Agenda

1
Agenda

• Nestle Case (continued)
• Data Management
• Chapter 5 Review
• Chapter 6 Review

2
Nestle Case

• Lessons Learned (last week)
• Discussion Lack of buy-in from functional
  management
• Causes?
• Solutions?

3
Data Storage

• Character Sets
• File Systems
• Storage Arrays and RAID
• Network File Systems
• Internet File Systems
• Storage Area Networks
• Network Attached Storage

4
Character Sets

• Representations of all text elements
• ASCII Standard for PCs, unix
• EBCDIC Mainframe standard
• UNICODE The web standard many more characters
• Sorting Implications
• ASCII 1Aa
• EBCDIC aA1

5
Ascii Character Set Example(with even parity)
0 0 1 1 0 1 0 1 5
0 1 0 0 0 0 0 1 A
1 1 1 0 0 0 0 1 a
_ _ _ _ _ _ _ _
6
File System Concepts and Issues

• File Allocation Table
• Disk Sectors and Fragmentation
• Hierarchical File Systems
• Journaled vs raw file systems
• Disk Caching
• Mechanism
• Reason

7
Hierarchical File System
src
pgm
exe
win
bin
data
txt
8
Caching Data
Process
Disks
Cache memory
Read-ahead Write to disk
9
Shared Storage Arrays

• Vendors StorTek, IBM, EMC, HP,
• Goals Management, flexibility
• Issues Connectivity
• Virtual Storage

10
RAID (redundant array of independent disks)

• RAID 0 (Striping)
• Consecutive blocks written to different disks
• Performance gain only
• RAID 1 (Mirroring)
• Writes all data to two separate drives
• Reads from either for performance
• RAID 5 (Parity disk)
• Any disk can fail without loss
• Parity drive allows failed drive to be recreated

11
RAID 0 Diagram(Striping)
Disk A
Disk B
Disk C
Disk D

Block 4
Block 1
Block 2
Block 3
12
RAID 1 Diagram(Mirroring)
Disk A
Disk B
Block 1

13
RAID 5 Diagram(with even Parity)
Disk A
Disk B
Disk C
Disk D
Parity Disk
block1 block2 block3 block4
paritydata
1 0 0
1 0 0 1
1 1
1 0 0 0
1 1 1
1 1 0
1
14
Remote Data

• Network File Systems
• Internet File Systems
• Considerations
• Performance
• Security
• Roaming

15
Storage Area Networks

• Fabric connectivity between servers and storage
• Access is still via owning server
• Used for structured data, data associated with
  specific application

16
Fabric Storage Area Network
Disk A
Disk B
Disk C
Disk D
Disk E
SAN Side 1 SAN Side 2
Server 1
Server 2
Server 3
Server 4
17
Network-Attached Storage

• File systems accessed via internet
• Appear as remote drives
• Used for unstructured data (e.g. documents), data
  used by many servers

18
NAS Diagram
NAS
Network
Server
Server
19
Discussion

• What might determine whether we will use SAN or
  NAS in a network design?

20
Chapter 5 Review

• Data availability, not processing power, is the
  limiting factor
• Distributing functions vs distributing control
  Scope
• Does distributed computing imply distributed
  control?
• Security and access implications

21
Chapter 6 Review

• Digital Convergence
• Satisfice vs excellence
• Meads law the logic behind the internet
• Messaging The next killer app