IT-212, How Computers Work Data Storage

Transcription

1 IT-212, How Computers Work Data Storage Adjunct Professor J Masiyowski Electrical and Computer Engineering Spring March 2003 IT 212 Part 4 : Data Storage (C10-14) 1

2 Course Outline Part 1: Boot-up Process Introduction Know the Hardware How Disk Boot Works Operating System s Role Part 2: How Software Works Programming Languages Windows Principles Applications Principles Part 3: Microchips Transistors RAM Microprocessors Part 4: Data Storage Disk Drives Optical & Removable Speed/Storage Part 5: I/O Devices Busses & Ports Keyboard & Displays Pointing Devices Scanners Part 6: Multimedia Sound Video Virtual Reality Part 7: How Internet Works LANs PC Connections Wireless & WWW Part 8: Printers Basic (Impact, LaserJets) Color (Inkjets) In-Class Exercise Apply How Computers Work to a practical situation 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 2

3 Lecture Agenda Chapter 10: How Disk Storage Works? History, concept of operations: electromagnetism Chapter 11: How Disk Drives Works principles of operation, floppy & hard disk Chapter 12: How Drives Increase Speed & Storage disk & file compression, disk defragmentation Chapter 13: How Optical Storage Works CD & DVD drives (Read Only, Write Once, R/W) Chapter 14: How Removable Storage Works Zip, Jaz, QIC tape Storage Device Demonstration 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 3

4 History (1) 1940 Vacuum tubes used for storage 1950 Tape drives replaced punch cards 1956 IBM developed first hard disk Winchester IBM introduced floppy disk (8 ) 1980 Sony introduced 3.5 Floppy, 720KB capacity Why is Persistent Data Storage Important? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 4

5 History (2) 1982 Drivetec announced 3.33MB (unformatted) 5.25 drive 1983 Philips and Sony developed CD-ROM as computer storage device 1992 Cost of 200MB drive dropped below $200 (less than $1 per megabyte) 1998 DVD-RAM drive 17GB (two movies) ~ 2001 affordable solid state storage devices 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 5

6 Introduction Focusing on non-volatile storage when the lights go out do you know what your data is doing? Storage types Online available to computer at all times Offline physically removed from computer Why is Data Storage so Important? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 6

7 External Non-volatile Media (1) Punch card Paper and Mylar Tapes with punched holes Magnetic Tape Large capacity Must be accessed sequentially hence slow Requires manual loading onto tape drive 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 7

8 External Non-volatile Media (2) Magnetic Cylinders Origin of cylinder/sector/track nomenclature on disk drives Automatically retrieved from jukebox Military airborne applications since more vibration resistant than disks Magnetic storage useful on military A/C since it is not susceptible to EMP 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 8

9 Internal Non-volatile Media Core memories (Old Days) Ferrite (magnetic oxide) doughnuts Small capacity, one bit per bead Storage by changing magnetic orientation Random Access Non-volatile Used in Space Shuttle At least to 1980s Radiation and vibration hard sense Flash Electronic Memory (Today!) Row adx Column adx 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 9

10 Hard/Soft Drives Non-volatile Random Access Where does hard come from? Hard drives: rigid aluminum alloy platter with magnetic coating Floppy: flexible mylar substrate Random versus Sequential Access 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 10

11 Next Generation Storage Device Solid State (Electronic Memory - Flash) High storage capacity fast data transfer rates (/w USB 2.0) increased flexibility portability of a floppy, with enhanced capacity Can provide for File/Drive Compression password protection for data security write protection prevents accidental erasure performance: ~ 900K byes/s 480M bps Data retention: ~ 10 years Write cycles: ~ 1,000,000 Trade-Offs? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 11

12 Example: Kanguru MicroDrive MicroDrive: Cost vs. Storage Capacity 2,000 1,500 Price ($) 1, Price ($) ,450 Storage Capacity (MByes) USB Mass Storage Device Data as of 27 Feb 03 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 12

13 Drive Read/Write Technologies Hard disk Disk head floats above fast rotating disk Head crash: r/w head strikes platter, physically destroying magnetic coating and data Denser magnetic coating gives higher storage density Floppy Head actually rides on slow moving disk Sizes: 8, 5.25, and March 2003 IT 212 Part 4 : Data Storage (C10-14) 13

14 Electromagnetism Magnetism and Electricity are interrelated Flow of current creates a magnetic field this is why motors work Movement of a wire in a magnetic field causes current to flow this is why generators work Direction of magnetic field is related to direction of current polarized (N or S) Chapter 10 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 14

15 Writing Data Bits to Disks Disk storage media Thin layer of magnetic particles on/in disk Disk head Ferrite core with a wire wrapped around it Electric current flowing in head generates polarized magnetic field Magnetizing of head magnetizes spot on disk Data actually stored as the change in polarization 1: -/+ +/- OR +/- -/+ 0: +/- +/- OR -/+ -/+ 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 15

16 Reading Data Bits from Disks Reading data Rotating disk causes magnetic field to move relative to read head which induces current in read head Head passes over changing magnetic field Direction of induced current flow reveals polarity of disk magnetic field Electronics reads direction of current flow as 1 or 0 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 16

17 Formatting a Disk Need addresses to locate data Disk segmented by electronics Tracks (concentric circles) Tracks divided into sectors (pie shaped wedges) typically 512 bytes Block or cluster OS defined minimum storage entity Typical cluster in DOS is 4kByte Why is Formatting Necessary? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 17

18 Writing Files to Disk (1) To save a working file from RAM (volatile) to a disk drive (non-volatile) OS reads FAT to find an unused cluster OS converts the number of the unused cluster it read from FAT to the physical track and sector OS sends desired physical ADX and data to BIOS 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 18

19 Writing Files to Disk (2) BIOS communicates with the disk drive electronics and copies data from RAM to the disk drive If more than 1 cluster, OS reserves another free cluster Can be anywhere on the disk FAT contains links from cluster to cluster FAT is updated when write complete 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 19

20 Reading Files From Disk (1) Accessing a file on a disk to work on it is called Opening the file Read only Read/write Application will often list files you recently worked on 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 20

21 Reading Files From Disk (2) To load a file from disk drive (nonvolatile) to RAM (volatile) After selecting file, the OS locates it in the FAT, and reads address of the first cluster The OS converts the cluster number to track and sectors OS requests the BIOS to retrieve the file to RAM; does not destroy original No need to update FAT 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 21

22 Deleting Files MS Windows Unix Files not actually deleted, just access to them Files overwritten to make unreadable by others Under Win95 or 98, OS can automatically move deleted files to recycle bin If a file is deleted from a floppy and not yet overwritten, may be able to retrieve the data using undelete command What are the Security/Privacy Issues? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 22

23 End of Chapter Next: Chapter 11 How Disk Drives Work 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 23

24 Why Floppy Diskettes? Convenient (small size, portable) Widely available; standardized Inexpensive Floppy drive evolution Disk size 8 (1973) 5.25 (1976) 3.5 (1980) Capacity: 90K to 1.4MB typical What about Security/Privacy Protection? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 24

25 3.5 Floppy Disk 80 tracks 18 sectors/track 512 bytes/sector Plastic case Read-only window writable = no hole not-writable = hole Rotation speed 300 RPM Density sensing hole double density = no hole high density (1.44 Mb) = hole Insertion key Metal shutter to protect access to Mylar disk 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 25

26 Hard Drives (1) Evolution Casing 10MB (1980) to 80 GB+ today Access time from 87 milliseconds to < 8ms Size from 3 to 4 thick to almost credit card size Entire mechanism is air-tight to prevent dust, etc. from getting between head and disk Head/disk gap is 2 millionths of an inch Dust is typically 10x gap Smoke particle is typically 5x gap Hair is 30x gap Any head/disk contact causes permanent loss of data 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 26

27 Hard Disks (2) What about Price? (Cost/Mbytes) 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 27

28 Disk Density Never enough storage available Bloated programs including OS Emphasize on-line help vs. printed manuals Large amounts of data Multimedia Graphics MPEG (movies) MP-3 music 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 28

29 Disk Performance Access Speed Time to get data to/from disk Around 7-9ms access time Latency depends on rotation speed Data transfer rate Number of heads Electronics RAM buffer in disk electronics IDE/SCSI Microprocessors getting faster, disk I/O holding steady (not increasing as quickly) Data Transfer Rate - How much/fast the data is exchanged 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 29

30 End of Chapter Next: Chapter 12 How Disk Drives Increase Speed & Storage 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 30

31 Improving Speed Access time - how long on average it takes to move the read/write head into position of the beginning of a file Performance Improvement Methods faster disk rotation faster electronics on device memory caching read ahead write buffering Data Transfer Rate - number of bytes or Mbytes transferred from one device to another in a specified amount of time 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 31

32 Improving Storage The capacity of a given disk can be better utilized by Data Compression Two Levels: File, Disk Some anti-virus programs don t like Some applications will not work with compressed files Defragmentation Utility which should be run occasionally 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 32

33 Directory Many files can be stored on a disk drive Hierarchical approach helps to organize them Directory A grouping of files and subdirectories that can be accessed collectively Special directories Root: First (top level) directory for a given disk. : Current directory.. : One directory closer to root than current directory 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 33

34 Directory Usage Most applications have a default storage directory User can organize files by various means School related IT212 ECE548 Hobbies Aquarium Genealogy 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 34

35 FAT: disk road map Unique table that describes stored on the disk itself in reserved locations Physical location of each file Linkages between clusters Empty locations on disk Locations which do not correctly store data (bad blocks) 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 35

36 FAT Types (1) FAT 12 Original DOS FAT 12-bit cluster address All floppies still use this fat for compatibility Hard disks < 16 Mbytes 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 36

37 FAT Types (2) FAT 16 Win95 16-bit cluster address Hard disks < 2 GB FAT 32 Win 95/98/2000/XP 32-bit cluster address Hard disks < 2048 GB 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 37

38 The FAT Problem Thinning Out the FAT FAT has space for only 65,536 entries (2 16 ) to describe all clusters on disk Larger disk, large cluster size cluster _ size Disk _ size # entries _ in _ FAT 256MB disk, cluster size=4k 512MB disk, cluster size=8k 1GB disk, cluster size=16k 2GB disk, cluster size=32k 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 38

39 FAT Limitation Solutions Disk compression: OS creates 1 large file which packs all files into 1 so that no space is wasted Revise FAT structure Windows 98 FAT32 Up to 6GB, cluster size=4k 2,048GB, cluster size=32kb Eliminate FAT structure: Master File Table Windows 2000, NT NT File system (NTFS) 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 39

40 File System Summary Characteristic FAT12 FAT16 FAT32 NTFS DOS,Windows9x, NT, DOS,Windows9x, Windows98, NT, WindowsNT, Operating System 2000, XP NT,2000, XP 2000, XP 2000, XP MaximumVolumeSize 16MBytes GBytes 2048GBytes nearly unlimited maximumnumberof files ,536 4,294,967,296 nearly <=255characters unlimited filename length 8+3(name+ext) <=255characters <=255characters &unicode set Security no no no access control Ecomony(cluster size) minimal minimal improved maximal Which File System is Best? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 40

41 Disk Storage Paradox Since the cluster is the minimum unit of space allocatable for a single file, all files must be some multiple of clusters long A 1 byte text file will still take up 1 cluster (32K) on a 2GB drive The Larger the disk, the more space will be wasted due to partially filled clusters. 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 41

42 Disk Compression Entire Disk Volume is created as a single file by Operating System Operating System treats this file as a virtual drive Used special device driver to manage access keeps track of how each cluster is used allows for cluster to be shared between files When files are stored, only the space needed to save the file is allocated, thus the cluster limit is bypassed and space is saved Is Disk Compression Necessary/Economical Today? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 42

43 File Compression Algorithm-based looks for recurring data, patterns of data creates dictionary of patterns found in file original data contains pointers to dictionary entries compressed file includes dictionary of patterns In some cases if data has few patterns, compressed file may be larger than original can compress binary files; also file folders trees Password protect compressed file cannot decompress file without password Is File Compression Necessary/Economical Today? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 43

44 Disk Defragmentation Files are stored as linked lists of clusters not necessarily sequentially Erased files leave randomly distributed unused areas New files must be made to fit in unused areas Defrag: Move clusters until each file contiguous How Often to Perform? Fragmented Disk Defragmentation 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 44

45 End of Chapter Next: Chapter 13 How Optical Storage Works 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 45

46 CD-ROMs Compact Disk, Read Only Memory Large Storage Capacity 650MB per side Use for Large Data Not Needing Update Clip art Photographs Large programs Book volumes 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 46

47 CD-R, CD-RW CD-R Compact Disk, Recordable Record only once (WORM) CD-RW Compact Disk, ReWritable Record and erase multiple times 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 47

48 DVD Digital Video Disk Digital Versatile Disk Large Capacity, 8.5 GB per side Movies Huge data storage Maps Phone directories Images 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 48

49 CD-ROM Technology Data Stored in Spiral Sectors Hard and floppy drives use concentric tracks Variable Rotation Speed Constant tangential velocity, i.e., data moves at same speed over optical sensor Faster near center 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 49

50 CD-ROM Technology Inner Circle: 4 sectors Outer Circle: 7 sectors When reading outer circle, need to rotate slower so that time to read each sector is same 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 50

51 CD-ROM Technology Buried Surface Contains Lands and Pits Detector Shines Laser Beam Through Plastic Protective Layer to Buried Data Layer Light that strikes pits is scattered and not reflected Light that strikes lands is reflected directly back towards detector and redirected via prism to photo-diode Photo-detector converts light/no-light to electrical signal indicating 1 or 0 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 51

52 CD-R Technology Disk Layers Clear Polymer (laser side) Clear lacquer Gold reflective Dye (usually green) Polycarbonate (Lexan) plastic structural layer (label side) 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 52

53 CD-R Technology Disk Surface Pregrooved ATIP: absolute timing in pregroove Groove is modulated Laser/photodetector measures modulation and adjusts motor to maintain constant medium/detector relative speed Laser Beam Heats Buried Layer and Produces Pits by Either Bleaching dye layer Distorting polycarbonate plastic layer Producing bubble in dye layer 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 53

54 DVD Technology Disk Layers Symmetric, read from both sides Multiple data layers on each side Laser focused through one data layer to next Spiral in on one layer, out on other No need to reposition head in middle of recording Shorter Wavelength Light Allows Focusing to Smaller Size 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 54

55 CD-RW and DVD RAM (R/W) Recording Layer Is Made of a Material Which Can Be Either Crystalline (Reflective) or Amorphous (Non-reflective) Write Erase Do nothing to make land To make pit, laser beam strikes buried layer and heats spot to F Changes crystalline spot to amorphous Amorphous spots can be annealed back to crystalline state using lower energy beam (400F) Which to Use: CD-R, CD-RW, DVD-RW? 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 55

56 End of Chapter Next: Chapter 14 Removable Data Storage 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 56

57 Zip Drives vs. 3.5 Floppy (1) Capacity MB vs. 1.4MB Speed Heads 3,000 rpm vs. 360rpm Dual read/write heads: lightly contact surface 1/10 size of floppy heads 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 57

58 Zip Drives vs. 3.5 Floppy (2) Track spacing 2,118 vs. 135 Sector Same physical size vs. same angular size Disk surface Higher density recording media 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 58

59 Removable Hard Drives (1) Have high density but need Portability work/school/home Expandability Individual drives for particular applications Backup/Security make copy of non-removable drive, store elsewhere requires more physical security handling 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 59

60 Removable Hard Drives (2) Typical hard drive is sealed to prevent particles causing head crash Removable alternative: Jaz drive Metal door seals drive when not in housing Filter around spindle prevents particles greater than 3 microns from being pulled into disk while spinning 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 60

61 Removable Hard Drives (3) Particle less than 2 microns, passes under head If particle more than 2 microns, contacts head, pushes head away. If write occurs (high write) drive detects and reads to see if OK Read/Write Arm Motion of Disk surface Read/Write Head 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 61

62 QIC Tape Drive Primary Purpose: Data Archival (i.e., Long Term) Large Capacity ~ 4G bytes + inexpensive sequential access number of standard recording formats directory {e.g., FAT} of entire contents placed first or in separate track recorded in spiral fashion on the tape media tracks are segments into blocks {similar to disks} includes Error Correction Codes with the data multiple read/write heads for write verification 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 62

63 Media Data Retention Rate How long will data stored on computer media remain intact? Media Years Stone Tablet 2200 Archival Acid-Free Paper Photographic Slides 100 Microfilm Magnetic Tape Digital Linear Tape Data Grade VHS Tape Newspaper CD Lifetime of Technology versus Lifetime of Medium 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 63

64 Future of Storage Hard drive Historically 60% capacity increase per year Improved cost/size $35 per megabyte to $.01 in last 10 years Future technologies Others: Near-field recording/magneto-optical Uses combination of laser and magnetism Optically assisted Winchester (OAW) drives Reflected laser beam affected by magnetic polarity Uses MEMS (microelectro-mechanical system) for mirrors, lenses Web/Internet/On-Line 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 64

65 End of Lecture Next Lecture: Input/Output Devices 3 March 2003 IT 212 Part 4 : Data Storage (C10-14) 65