Lecture 3:Availability

Mechanism
REDUNDANCY
Learning Outcomes

 To define redundancy as availability mechanism

 To explain RAID
Availability Mechanisms

 Backup & Restore

 On-line indices allow quick retrieval. Backups to disk allow quick restoring.
 Hardware and software compression can reduce backup times, reduce network load
and reduce the space needed.

 Network backups load the network significantly, it may not be possible, for example
to backup 100 4GB file servers each night over the network. Planning is important.

 Environment
 The computing environment can be protected with Air Conditioning, locked server
rooms and UPS.

 Redundancy
Redundancy
It is one of the Availability Mechanisms
Redundancy

Redundancy increases availability and may be implemented in

hardware (RAID) disk drivers or
OS (RAID)
at the application/service level e.g.:
Replication.
transaction monitors.
backup domain controllers.
 Best RAID for Your
Application:
A. Pure Software RAID solutions: Entry-level RAID 0 or 1 for performance or data availability is key.
 Targeted Applications:
 1. Workstations without large data storage requirements
 2. Entry-level servers without boot protection requirements

B. Hybrid solutions: Cost-sensitive solutions similar to pure software RAID, but with bootability
requirements.
 Targeted Applications:
 1. Entry-level servers without large storage requirements
 2. Compute engines connected to networked storage

C. Hardware RAID solutions: Most feature-rich and highest performance solution. This may be
implemented as RAID on the Motherboard (ROMB) or with plug-in cards.
 Targeted Applications:
 1. High Performance workstations with large data storage requirements
 2. Entry level to enterprise servers requiring performance and scalability from the storage subsystem.
 Application/service
redundancy
 This is often the cheapest and easiest to implement, where available.
 The principle problem is that few applications support this type of redundancy.
 In this type, Clients connecting to these servers automatically look for a backup or duplicate server if the
primary is not available.
 Examples
 Naming servers (NIS+, DNS, NIS, WINS, Lan Manager...) often have this capability in-built and its use is
highly recommended.
 RAID / mirroring is not necessary for these servers, unless the cost of RAID is cheaper.
 Filesystem servers can increase availability (cheaply) by replicating files (e.g. user home directories) to
another system or to another local disk regularly.
 If a major crash of the primary file server occurs, users can mount their files from the second system, but
changes made since the last replication/ synchronisation will be lost.
=> Typical products: Rdist (UNIX), File replicator (NT).
=> To reduce down time, the above method can be used to keep a synchronised copy of the system disk
available on important servers, without adding RAID.
RAID / Mirroring
 The classical method of increasing system availability is by duplicating one of the
weakest part in a computer: the disk.

 RAID (Redundant Array of Inexpensive Disks) is a de-facto standard for defining how
standard disks can be used to increase redundancy. The top RAID systems duplicate
disks, disk controllers, power supplies and communication channels. The simplest
RAID systems are software-only disk drivers which group together disparate disks
into a redundant set.

 There are several RAID levels:

 RAID 1: This is basically mirroring.
 RAID 1+: This is RAID 1 with the addition of parity checking.
 RAID 5: Striping
 RAID 5+: Striping with parity (most commonly used RAID level).
 Actual storage
Available storage
space

RAID1 2 TByte 2 TByte

RAID5 5 TByte 4 TByte

RAID6 8 TByte 6 TByte

System Redundancy
 If applications do not provide built in redundancy, special software (and perhaps
hardware) can be installed on two systems to provide Hot Standby functionality.

 The principle is as follows:

Both systems can access shared (high availability, dual ported) disks and have
duplicate network connections. The backup machine monitors the primary constantly
and if it notices that the primary is no longer functioning, it takes control of the shared
disks, reconfigures it self to have the same network address as the primary and starts
up the applications that were running on the master. Of course this will only work with
certain applications e.g. if the primary crashes and it's principal application thrashes it's
configuration or data files in doing so, the backup server will not be able to start the
application.

 A example of this is IBM's HACMP product, or Sun's HA cluster.

Full Hardware redundancy
 Specialized computer systems offer compete
redundancy in one system i.e. CPU, memory,
disks etc.. are fully duplicated. A single point of
failure should not exist.
These systems often require specially adapted Operating
Systems, cost a fortune and are rarely compatible with
mainstream systems. Rarely used in the commercial arena, they
are most reserved for military or special financial use.

 An example is the Stratos line of systems.

RAID

Originally, the term RAID stood for "redundant array of

inexpensive disks," but now it usually refers to a "redundant
array of independent disks."

While older storage devices used only one disk drive to store data,
RAID storage uses multiple disks in order to provide fault tolerance,
to improve overall performance, and to increase storage capacity in
a system.
How RAID Works
With RAID technology, data can be mirrored on one or more
other disks in the same array, so that if one disk fails, the data is
preserved.
Thanks to a technique known as "striping," RAID also offers the
option of reading or writing to more than one disk at the same
time in order to improve performance.
In this arrangement, sequential data is broken into segments
which are sent to the various disks in the array, speeding up
throughput.
Also, because a RAID array uses multiple disks that appear to be
a single device, it can often provide more storage capacity than
a single disk.
RAID 0
 A RAID 0 (also known as a stripe set or striped
volume) splits data evenly across two or more
disks (striped) without parity information for
speed. RAID 0 was not one of the original RAID
levels and provides no data redundancy. RAID 0
is normally used to increase performance,
although it can also be used as a way to create a
large logical disk out of two or more physical
ones.

 A RAID 0 can be created with disks of differing

sizes, but the storage space added to the array
by each disk is limited to the size of the smallest
disk. For example, if a 100 GB disk is striped
together with a 350 GB disk, the size of the array
will be 200 GB (100 GB × 2).
RAID 1
An exact copy (or mirror) of a set of data on two
disks. This is useful when read performance or
reliability is more important than data storage
capacity. Such an array can only be as big as the
smallest member disk. A classic RAID 1 mirrored
pair contains two disks.
RAID 2
A RAID 2 stripes data at the bit (rather than block)
level, and uses a Hamming code for error
correction.
The disks are synchronized by the controller to spin
at the same angular orientation (they reach Index
at the same time), so it generally cannot service
multiple requests simultaneously.
Extremely high data transfer rates are possible. This
is the only original level of RAID that is not currently
used
RAID 3

A RAID 3 uses byte-level striping with a

dedicated parity disk. RAID 3 is very rare in
practice. One of the characteristics of RAID 3 is
that it generally cannot service multiple requests
simultaneously.
This makes it suitable for applications that
demand the highest transfer rates in long
sequential reads and writes, for example
uncompressed video editing.
RAID 4

 A RAID 4 uses block-level striping

with a dedicated parity disk.
 In the example on the right, a read
request for block A1 would be
serviced by disk 0. A simultaneous
read request for block B1 would
have to wait, but a read request for
B2 could be serviced concurrently
by disk 1.
RAID 5
A RAID 5 comprises block-level striping with
distributed parity. Unlike in RAID 4, parity
information is distributed among the drives.
It requires that all drives but one be present to
operate. Upon failure of a single drive, subsequent
reads can be calculated from the distributed parity
such that no data is lost.
RAID 5 requires at least three disks.
RAID 6
RAID 6 extends RAID 5 by adding an
additional parity block; thus it uses
block-level striping with two parity
blocks distributed across all member
disks.
RAID Levels- Summary
Level 0: striped disk array without fault tolerance
Level 1: mirroring and duplexing
Level 2: error-correcting coding
Level 3: bit-interleaved parity
Level 4: dedicated parity drive
Level 5: block interleaved distributed parity
Level 6: independent data disks with double parity
Level 10: a stripe of mirrors
Level 50:
Level 53:
Level S:
 Parity
 In computers, parity refers to a technique of checking whether data has been
lost or written over when it's moved from one place in storage to another or
when transmitted between computers.
 How it works:
An additional binary digit, the parity bit, is added to a group of bits that are
moved together. This bit is used only for the purpose of identifying whether the
bits being moved arrived successfully.

Before the bits are sent, they are counted and if the total number of data bits is
even, the parity bit is set to one so that the total number of bits transmitted
will form an odd number. If the total number of data bits is already an odd
number, the parity bit remains or is set to 0.

At the receiving end, each group of incoming bits is checked to see if the group
totals to an odd number. If the total is even, a transmission error has occurred
and either the transmission is retried or the system halts and an error message
is sent to the user.
Parity (cont…)

 It's possible that two bits could be in error in a

transmission, offsetting each other.

 In some large computer systems where data integrity

is seen as extremely important, three bits are
allocated for parity checking.

 No parity also usually means that the parity bit can be

used for data, speeding up transmission
Parity in RAID5
Summary

You have learnt:

 redundancy as availability mechanism
 RAID
THE END