1. Ch. 2-2 Computer Clusters1 2.3 컴퓨터 클러스터의 설계 원칙 2.3.1 Single-System Image Featues  It means the illusion of a single system, single control, symmetry, and transparency.  Single system: the entire cluster is viewed by users as one system that has multiple processors.  Single control: Logically, an user or system user utilizes services from one place with a single interface.  Symmetry: All clusters services and functionalities are symmetric to all nodes and all users, except those protected by access rights.  Location-transparent: The user is not aware of the where about of the physical device that eventually provides a service.  Cluster nodes  home node  local node  remote nodes  The illusion of an SSI can be obtained at several layers: application software layer, hardware or kernel layer, middleware layer.

2. Ch. 2-2 Computer Clusters2  Single Entry Point  The single entry point enables users to login to a cluster as one virtual host.  The system transparently distribute the user ’ s login and connection requests to different physical hosts to balance the load.  Realizing a Single Entry Point in a Cluster of Computers Fig. 2.13  Single File Hierarchy From the view-point of any process, files can reside on three types of locations in a cluster, as shown in Fig. 2.14. A stable storage requires two aspects: persistent, fault-tolerant. Stable storage (global files) could be implemented as one centralized, large RAID disk. But it could also be distributed using local disks of cluster nodes.  Single I/O Space over Distributed RAID for I/O-Centric Clusters Fig. 2.16

3. Ch. 2-2 Computer Clusters3  RAID 2.3.2 High Availability through Redundancy  When designing robust, high available systems three terms are often used together: reliability, availability, and serviceability (RAS).  신뢰성 : 시스템이 고장 없이 얼마나 오래 동작할 수 있는지를 측정  가용성 : 시스템이 사용자에게 가용인 시간 백분율  서비스 가능성 : 시스템을 서비스 ( 유지, 보수, 업그레이드 ) 하는 것이 얼마나 쉬운지를 말한다.

4. Ch. 2-2 Computer Clusters4  Availability and Failure Rate  Availability=MTTF/(MTTF+MTTR)  MTTF (mean time to failure)  MTTR (mean time to repair)  Planned vs. Unplanned Failures  Transient vs. Permanent Failures  Partial vs. Total Failures  Single Point of failure in an SMP and in Clusters of Computers, Fig. 2.19.  Redundancy Techniques  Table 2.5 Availability of Computer System Types  Isolated Redundancy  When a component (the primary component) fails, the service it provided is take over the another component (the backup component).  The primary and the backup components should be isolated from each other.  Benefits not a single point of failure 고장 된 구성요소는 나머지 시스템이 작동 중 일 때, 수리될 수 있다. 주된 구성요소와 백업 구성요소는 서로 테스트하고 디버거 할 수 있다.

5. Ch. 2-2 Computer Clusters5  N-Version Programming to Enhance Software Reliability  The software is implemented by N isolated teams who may not even know the other exist.  Different teams are asked to implement the software using different algorithms, programming languages, environment tools, and even platform.  In a fault-tolerant system, the N versions all run simultaneously and their results are constantly compared. If the results differ, the system is notified that a fault has occurred. 2.3.3 Fault-Tolerant Cluster Configurations  Three ascending levels of availability  Hot standby server clusters  Active-takeover clusters  Failover cluster 시스템 대체작동은 다수의 기능들 : 고장 진단, 고장 공지, 고장 복구를 제공해야 한 다.  Recovery Scheme  Backward recovery Checkpoint Rollback

6. Ch. 2-2 Computer Clusters6 2.4 클러스터 작업 및 자원 관리 2.4.1 Cluster Job Scheduling Methods  Cluster jobs may be scheduled to run at a specific time (calendar scheduling) or when a particular event happens (event scheduling).  Table 2.6 Job Scheduling Issues and Schemes for Cluster Nodes  Space Sharing  Multiple jobs can run on disjointed partitions of nodes simultaneously.  At most, a process is assigned to a node at a time.  Job Scheduling by Tiling over Cluster Nodes, Fig. 2.22  Time Sharing  Independent scheduling (local scheduling)  Gang scheduling The gang scheduling scheme schedules all processes of a parallel job together. When one process is active, all processes are active.  Competition with foreign jobs

7. Ch. 2-2 Computer Clusters7 2.4.2 Cluster Job Management Systems  A Job Management System (JMS) should have three parts:  user server  job scheduler  resource manager: 자원 할당 / 감시, 스케줄링 정책 시행, 회계정보 수집  JMS Administration  Cluster Job Types  Characteristics of a Cluster Workload  NAS 벤치마크 경험에 기초한 작업 부하 특성, p. 108 참조  Migration Schemes  Node availability  Migration overhead  Recruitment threshold The recruitment threshold is the amount of time a workstation stays unused before the cluster considers it an idle node. 2.4.3 Load Sharing Facility for Cluster Computing