Scribd
Upload
9 views

Fault Tolerance - Lecture 3

fault tolerance

Uploaded by

Ebenezer Komla Gavua
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
9 views

Fault Tolerance - Lecture 3

fault tolerance

Uploaded by

Ebenezer Komla Gavua
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 21

Storage

• Data storage: also has issues of availability and fault tolerance


• Different methods are available for making sure shared data and applications
are never lost or irretrievable
RAID (Redundant Array of Independent [or
Inexpensive] Disks)
• Collection of disks
• Provide shared data, application fault tolerance
• Disk array (drive)
• Group of hard disks
• RAID drive (RAID array)
• Collection of disks working in a RAID configuration
• Single logical drive
RAID (cont’d.)
• Hardware RAID
• Set of disks, separate disk controller
• RAID array managed exclusively by RAID disk controller
• Attached to server through server’s controller interface
• Software RAID
• Software implements, controls RAID techniques
• Any hard disk type
• Less expensive (no controller, disk array)
• Performance rivals hardware RAID
RAID (cont’d.)
• RAID Level 0 - Disk Striping
• Simple RAID implementation
• Data written in 64-KB blocks equally across all disks
• Not fault-tolerant
• Does not provide true redundancy
• Best RAID performance (in this chapter)
• Uses multiple disk controllers
RAID (cont’d.)

Figure 14-6 RAID level 0 - disk striping


RAID (cont’d.)
• RAID Level 1- Disk Mirroring
• Disk mirroring provides redundancy
• Data from one disk copied automatically to another disk
• Advantages
• Simplicity, automatic and complete data redundancy
• Disadvantages
• Cost of two disks
• CPU usage because software does the mirroring
RAID (cont’d.)
• Disk duplexing
• Similar to disk mirroring
• Data continually copied from one disk to another
• Separate disk controller used for each disk
• Provides added fault tolerance
RAID (cont’d.)

Figure 14-7 RAID level 1 - disk mirroring


RAID (cont’d.)
• RAID Level 3 - Disk Striping with Parity ECC
• ECC (error correction code)
• Algorithm to detect, correct errors
• Known as parity error correction code
• Parity
• Mechanism to verify data integrity
• Number of bits in byte sum to odd, even number
• Use either even parity, odd parity, not both
Table 14-1 The use of parity bits to achieve parity

• RAID Level 3 - Disk Striping with Parity ECC (cont’d.)


• Parity tracks data integrity
• Not data type, protocol, transmission method, file size
• Parity error checking
• Process of comparing data parity
Figure 14-8 RAID level 3 - disk striping with parity ECC

• RAID Level 3 - Disk Striping with Parity ECC (cont’d.)


• Advantage
• High data transfer rate
• Disadvantage
• Parity information appears on single disk
RAID (cont’d.)
• RAID Level 5 - Disk Striping with Distributed Parity
• Most popular data storage technique
• Data written in small blocks across several disks
• Parity error checking information distributed among disks
• Advantages over RAID level 3
• Writes data more rapidly
• Uses several disks for parity information
• Disk replacement causes little interruption
• Controlling software regenerates failed file parts
RAID (cont’d.)

Figure 14-9 RAID level 5 - disk striping with distributed parity


RAID (cont’d.)
• RAID Level 5 - Disk Striping with Distributed Parity
• Hot spare (failover component)
• Array disk, partition used only when one RAID disk fails
• See link Ch 14i
• Cold spare (hot swappable component)
• Duplicate component
• Not installed
NAS (Network Attached Storage)
• Specialized storage device, storage device group
• Provides centralized fault-tolerant data storage
• Difference from RAID
• Maintains own interface to LAN
NAS (cont’d.)
• Advantages
• NAS device contains own file system
• Optimized for saving, serving files
• Reads, writes fast
• Easily expandable
• No service interruption
• Disadvantage
• No direct communication with network clients
• Use
• Enterprises requiring fault tolerance, fast data access
NAS (cont’d.)

Figure 14-10 Network attached storage on a LAN


SANs (Storage Area Networks)
• Distinct networks of storage devices
• Communicate directly
• With each other, other networks
• Multiple storage devices
• Connected to multiple, identical servers
SANs (cont’d.)
• Advantages
• Fault tolerant
• Fast
• Special transmission method
• Special protocols, like Fibre Channel
• Despite the name, Fibre Channel can run over both copper and fiber media (link Ch 14k,
14l)
• SAN can be installed in location separate from LAN served
• Provides added fault tolerance
• Highly scalable
• Faster, more efficient method of writing data
SANs (cont’d.)
• Drawbacks
• High cost
• Small SAN: $100,000
• Large SAN: several million dollars
• More complex than NAS, RAID
• Training, administration efforts required
• Use
• Environments with huge data quantities requiring quick availability
Figure 14-11 A storage area network

You might also like