COMP 7500 Advanced Operating Systems Dr Xiao Qin

COMP 7500 Advanced Operating Systems Dr. Xiao Qin Auburn University http: //www. eng. auburn. edu/~xqin@auburn. edu Spring, 2012

Your Background Not-A-Quiz • • Have you taken the operating systems class? What lab assignments have you completed in the OS class? What is your on-going dissertation or thesis research project? Is your current research project related to operating systems?

COMP 7500: Semester Calendar See the class webpage for the most up to date version! http: //www. eng. auburn. edu/~xqin/courses/comp 7500

Course Syllabus • Prerequisite: COMP 3500 Operating Systems • 1 midterm exam and 1 final exam • Grading – Class Participation – Midterm – Final – Research Projects – Presentation 10% 20% 40% (Two projects) 10%

Course Syllabus (cont. ) • Scale – Letter grades will be awarded based on the following scale. This scale may be adjusted upwards if it is necessary based on the final grades. – A 90 B 80 C 70 D 60 F < 60

Office Hours and Exams Office hours: WF 3: 00 -4: 00 pm Mid-term W 2/29/2012

Am I going to read the papers to you? • NO! • Papers provides a framework and complete background, so lectures can be more interactive. – You do the reading – We’ll discuss it • Projects will go “beyond”

Questions Please ask at any time!

Focus of comp 7500 • • • Load Balancing Pretching and Caching Security Issues in Operating Systems Energy Conservation Embedded Operating Systems Performance Evaluation

Perspectives of the Computer print cut save send open() malloc() fork() read-disk start-printer track-mouse Application Software System Software Hardware (b) Application Programmer View (c) OS Programmer View (a) End User View

System Software • Independent of individual applications, but common to all of them • Examples? –C library functions –A window system –A database management system –Resource management functions –The OS

Disk Abstractions Application Programmer OS Programmer load(…); seek(…); out(…); (a) Direct Control void write() { load(…); seek(…) out(…) } (b) write() abstraction int fprintf(…) {. . . write(…) … } (c) fprintf() abstraction

Abstract Resources User Interface Application Abstract Resources (API) Middleware OS Resources (OS Interface) OS Hardware Resources

Resource Sharing • Space- vs time-multiplexed sharing • To control sharing, must be able to isolate resources • OS usually provides mechanism to isolate, then selectively allows sharing – How to isolate resources – How to be sure that sharing is acceptable • Concurrency

COMP 7500 Advanced Operating Systems I/O-Aware Load Balancing Techniques Dr. Xiao Qin Auburn University http: //www. eng. auburn. edu/~xqin@auburn. edu Spring, 2012

Technology Trend In reality: Big Fish Eat Little Fish 18

Technology Trend 1988 Computer Food Chain Mainframe Supercomputer Minisupercomputer Work- PC Ministation computer Massively Parallel Processors 19

Technology Trend Clusters Minisupercomputer Minicomputer 1998 Computer Food Chain Mainframe Server Supercomputer Work- PC station Now who is eating whom? 20

Supercomputer Trends in Top 500 SIMD Single processor Cluster Constellations SMP MPP www. top 500. org Nov. 2004 CPU CPU CPU M PC PC BUS/CROSSBAR network MEMORY Symmetric Multiprocessing (SMP) cluster Massively Parallel Processor (MPP) 21 PC

Growth in Microprocessor Performance Observations? 22

Six Generations of DRAMs 23

Technology dramatic change • Processor – transistor number in a chip: about 55% per year – clock rate: about 20% per year • Memory – DRAM capacity: about 60% per year (4 x every 3 years) – Memory speed: about 10% per year – Cost per bit: improves about 25% per year • Disk – capacity: about 60% per year – Total use of data: 100% per 9 months! • Network Bandwidth – 10 years: 10 Mb 100 Mb – 5 years: 100 Mb 1 Gb 24

Updated Technology Trends (Summary) Capacity Logic 4 x in 4 years Speed (latency) 2 x in 3 years DRAM 4 x in 3 years 2 x in 10 years Disk 4 x in 2 years 2 x in 10 years Network (bandwidth) 10 x in 5 years 25

I/O-intensive Applications remote-sensing database systems long running simulations biological sequence analysis 26

Motivation Faster! Memory W: 592 MB/s R: 464 MB/s C: 316 MB/s PCI Bus 264 MB/s W: 209 MB/s R: 236 MB/s Disk Write: 32 MB/s Read : 26 MB/s disk • I/O-intensive Applications require input and output of large amounts of data. • I/O performance can be a potential bottleneck. 27

Current Solutions • Disk I/O Systems – Caching – Prefetching – Parallel I/O • Limitation – Low level – Not Portable 28

Current Solutions (Cont. ) Scheduling/Load balancing Space-sharing (PBS, Backfilling) Time-Sharing Centralized Control (PBS) Distributed Control Coordinated Scheduling (Gang) Disk-I/ONetwork-I/O-aware Non-I/O-aware load balancing (Condor, Mosix, DQS, LSF) Support Parallel Disk-I/O Buffer Sequential Jobs Management Support Homogeneous Clusters Support Heterogeneous Clusters 29