Linux, FreeBSD, Solaris?
3. How many years ago did you start programming?
   何年前にプログラ ミングを始まったでしょう。
4. Have you administered a multi-user computer system before, such as a web server or file server?
   マルチユーザーのコンピューター（ウェブサーバー、ファイルサーバーなど）を管理したことありますか。
5. Have you managed the following services:
   以下のサービスを管理したことありますか。
   networking, file, web, compute, DNS, public key?
6. List the languages you have programmed in:
   今までの使ったことがある言語をリストアップしてください：
   assembler, BASIC, C,　C++, Fortran, Java, Lisp, Perl, Python, Ruby, sh, ...?
7. Which tools have you used:
   このリストの中にはどのツールを使ったことがある：
   make, git, svn, CVS, RCS, Microsoft VisualC, etc.,
   automake, grep, sed, awk, lex/flex, yacc/bison, Spirit, ...?
8. What is the longest program you have written, measured in lines?
   ひとつのプログラムは何行まで書きましたか。
9. When you write code, do you write comments?
   プログラムを書いているとき、コメントを書きますか。
10. Have you worked in a team to write one program? What source code control system did you use? How did you divide up the work, and how did you define and test interfaces between members of the team?
    一 人以上のグループでプログラムを書いたことありますか。ソースコードコント ロールはどうやってしましたか。仕事がどうやって分割しましたか。インター フェースの定義のしかたには何をしましたか。

Course Outline

教科書 Textbook

Lecture by Lecture

• 第1回 4月9日 OSの歴史と基本概念
  Lecture 1, April 9: OS History and Basic Principles
  • Levin and Redell, An Evaluation of the Ninth SOSP Submissions -or- How (and How Not) to Write a Good Systems Paper
  • 配布資料 Lecture Notes
• 第2回 4月16日 システムコール
  Lecture 2, April 16: System Calls
  • Lampson, Hints for Computer System Design
  • 配布資料 Lecture Notes
• 第3回 4月23日 プロセスとスレッド
  Lecture 3, April 23: Processes and Threads
  • 配布資料 Lecture Notes
• 第4回 5月7日 プロセス間通信
  Lecture 4, May 7: Inter-process Communication
  • 配布資料 Lecture Notes
• 第5回 5月14日 プロセススケジューリング
  Lecture 5, May 14: Process Scheduling
  • 配布資料 Lecture Notes
• 第6回 5月21日 メモリ管理と仮想記憶
  Lecture 6, May 21: Memory Management and Virtual Memory
  • 配布資料 Lecture Notes
• 第7回 5月28日 ページ置換アルゴリズム
  Lecture 7, May 28: Page Replacement Algorithms
  • 配布資料 Lecture Notes
• 第8回 6月4日 ネットワーク＆セキュリティ
  Lecture 8, June 4: Networking & Security
  • 配布資料 Lecture Notes
• 第9回 6月11日 ファイルシステム
  Lecture 9, June 11: File Systems
  • 配布資料 Lecture Notes
• 第10回 6月18日 ファイル・システムの実装
  Lecture 10, June 18: File System Implementations
  • 配布資料 Lecture Notes
• 第11回 6月25日 入出力
  Lecture 11, June 25: Input/Output Systems
  • 配布資料 Lecture Notes
• 第12回 7月2日 ハイパーバイザー
  Lecture 12, July 2: Hypervisors
  • 配布資料 Lecture Notes
• 第13回 7月9日 ライルタイムと組み込み用のOS
  Lecture 13, July 9: Real-time and embedded OSes
  • 配布資料 Lecture Notes
• 第14回 7月16日 OS事例研究
  Lecture 14, July 16: Operating Systems Research
  • 配布資料 Lecture Notes

前提知識
Prerequisites

• The basic hardware parts of a computer:
  • CPU
  • memory
  • disk drive
• Programming skills: you must be able to program in some language, preferably C.
• You must know the basic concept of memory addresses (in C, pointers).
• You must be comfortable with hexadecimals and binary manipulation of numbers.

Requirements

成績の仕方
Grading

• 授業の討論/Class participation: 10%
• 宿題/Homework: 40%
• 学期のプロジェクト/Term project: 50%

This year, heavier emphasis will be placed on the term project than on homeworks. You should expect your project to take 40-60 hours. There will be five homework assignments over the course of the semester, and readings almost every week. At least one of the programming assignments will involve parallel programming.

Term Project

ソフトウエアプロジェクトは必要です！一人から三人までできます。来週は説 明と去年のプロジェクトの話です。

You will complete a moderately large software project during the term. Projects may be individual, or teams of up to three students. We will talk about ideas for specific projects next week. The following schedule applies:

• Project and team proposals due: April 23
• Review of proposals returned: May 7
• Revised project proposals due: May 14
• Final approval of projects, implementation begins: May 21
• Mid-term progress review: June 18
• Final evaluation of projects (face-to-face): July 16-26

Contacting Me/Office Hours/Class Notes
連絡先/オフィスアワー

Office Hours, Spring 2013春のオフィスアウアー：Monday （月曜日）, 10-12, Delta N211. You may come to my office during this time without an appointment. If you wish to see me otherwise, you can attempt to find me directly, or send me email to arrange an appointment.

The lecture notes for each week are posted in SFC-SFS shortly before the lecture.

Operating System Principles
オペレーティングシステムの概念

ふたつの概念：全ての情報は分散である、とすべてのことは同時でやる。 特殊相対性理論はソフトウエアでも影響する。何かを処理しようとすると き、必要なデータが保存してあるところから現在地にはこんでこなければいけ ない。ほかのところでほか関数されることも考えておかなければいけない。

You must consider the light cone of information.

光円錐をかんがえなければならない。

オペレーティングシステムの定義として、Tanenbaumの観点はふたつである： オペレーティングシステムはリソースマネジャ、とオペレーティングシステム は拡張マシンです。私の意見で、もうふたつのこと同じぐらいのレベル重要だ と思います：ネーミングとデータの転送。

• Resource management/リソース管理
  • Synchronization: correctness, efficiency, deadlock avoidance
  • Fair and/or appropriate allocation of resources
  • Security
• Extension of the machine/拡張マシン
  • Expand memory through virtual memory
  • Expand non-volatile storage through file systems, volume managers, and network/distributed file systems
• Naming/ネーミング
  Naming is one of the principle functions of all computer systems.
  • Storage: registers, cache, main memory, disk memory, removable devices
  • Hardware: devices, especially dynamic ones (USB, iSCSI, etc.)
  • Virtual objects: processes, pipes
  • Data: URLs, URIs
  • Sharing names
• Data movement/データ転送
  Generally, user programs do not move data directly; they must ask the operating system to do it for them.
  • Moving data: buses, networks
  • Controlling the movement: registers, RPC, third-party transfers, etc.
  • Performance: data layout, resource allocation at multiple levels (e.g., I/O bus and memory bus)
  • Real-time constraints: buffering, jitter, missed revolutions, tape stalls

These will be considered, along with papers on how to structure a system. Although it is important to understand a system as thoroughly as possible, students will also begin to recognize the value of being able to use, modify, and debug systems they do not completely understand, and how to build systems in which this is possible.

Operating System History
オペレーティングシステムの歴史

In the mid-twentieth century, there was much excitement over the design of digital computers. Early important names include John von Neumann, Konrad Zuse, J. Presper Eckert, John William Mauchly, and Maurice Wilkes; I encourage you look up those names and study their contributions. The term "von Neumann architecture" is generally used to refer to a computer that can store a program as a form of data, rather than in the wiring of the computer itself, though it is generally acknowledged that all of the above contributed to the basic ideas. Without this innovation, the field of computer software would be very dry!

By the mid-1950s, computers made of transistors and capable of storing their programs were becoming common. At first, each program controlled the entire computer; it was like reformatting your hard disk each time you wanted to run a different program -- except, of course, this was before the invention of the hard disk! Programs were on decks of punch cards instead. Individuals used the computer for a single program for hours at a time. Each program had to know how to control all of the hardware of the entire computer that it intended to use. (The first hard disk, the IBM 305 RAMAC, held five megabytes and was introduced in 1956.)

At the same time, the first high-level languages and compilers, of which FORTRAN is the most prominent example, were being developed. During this time, batch system operating systems were created. The computer was run by operators, a specialized profession of people who did not necessarily write programs, but knew how to run a progam on the computer. They took card decks from programmers, fed them to the computer, and returned the original program and data card decks and the output data card deck to the programmers.

On April 7, 1964, IBM introduced the System/360. It was a watershed event in computing history, but for our purposes there are two major innovations:

• Portable software.
• Multiprogramming.

Around 1960, it was realized that more than one user could be seated at a terminal at the same time, and that each user spends most of his time thinking, which would allow the computer to be idle. Fernando J. Corbató led the development of the Compatible Time Sharing System (CTSS), but Corbató credits John McCarthy with the original idea. (Both Corbató and McCarthy are alumni of Caltech, where I went to school.) Corby went on to lead the development of Multics, which would ultimately lead to the development of Unix. With these systems, and others such as some of the operating systems from Digital Equipment Corporation (DEC), especially VMS, the idea of computers that could do many things at once and support many users at the same time became common.

In the late 1970s, personal computers began to appear; their history followed the same sequence as above, only faster. Initially a personal computer could run only a single program. Later, systems could run more than one program, but with no hardware protection for the separate programs, crashes were common as programs interfered with one another. Finally, more "modern" PC operating systems adopted the widely-accepted notions of a kernel and common libraries that allowed programs to more efficiently and safely share a computer. Programs in these systems are run in a process. Today's PC operating systems, such as Linux, Microsoft Windows, and MacOS, are heavily derived from the systems mentioned above. They also commonly ship with a GUI and many utilities, including shells, editors, compilers, and graphics tools, that are not strictly part of the OS itself.

More recently, a similar evolution has happened in the embedded device world; OSes for cell phones, for example, didn't really exist initially. Firmware was hand-coded from the ground up. Later, OSes, even with simple forms of multitasking, arrived (Symbian and PalmOS are probably the two most prominent). Nowadays, with iOS and Android, full-featured, Unix-like system services are available with sophisticated development tools and GUI toolkits.

Computing Today

• We are in the Late Moore's Law Period.
• no more clock speed increases
• multicore (& radical ideas such as TRIPS, WaveScalar)
• 1000x in chip density to come before we exhaust Moore
• heat management critical
• fault tolerance at all levels critical
• CRA architecture "Grand Challenge" 1TOP/W (10000x improvement, 0.001 nJ/op (1E-12 Joules); I think that's per instruction, not gate (binary op); leaves room for 3.5E8 gates/"op"; it's gotta be right up against the Landauer limit)
• The Google era: we can create enormous amounts of data, but can we manage it, both locally and globally? Organizing and finding data is now harder than processing it.
• Mobile and ubiquitous computing.
• Web 2.0: AJAX and asynchronous computing.
• Virtualization of machines, and even networks.

Readings

• Tanenbaum, Sec. 1.1-1.3
• Lampson, Hints for Computer System Design
• Levin and Redell, An Evaluation of the Ninth SOSP Submissions -or- How (and How Not) to Write a Good Systems Paper

Homework

1. Demonstrate that you have a working compiler setup where you will be able to write programs for class. Capture the output of the compilation and execution of a simple program (such as "Hello, world") and submit via SFS. Include the amount of time it took you to complete this exercise.
   C言語のコンパイラがあると証明して。簡単なプログラム("Hello, world" など)を書いて、ソースと実行のアウトプットをSFSに。かかった時間 も記録してください。
2. Read the paper above (Levin) and write a summary of the paper. (It is not necessary to write up notes about the readings from the textbook.)　
   上のLevinの論文を読んで、自分の意見、学んだことなどの ノートを取って。（教科書で読んだところのログは必要あり ません。）
3. Start reading the Lampson paper; for non-native English speakers, it is apparently a long paper, so I will give you two weeks.
   Lampsonの論文読み始まって。「長い」という意見を持っている人います ので、二週間で読む。
4. If you have additional thoughts about the above discussion of programming experience, please email me.
   他の意見があれば、メールでおねがいします。
5. What are the following hexadecimal numbers, in binary and in decimal?
   以下の十六進の数が二進と十進で書いてください。
   1. 0x64
   2. 0x1000
   3. 0x100
   4. 0x400
   5. 0x1049
6. Define the words "pointer" and "memory address".
   「ポインター」と「メモリアドレス」を定義していください。
7. Read the text for next week.
   来週の教科書のことを読んでください。

Next Lecture

第2回 4月16日 システムコール
Lecture 2, April 16: System Calls

• Discussion of Lampson and Levin: What makes a good system?
• System Calls
• Discussion of term projects

• Tanenbaum, 1.5-1.6
• If you are not familiar with hardware, Tanenbaum 1.4
• The Levin paper, cited above

その他 Additional Information

• Class Syllabus and Top Page 授業のシラバスとホームページ
• Rodney Van Meterのホームページ (home page)
• 村井研究室 Murai Lab