1. COMP 1321 Digital Infrastructure
Richard Henson
University of Worcester
October 2017

2. Week 5: Data representation Data Transfer on/off motherboard
Learning Objectives:
Explain what stored data could represent
Explain how storage media store data and how it can be retrieved
Explain mechanisms for data transfer on/off motherboard
Explain how data on a storage medium can be found & displayed by named software tools

3. CPU, Memory and Boot-up BOOT process requires CPU instructions…
have to be stored somewhere (in files)
Booting up is a matter of loading all these files & their programs
Where to store?
need to be accessed & processed as quickly as possible!

4. Memory… or Storage? Lot of confusion…
memory interfaces directly with CPU
also called primary storage
held on motherboard, controlled by system clock
fast (dynamic RAM), very fast (static RAM)
quite fast (ROM)

5. Primary & Secondary Storage
any other form of data storage not directly interfacing with CPU via bus
accessible to CPU via i/o calls
e.g. INT 21 (Intel 8086…)
uses ports to connect electrically with CPU
external e.g. USB, Ethernet
internal e.g. SATA
Slower than primary storage

6. Virtual Memory
Use of fast(ish) secondary storage device locations as if they were primary storage locations…
hard disk (especially SSD)
storage addressed directly by the CPU
Requires programmed mapping between extra primary storage locations & secondary storage locations
adv: unlikely that the CPU will run out of “memory”
disadv: hard disk performance “falls of a cliff” when virtual memory interfaces with CPU

7. “Firmware” Needs updating (like all software)…
Software held on EPROM (erasable, programmable) chip
can’t easily be tampered with
IDEAL for low-level operating system programs, ensures rapid boot-up
also embedded applications
Needs updating (like all software)…
some flexibility to overwrite

8. Questions Is virtual memory primary or secondary storage?
What about firmware?

9. What could Stored Data Represent…
With one 1 byte word:
i.e. 1 byte per memory location
Could be many things!!!
Can be difficult to decide what the data really does represent…
e.g: data has been recovered from a location; presented as 4E (hex)
what is it… ?

10. What could “4E” represent?
Could be part of a program instruction
in assembly language
or source code (as ASCII code)
Program data
as a number
as an ASCII character…

11. More possibilities for “4E”
Over to you…
In groups…
Five minutes…

12. Putting meaning onto raw data… (1)
Single item of data… at a single location… (e.g “peeked” as 4E) could be anything!
Only find out context by studying other bytes around it…
if next byte is… 4F (hex)
and byte after that is… 57 (hex)
the ASCII codes together spell “NOW”
so the bytes are probably all ASCII codes

13. Putting meaning onto raw data… (2)
What if the next bytes were 6B and 7D?
ASCII codes would deliver… Nk}
not a proper word
data probably not ASCII codes
What else?
could be integers between 0 and 255
maybe stored variables, or constants…
Wrong to assume… keep an open mind!

14. Use of “Control Bits”
The byte could also be broken up into two nibbles of data
0100… could be an integer of value 4
1110… could be an integer of value 14
It could also be made up of 8 “Boolean” values, which could control outputs to machinery
i.e. 0 = off; 1 = on

15. Looking at Memory locations… (Peeking)
Intel 8086 tool… debug
available since early days of DOS
Debug needs –d parameter to peek…
shows 128 bytes at a time (& corresp ASCII codes)
default memory location is the start of “free” memory
locations may still contain data from previous usage
Specified memory locations can be peeked
e.g. –d 0200 for next 128 bytes starting from &0200

16. “Peeking” and “Poking”…
& … represents address (as opposed to data)
Debug –d can be used to present a whole range of memory
e.g. –d would display every byte between addresses &0200 (hex) and &0300 (hex)
Debug –e can overwrite contents of a specified location (or sequence of locations) with new data
called “poking”
potentially can crash the computer… (!)

17. Protection against memory overwrite
Operating system protects memory addresses used by “active” processes
Use of debug -e bypasses protection!!!
only protection for computer’s primary memory is to disable the debug program
but could in theory still be executed remotely, if administrative access to local computer has been granted… (!)

18. How does data get between devices?
Data usually needs to go in both directions…
DEVICE A
DEVICE B

19. Three Data Communication Alternatives
Simplex
one direction only
Example:
Broadcast data from a radio or TV mast

20. Data Transfer Half Duplex Example: one direction only at a time
Data sent along a single copper wire
first
then

21. Data Transfer Full Duplex Example: both directions simultaneously
Broadband telephone communications

22. i/o connections with the motherboard
Normally connect digital i/o devices to the motherboard via:
Direct connections through “ports”
Click in expansion or “daughter” cards with their own ports

23. i/o Buses used with older expansion Cards
ISA = Industry Standard Association
early ( ) communications standard
speed: up to 16 MB s-1
8 or 16-bit parallel connections
PCI = Peripheral Component Interconnect
later ( ) communications standard
speed: up to 133 MB s-1
32-bit parallel connection
‘Plug and play’ – no set-up software needed, (depending on the operating system used…)

24. Legacy Motherboard: PCI & ISA slots from http://www. ibase-i. com
PCI slot
ISA slot

25. Peripheral Connectors on the Motherboard
On-board IDE slot (now legacy)
up to TWO hard disk or DVD-ROM
40-pin “ribbon” cable
On-board SCSI slot (server board)
connects a much larger number
of devices

26. Other Hard Disk connections
On-board SATA slot
thinner ribbon cable
3.5” SATA hard disk
2.5” SATA hard disk
External SATA hard disk
Connected to motherboard
via USB

27. STAR motherboard architecture
Copied from “star” arrangement for networking computers
one hub (MCH) connects fast components
hub at centre; components at ends of ‘spokes’
other hub (IOCH) connects slower components and peripherals
hubs communicate directly with each other

28. Motherboard Hubs MCH = Memory Control Hub
connects very fast devices together in a ‘star’ configuration
I(O)CH = Input-output Control Hub
connects together slower devices, also in a star configuration

29. MCH and I(O)CH from http://www.3dnews.ru/motherboard/intel-ht-chipset/

30. Motherboard with MCH and ICH from http://www. tomshardware

31. Why arrange motherboard components like this?
Longer wires…
more time to send messages (good)
degradation of message at high speed (bad)
Therefore…
important for fast components to be close together
slower components can be further apart

32. Motherboard with MCH and ICH from http://www. tomshardware
AGP slot
Socket for processor
MCH
ICH
Slots for RAM cards

33. Another PC Motherboard… from http://www. techiwarehouse

34. Finding data on Secondary Storage (1)
“file” ~ conventional name for a package of bytes of data
Primary storage : controlled directly by CPU instructions
Secondary storage: controlled by hard disk controller programs & file system manager

35. Primary/Secondary Storage of data as files
Secondary storage devices organise data for quick access
logically structured into “partitions”
If Windows = operating system, each partition allocated a letter (e.g. C:, D:, etc.

36. File Organisation on Disk
Disk could be stretched out to form a long line of sectors
size of sectors depends on formatting type
512 locations (i.e. bytes)
2048 locations
files laid down in sectors

37. Finding data on Secondary Storage
Essential for each partition to create a table or catalogue for starting address of files that are written to it
otherwise the file becomes very difficult to retrieve…
Method depends on filing system chosen when partition formatted…
As well as formatting each filing system structures the media to receive data in its own unique way…

38. “Boot Sector” Important for hard disk boot up
Process of loading operating system from on secondary storage starts from…
provides configuration information for effective communication with CPU
if damaged, boot up halted!
should have a backup… needs to be copied to boot sector to overwrite corrupted data

39. Partitions Created by special program
Areas of hard disk managed by a file system
different partitions can use different file systems
single boot partition containing boot sector
can point to different operating systems
Selectable via screen menu

40. Booting up: loading an Operating System…
Needs to be loaded into RAM
some operating systems load everything from ROM
others use a combination…
some loaded first from ROM
rest from hard disk or other source
Hard disk needs a bootable partition to load rest of operating system into RAM

41. Data Storage on Disk Partition
Sectors numbered
Files stored in specified sector address ranges

42. Disk Catalogues Organise files into directories/folders
Top folder (C:) = root
Rest of folders link hierarchically
Catalogue logically allocates each file to a folder for ease of retrieval

43. Fragmentation of Data on Secondary Storage
General problem with hierarchical data storage…
deleted data items leave holes in the structure
New data items saved try to fill the gaps
large files can be broken into fragments
fragments linked by address pointers
slows down retrieval

44. Removing Fragmentation
If disk only partly fragmented…
defragmented files copied into memory
remaining files moved around to close up holes
Previously defragmented files copied back to disk as complete files
If disk >75% fragmented
most effective solution is to copy all files to another partition
can copy back later once original partition has all data deleted

45. Is it true that deleted files aren’t really deleted?
Absolutely!
Two things happen when a file is deleted:
the first data item stored in the file (first character of filename) is changed to “?”
the catalogue entry ceases to recognise & display the filename
starting address shows “?” Character
file system is programmed to ignore ? at such a location
Rest of the data is untouched…
easily demonstrated through use of a Hex editor program: can show file contents “before” and “after”

46. “Normal” Loading of a File from Secondary Media
File catalogue essential for data retrieval
application reads file catalogue
displays folders and files
user chooses file, application uses disk addresses to load into memory
What if file catalogue corrupted?
backup copy on disk…
what if both become corrupted?

47. Direct access by address on Secondary Media
If both file catalogues are damaged…
file (and its data) cannot be located
“Hex editors” available to do the equivalent of debug –d (peek) and –e (poke)
enables full search of all addresses for particular ASCII string(s)
essential for recovery of data…
Also used for restoring recently deleted files
“?” character restored to a real character
then be picked up & shown on catalogue display

48. WinHex Probably the most popular tool to examine hard disks
readout quite similar to debug –e
data presented byte-by-byte according to catalogue address
range of options for extracting, overwriting data, and (like debug) writing consecutive raw data items to a file