Supervised by Prof
Download 281.3 Kb.
|
- Navigate this page:
- Table of Contents Introduction 6 Section 1: CMS and its Implementation 8
- I: Direct Manipulation Interfaces 17 II: Interfaces for Systems Administration 19
- IV: Similar Projects 36
- II: Description of Experiments 54
- III: Evaluation Techniques 58 IV: Real-world Experimentation 59 Future Directions 61 Acknowledgements 64
| האוניברסיטה העברית בירושלים
הפקולטה למתמטיקה ולמדעי הטבע- המכון למדעי המחשב מערכת קוביות לניהול (Cube Management System): מחקר בנושא ממשקים מוחשיים לניהול מערכות
בהנחיית פרופ' סקוט קירקפטריק עבודת גמר לתואר מוסמך במדעי הטבע י"א טבת תש"ע The Hebrew University of Jerusalem, Faculty of Mathematics and Natural Sciences Selim and Rachel Benin School of Computer Science and Engineering The Cube Management System: A study in Tangible User Interfaces for Systems Administration Aviva Dayan Supervised by Prof. Scott Kirkpatrick Thesis submitted in partial completion of the requirements for the degree of Master of Science December 27, 2009 תקציר יחד עם עליית רמת המרוכבות של מערכות מחשב, דרגת הקושי של ניהול מערכות אלה משימת הבקרה עליהם גם עלתה – דבר שהביא איתו איום תמידי של איחורים באחזקת המערכות ותיקונם, כאשר כל שנייה של השבתת מערכת במימדים אלה עלולה להביא להפסדים רבים לחברה המתחזקת אותה. ייתכן וישועה מסוימת תבוא משינוי בכלים המשמשים לבקרה על מערכות אלה. תחום הממשקים המוחשיים מבטיח במיוחד בעניין זה: ייתכן והוספת אספקט המוחשיות תוכל להוסיף יתרונות כגון הורדת עומס הקוגניטיבי של מנהלי הרשתות, שמירת ‘צילומים נישאים’ של מצבי המערכת, ומאוד ייתכן גם שתוסיף לייעול העבודה בצוותים. בנוסף, שינוי צורת הצגת המידע עשויה להקל על גילוי בעיות במערכת בהקדם האפשרי. תזה זו מציגה אפוא ממשק מוחשי מקורי לבקרה על מערכות גדולות: מערכת בקרה בקוביות - או כפי שמכונה באנגלית, ה Cube Monitoring System )או בקיצור CMS ) המייצגת מערכת באופן הירארכי ודינאמי בעזרת קוביות המונחות על מגש בעל אזורים שונים. בהיבטיו השונים של ה CMS מוצגים בהקשר של מחקר קיים בתחומי עיצוב ממשקים וניהול מערכות. סקירת המחקר הקיים נותן שפה לתיאור ממשקים, בנוסף להנחיות לעיצובן, וכלים להערכתן. הרקע התיאורטי מאפשר ניתוח מעמיק יותר של מערכת ה-CMS והערכת מידת התאמתן של תכונות ה CMS לצרכי מנהלי המערכות. ניתוח זה הנחה את תכנונם של קבוצת ניסויים שמטרתם היא בדיקת הצלחתה של ממשק ה CMS לספק פתרון לבעיות הניצבות בפני משתמשיה הפוטנציאליים. בסוף העבודה מוצגים כיוונים אפשריים למחקר עתידי, כולל בעיות וסיבוכים פוטנציאליים, וגישות נוספות למחקר, ומוצגות גם דרכים אפשריות לשפר מימושים עתידיים של מערכת ה- CMS. Abstract As computer systems have become increasingly complex over the years, the task of monitoring and administrating these systems has also increased in complexity, leading to the threat of potential costly and dangerous errors and delays in their maintenance and repair. Changing the type of tools used for monitoring the system could potentially bring some relief. The emerging field of tangible user interfaces shows some promise in this respect: adding the aspect of tangibility may offer advantages such as potentially lessening the cognitive burden of administrators, preserving 'portable snapshots' of system state, and also possibly facilitating teamwork. Tangibility may also allow the display of information to be rearranged in a manner that could facilitate the discovery of looming – and burning – problems, as early as possible. This thesis thus describes and analyzes an original, tangible user interface for large systems monitoring – the Cube Management System (CMS), which represents the system in a dynamic, hierarchical manner by means of cubes on a sectioned tray. CMS is then discussed in the context of preexisting research in the fields of interface design and systems administration. The survey of prior research provides the language for describing interface systems, as well as guidelines for their design, and tools for evaluation. This theoretical background allows for a deeper analysis of the CMS system and an assessment of the extent to which these features fulfill the needs of system administrators; this analysis then serves to guide the construction of a set of experiments to assess the usefulness of the CMS system in fulfilling its purpose. The paper concludes by presenting possible directions for future research, including potential issues and complications, additional research approaches, and presents possible ways to improve future incarnations of the CMS system. Table of Contents Introduction 6 Section 1: CMS and its Implementation 8 I: Overall Architecture 8 II: CMS Functionality 10 III: Initial Prototype 13 Section 2: Background 16 I: Direct Manipulation Interfaces 17 II: Interfaces for Systems Administration 19 Ecological Interface Design 20 Haber & Bailey's Design Guidelines 24
Fitzmaurice’s Thesis 28 Hiroshi Ishii & Brygg Ullmer’s 2000 Paper 32 Holmquist, Redström and Ljungstrand’s Terminology 35 IV: Similar Projects 36 Triangles 37 mediaBlocks 38 Lego Wall 39 Bricks 41 Section 3: An Analysis of CMS 44 I: CMS as a TUI 44 II: CMS as a Tool for Systems Administration 48 III: A General Analysis of CMS 51 Section 4: Testing 53 I: Goals of Experiments 53 II: Description of Experiments 54 Experiment set I: Shape-based affordances 54 Experiment Set II: Dynamic Layout 54 Experiment Set III: Alerts 54 Experiment Set IV: Interdependence of Group actions 55 Experiment Set V: Aggregation / De-aggregation 55 Experiment Set VI: Alternative Hierarchies 55 Experiment Set VII: Multiple Display Facets 56 Experiment Set VIII: Cube Size and Shape 56 Experiment Set IX: Hybridity 56 III: Evaluation Techniques 58 IV: Real-world Experimentation 59 Future Directions 61 Acknowledgements 64 IntroductionMaintaining a fully-operational system is usually a matter of crucial importance for its owners, where even several seconds' downtime can carry huge costs in terms of unprocessed transactions, chain reactions, and direct losses. However, these systems are continually growing in complexity, leading to an ironic situation wherein even as organizations become more reliant on such systems, they become more difficult to maintain. The administrator's burden could potentially be somewhat alleviated through the use of a tangible interface. The Cube Management System (CMS for short) is an original system1 specifically for this purpose. CMS offers a Tangible User Interface (TUI) for monitoring large-scale systems, and its tangibility promises to offer several advantages over traditional Graphical User Interface (GUI) - or Command Line Interface (CLI) - based systems. Such assumed advantages include the reduction of system administrators' cognitive load, and the facilitation of group interaction - among other advantages which could potentially save precious seconds in maintaining these complex and often mission-critical systems. CMS uses cubes to represent the components of its underlying system. The assignment of a cube-component representation is determined based on the location of the cubes on a specially sectioned tray. The tray is connected to a live, logical representation of the underlying system, so that the cubes display the current status of the components that they represent. However, assumptions about CMS's viability need to be evaluated: This would assess just how much – if any – advantage is gained from the switch to tangibility, and if so, what aspects of the tangibility procure the advantage. Such an assessment would also evaluate whether any – and if so, what - disadvantage might result from the introduction of tangibility as an element to the system. A decision then needs to be made as to whether the advantage outweighs the disadvantage and if any worthwhile improvements can be brought about by redesign. Such an evaluation would elucidate not only the usefulness of CMS as a tool in systems design, but would also clarify issues in cognitive psychology, perception and tool-usage, as applied to TUI design and to interface design in general. The first step in the evaluation process undertook a survey of prior research in the field of tangible interfaces, and studied the language used for analyzing these tools. This paper presents a summary of this survey, and then undertakes to analyze CMS using the language of TUI analysis as it emerges from the existing research. It then outlines proposed experiments, together with an analysis of how these experiments might be useful in evaluating the features of tangibility in general and of CMS in particular. After undertaking the experiments outlined in this paper, our results should serve as a coherent and unified assessment of the goals and advantages of TUI design for systems administration and of the overall needs of system administrators as human users, and thus offer a significant contribution to the research and development efforts in these two fields. The discussion of CMS ends with proposed future directions, which merit evaluation as alternative features of the CMS system. Download 281.3 Kb. Share with your friends:
|