What's Raid?
So what exactly is RAID? Nope, it's not the bug spray I'm talking
about here. It is a technique that was developed to provide speed,
reliability, and increased storage capacity using multiple disks,
rather than single disk solutions. RAID basically takes multiple hard
drives and allows them to be used as one large hard drive with benefits
depending on the scheme or level of RAID being used. Depending on
your needs, there are many different RAID variations and implementations
available with prices ranging from less than $100 to over $25,000.
Of course, the better the RAID implementation, the more expensive
it's probably going to be. There is really no one best RAID implementation.
Some implementations are better than others depending on the actual
application.
Levels of Raid
RAID 0 has the best performance and capacity,
but the lowest availability (no fault tolerance). If one drive fails,
the entire array would fail because part of the data is missing without
any ways in recovering them.
RAID 1 has the highest availability but
lowest capacity, it requires using twice the number of drives . Performance
is roughly the same as when using a single drive, although in some
cases the dual write may be somewhat slower.
RAID 0+1 offers some performance improvements
by combining striping and mirroring, but capacity is low since the
mirror requires a duplicate set of drives.
RAID 3 has high performance and middle
capacity, but the availability is lower when comparedtoRAID 1
RAID 5 has moderate benefits in all three areas.
Read performance can be as fast as RAID 0, but write performance is
slower, since the parity information must be calculated and written
along with the data. Capacity is higher than for RAID 1 but it does
not use striping, since the array use additional space for parity
information, Availability is high with RAID 5 because of fault tolerance?If
a drive fails, the missing data is recalculated from remaining operation
drives.
DISK MIRRORING
DATA STRIPPING
Conclusion
So what have we learned here? Well we've learned that RAID is not
just a bug spray. RAID is a good solution for companies or individuals
craving more transfer performance, redundancy, and storage capacity
in their data storage systems. There are many levels of RAID, which
range from very simple and cheap to extremely complex and expensive.
The benefits of having RAID in your system are obvious, however, RAID
is not for everyone. Performance freaks like myself will always like
what RAID has to offer, but the price tag of the better RAID implementations
is still a hurdle to overcome. I tried to cover as much as I possibly
could about what RAID is, it's benefits, and the various implementations.
I tried not to focus on any specific company or hardware implementation,
but looked at RAID in general. I hope you've enjoyed reading this
article and maybe even learned something along the way. This isn't
the definitive guide to RAID, but I hope it's helped you understand
RAID more.
What is Seria ATA (SATA)
?
Basics
Serial ATA (SATA) is the next-generation interface standard for low-cost
direct-attached storage in desktop PC, workstation, and entry-level
server environments. As a serial technology (bits transmitted in a single
stream, rather than along parallel paths) SATA eliminates the restrictions
on performance, reliability, and scalability that are inherent in today?s
parallel ATA (IDE) standard. Because SATA cost-effectively enables RAID
protection, is easily scalable, and has a high performance roadmap,
it will become the dominant direct-attach storage interface for budget-conscious
users.
Storage Interface Evolution
When it was introduced nearly twenty years ago, parallel ATA, also known
as IDE, provided a simple low-cost storage interface standard that met
the performance and flexibility needs of desktop PCs. Concurrently,
the more robust and costly SCSI interface evolved to fulfill the higher
performance, reliability, and scalability requirements of enterprise-class
applications. However, as CPU capabilities and the complexity of applications
and data types accelerate, technology restrictions limit the future
applicability of parallel interfaces. To better meet future processing
needs, both ATA and SCSI are moving to more flexible and capable serial
technology. Over time, SATA for the low-end and Serial-Attached SCSI
(SAS) for the high-end will become the industry standard storage interfaces.
Parallel vs. Serial Interfaces
| Interface |
Technology |
Transfer Rate |
Cabling
Connectivity |
Connectivity |
|
Current
|
Planned
|
| ATA/IDE |
- Parallel |
- 133 MB/s |
- At max today |
- Wide ribbon
- 40-pin
- 18-inch length |
- 2 drives per
channel
- Master/slave relationship
- Shared bandwidth among drives |
| SATA |
- Serial |
- 150 MB/s |
- 600 MB/s |
- Thin, round ribbon
- 4-pin
- 1-meter length |
- Single drive
per channel
- Point-to-point connection
- Full bandwidth per drive |
| SCSI |
- Parallel |
- 320 MB/s |
- None planned |
- Wide, round ribbon
- 68-pin
- 12.5 meter (LVD) length |
- Up to 15 devices
per channel |
| SAS |
- Serial |
- 300 MB/s |
- 1200 MB/s |
- Thin, round ribbon
- 6-meter length |
- 128 devices
- Expanders allow up to 16,000 devices |
Why SATA?
SATA has been developed as a backward compatible, evolutionary replacement
for ATA. Employing a serial technology version of the ATA design, SATA
offers compelling technology, performance, and usability benefits for
data-intensive applications in direct-attached storage environments.
Within the next three years SATA will replace ATA/IDE as the low-cost
interface-of-choice.
Serial ATA
| Features |
Benefits |
- High performance roadmap
(1.5 to 6.0 gigabits/sec) |
- Scalable performance
growth |
| - Lowest-cost per megabyte |
- Wide market appeal |
| - Command optimization |
- Makes SATA RAID more
practical |
| - Point-to-point connections |
- Greater data reliability |
| - Full backward compatibility |
- Easier, faster, cheaper
migration |
| - Single thin 1-meter cable |
- Greater flexibility;
space savings |
| - Backplane connection
Hot-plug/hot-swap flexibility |
- Hot-plug/hot-swap flexibility |
When to Choose SATA
The low-cost/high-benefit nature of SATA makes it an ideal fit for budget-conscious
desktop, high-end workstation, and entry-level server users whose application
needs require high performance without either the additional robustness,
or the external connectivity features of SCSI technologies.
| Interface |
Cost |
Optimal Data Type |
Storage Environment |
Application Environments |
| ATA/IDE
|
- Low |
- Reference
data: low frequency access, sequential data; e.g., file sharing,
email, web, backup, archive |
- Internal
DAS |
- Desktop PCs |
| SATA |
- Desktop PCs,
workstations,
entry-level servers |
| SCSI |
- Moderate |
- High-frequency
transactional & random access data; e.g., database, online
purchases, OLTP, CRM |
- Internal
DAS & External NAS/SAN |
- Mission critical
enterprise servers, networked storage |
| SAS |
- Mission critical
enterprise servers, large-scale networked storage |
Glossary
| ATA |
"Advanced Technology
Attachment," a storage interface designed over 15 years ago and
now the de facto I/O standard for desktop PCs. Though adequate
for low data-demand applications, the combination of increased
CPU capabilities, greater application through-put demands, and
faster, more capable hard drives, severely limits the future usefulness
of this interface. |
| Command
Optimization |
Commands to a
device are queued for immediate execution, without having to wait
for responses, increasing performance and making RAID more practical. |
| CRM |
"Customer Resource
Management" software. |
| IDE |
"Integrated Device
Electronics," the current low-cost storage interface standard
for desktop and portable PCs, synonymous with ATA. |
| NAS |
"Network Attached
Storage," a storage design that connects a server to externally
enclosed hard drives via a local area network. |
| OLTP |
"Online Transaction
Processing" database. |
| Parallel
Technology |
A design that
allows a device (hard drive) to receive multiple bits of information
at the same time. Parallel interfaces use short, wide cables carrying
multiple signals, and pose inherent design limitations on data
transfer speed and multiple device connections. |
| Point-to-Point |
Direct connection
between the backplane and the storage device, allowing for the
high-performance, full utilization of bandwidth. |
| SAN |
"Storage Area
Network," a storage design that connects all the storage devices
on a network with all the servers on a network for enhanced reliability
and performance. |
| SAS |
"Serial Attached
SCSI," the serial imple-mentation of the SCSI standard, providing
greater flexibility, performance, reliability, and connectivity. |
| SATA |
"Serial ATA" is
an evolutionary replacement for the Parallel ATA physical storage
interface. |
| SCSI |
"Small Computer
System Interface," the predominant storage I/O technology for
high-reliability, high-performance server applications. |
| Serial
Technology |
A design that
allows data to be sent one bit at a time. Serial interfaces use
thin cables, and are capable of faster speeds, greater reliability,
and more flexibility in attaching multiple drives. |
What's Hot-Swap?
The term "Hot Swap" refers to the common practice of either
inserting, or removing SCSI disk drives in an operating bus, typically
used in RAID subsystems or JBOD (just of a bunch of disks) environments.
The ability to "Hot Swap" a disk drive is beneficial to customers.
It allows them to remove potentially defective drives from the system,
or upgrade capacity without having the inconvenience and expense of
taking the entire system down to replace the drive. The ANSI documents
cover this function under the chapter heading "Removal and Insertion
of SCSI devices". Four distinct levels of functionality are defined
in Table A.
The term "Hot Swap" is not actually defined in the ANSI standards,
or the draft standards under development. It is interpreted as "the
very restrictive Level 4 Removal and Insertion of disk drives." To
avoid confusion, the two ter ms are linked together as "Level 4 Hot
Swap."
The main difference between Level 4 and the easier levels is that
the bus is allowed to operate (move data or operate in any legal SCSI
bus phase). Since inserting a disk into any powered bus will result
in some level of electrical transients, it is necessary to insure
that those transients do not interfere with, or corrupt the control
of data signals present on the bus.
Defining transfer rate
ATA-33: 33 MB/ sec.
ATA-66: 66 MB/ sec.
ATA-100: 100 MB/ sec.
What is SCSI? - Small Computer
System Interface
SCSI (pronounced "scuzzy") stands for Small Computer System
Interface, the technology that allows you to connect various internal
and external devices to your PC or PC server. This connection is made
using a SCSI card that fits inside your computer.
The Types of SCSI
| Type |
Speed |
Hard drive/peripheral
connections |
Ultra320 SCSI
(16-bit Wide) |
320 MByte/sec |
State-of-the-art hard drives |
Ultra160 SCSI
(16-bit Wide) |
160 MByte/sec |
Hard drives |
Ultra2 SCSI
(16-bit Wide) |
80 MByte/sec |
Hard drives |
Ultra Wide SCSI
(16-bit Wide) |
40 MByte/sec |
Hard drives
and tape drives |
Ultra SCSI
(8-bit Narrow) |
20 MByte/sec |
CD-R, CD-RW,
tape, removable storage (Jaz), and DVD drives |
SCSI-2, Fast
SCSI
(8-bit Narrow) |
10 MByte/sec |
Scanners, Zip
drives, and CD-ROM |
The Benefits
of SCSIPerformance
- Supports up to 320 MByte/sec transfer rates per channel with Ultra320
SCSI
- Connects high-performance devices such as hard disk drives, CD-RWs,
and other high-speed peripherals to your PC
Connectivity
- Connectivity for internal and external SCSI devices
- Single SCSI card can connect up to 15 devices per channel
Compatibility
- Accommodates previous generations of the same technology
- SCSI allows older peripherals to co-exist with the latest technology
without hampering speed or performance
Reliability
- SCSI has traditionally been the most reliable choice for IT professionals
with regard to data integrity, component failure, and product quality
- SCSI presence as an I/O choice over many years attests to its
continued execution
The Newest SCSI
Features
Features added with Ultra320 SCSI:
- 320 MByte/sec performance per channelPacket Protocol and its reduction
in command overhead allows increased speed without bandwidth issuesQuick
Arbitration Select (QAS) increases bus utilization by streamlining
release and re-use of the bus by the various peripherals
- Cyclic Redundancy Check (CRC) for all SCSI bus phases. CRC improves
data integrity by detecting data integrity errors. Previous versions
of SCSI only checked the data phase.
Server Technology Comparison
| SATAFibre Channel |
SCSI |
| Best suited for |
Entry-level
to mid-range serversServer-to-server, campus networks |
Mid-range to
enterprise servers |
| Advantages |
Performance
of first-generation Serial ATA products:
150 MByte/sec
Expected low cost, but still an emerging technology Performance:
200 MByte/sec
Hard drive reliability
Highest hard disk drive expandability
|
Performance:
320 MByte/sec per channel
High hard drive and peripheral reliability
Connectivity to the largest variety of peripherals
Expandability
|
Single-User Technology Comparison
| USBATA |
SCSI |
| Best suited for |
Basic desktopBasic
desktop |
Performance
desktop/workstation |
| Advantages |
No added cost*
Easy, external connectivity for simple devices like joysticks,
keyboards, mice, and entry-level scanners No added cost*
Industry-standard interface for connecting internal devices
such as hard disk drives and CD-ROMs
|
Highest performance
Highest device reliability Connection to the largest variety
of peripherals
Expandability
|
*ATA and USB connections are standard on all
new WindowsR computer systems.
Glossary
Basic Input/Output System (BIOS): A motherboard
BIOS controls the basic functions of the computer such as the keyboard,
monitor, etc. The BIOS on a SCSI card is used to control SCSI hard
disk drives and perform the hard disk boot function.
Bus Mastering: The ability to process SCSI commands
on the SCSI card due to its built-in processor without using the system?s
CPU.
Direct Memory Access (DMA): The fastest method of
data transfer available for multitasking operating systems. Data is
transferred from SCSI devices to system memory (RAM) via the SCSI
card without using the system?s CPU.
Daisy Chain: A cable configuration in which internal
and external SCSI devices such as hard drives, CDs, scanners, and
tape drives are connected in a series to the SCSI card.
Input/Output (I/O): An operation, program, or device
that enters data into or extracts data from a computer.
Low Voltage Differential (LVD): Introduced with
Ultra2 SCSI, and further enhanced with Ultra160 and Ultra320 SCSI,
LVD technology enables data transfer to 320 MByte/sec and supports
cable lengths to 12 meters.
Quick Arbitration Select (QAS): An Ultra320 SCSI
feature that improves the control release between SCSI devices, reducing
command overhead.
Packet Protocol/Packetized SCSI: This Ultra320 SCSI
feature speeds the transfer of data, command, and status packets over
earlier generations of SCSI.
SCSI Bus: A host adapter and one or more SCSI peripherals
connected by cables in a linear chain configuration. The host adapter
may exist anywhere on the chain, allowing connection of both internal
and external SCSI devices. A system may have more than one SCSI bus
by using multiple host adapters.
SCSI ID: A unique identification number used for
each device on the SCSI chain.
Termination: A feature that stops the data signal
at the beginning and the end of the SCSI bus. The first and last devices
on the SCSI bus must be terminated.
What's Fiber?
A high-speed, high-bandwidth serial protocol for channels and networks
that interconnect over twisted-pair wires, coaxial cable or fiber
optic cable. The "fabric" topology of Fiber Channel offers up to 16
million ports with cable lengths of up to 10 kilometers. SCSI will
use the lower cost "Arbitrated Loop" topology (FC-AL) of Fiber Channel.
FC-AL using fiber optic media offers speeds of up to 100 MBytes/sec
and up to 127 ports all connected in serial with up to 25 meters between
ports. Fiber Channel on copper wiring is available in several versions
from 12.5 MBytes/sec with up to 100 meters of cable to 100 MBytes/sec
with up to 25 meters of cable. Does not require ID switches or terminators.
The FC-AL loop may be connected to a Fiber Channel "fabric" for connection
to other nodes. SCSI on FC-AL will be expensive and will require some
changes to software as well as hardware.
Fibre Channel SCSI
This refers to products with fibre channel physical and protocol
layers using the SCSI command set. The Fibre Channel interface is
completely different from parallel SCSI in that it is a serial interface,
meaning command and data information is transmitted on one signal
stream organized into packets. The fibre may be either a copper coaxial
cable or a fiber optic cable. The signal on the first implementation
of fibre channel uses a 1 GHz rate, thereby achieving 100 Mbytes/sec
over the cable. Fibre channel also implements increased software control
of configuration and pushes the total device count on the bus to 126
IDs, as opposed to only 8 or 16 on a parallel bus.
Data Transfer Defined:
SCSI I & II: 4 MB/ sec.
Fast SCSI: 10 MB/ sec.
Ultra SCSI & Fast Wide SCSI: 20 MB/ sec.
SCSI III (or Ultra Wide SCSI): 40 MB/ sec.
Wide ULTRA II SCSI (LVD): 80 MB/ sec.
ULTRA III & 160: 160MB/ sec.
Connector Interfaces
IDE: uses 40 pin connector
SCSI I: uses 25 pin connector
SCSI II: uses 50 pin connector
SCSI III: uses 68 pin connector
SCA or SCA2: uses 80 pin connector
How much memory do I need?
How can you tell when a server requires more memory? Quite often,
the users of the network are good indicators. If network-related activity
such as email, shared applications, or printing slows down, they'll
probably let their Network Administrator know. Here are a few proactive
strategies that can be used to gauge whether or not a server has sufficient
memory:
- Monitor server disk activity. If disk swapping is detected, it
is usually a result of inadequate memory.
- Most servers have a utility that monitors CPU, memory, and disk
utilization. Review this at peak usage times to measure the highest
spikes in demand.
Once it's determined that a server does need more memory, there are
many factors to consider when deciding on how much is enough:
What functions does the server perform (application, communication,
remote access, email, Web, file, multimedia, print, database)?
Some servers hold a large amount of information in memory at once,
while others process information sequentially. For example, a typical
large database server does a lot of data processing; with more memory,
such a server would likely run much faster because more of the records
it needs for searches and queries could be held in memory - that is,
"at the ready." On the other hand, compared to a database server,
a typical file server can perform efficiently with less memory because
its primary job is simply to transfer information rather than to process
it.
What operating system does the server use?
Each server operating system manages memory differently. For example,
a network operating system (NOS) such as the Novell operating
system handles information much differently than an application-oriented
system such as Windows NT. Windows NT's richer interface requires
more memory, while the traditional Novell functions of file and print
serving require less memory.
How many users access the server at one time?
Most servers are designed and configured to support a certain number
of users at one time. Recent tests show that this number is directly
proportional to the amount of memory in the server. As soon as the
number of users exceeds maximum capacity, the server resorts to using
hard disk space as virtual memory, and performance drops sharply.
In recent studies with Windows NT, additional memory allowed an application
server to increase by several times the number of users supported
while maintaining the same level of performance.
What kind and how many processors are installed on the server?
Memory and processors affect server performance differently, but they
work hand in hand. Adding memory allows more information to be handled
at one time, while adding processors allows the information to be
processed faster. So, if you add processing power to a system, additional
memory will enable the processors to perform at their full potential.
How critical is the server's response time?
In some servers, such as Web or e-commerce servers, response time
directly affects the customer experience and hence revenue. In these
cases, some IT Managers choose to install more memory than they
think they would ever need in order to accommodate surprise surges
in use. Because server configurations involve so many variables,
it's difficult to make precise recommendations with regard to memory.
The following chart shows two server upgrade scenarios.
SERVER MEMORY MAP
| WINDOWS® 2000 SERVER
|
Designed to help businesses
of all sizes run better, Windows 2000 Server offers a manageable,
reliable and internet-ready solution for today's growing enterprises.
For optimal performance, consider adding more memory to take advantage
of Windows 2000 Server's robust feature set. Windows 2000 Server
is internet-ready and promises to run today's and tomorrow's applications
better.
Baseline: 128MB
Optimal: 256MB - 1GB
|
|
| Application Server Houses
one or more applications to be accessed over a wide user base |
256MB - 4GB |
| Directory ServerCentral Management
of network resources |
128MB - 1GB |
| Print ServerDistributes print jobs
to appropriate printers |
128MB - 512MB |
| Communication ServerManages a variety
of communications such as PBX, Voicemail, Email, and VPN |
512MB - 2GB |
| Web ServerInternet and intranet
solutions |
512MB - 2GB |
| Database ServerManages simple to
complex databases of varying sizes |
256MB - 4GB |
|
Microsoft Windows ServerTM 2003
|
| Memory |
Minimum |
Recommended |
Maximum |
| Standard |
128MB |
256MB |
4GB |
| Enterprise |
128MB |
256MB |
32 GB for x86-based computers
512 GB for Itanium-based computers*
|
| Datacenter |
512MB |
1GB |
64 GB for x86-based computers
512 GB for Itanium-based computers*
|
| Web |
128MB |
256MB |
2GB |
* Important: Per Microsoft, the 64-bit versions of Windows Server
2003, Enterprise Edition and Windows Server 2003, Datacenter Edition
are only compatible with 64-bit Intel Itanium-based systems. They cannot
be successfully installed on 32-bit systems.
| LINUX® |
Linux is a reliable, cost-effective
alternative to traditional UNIX servers. Depending on the distribution,
the Linux server platform features a variety of utilities, applications,
and services.
Baseline: 64MB - 128MB
Optimal: 256MB - 1GB
|
|
| Application ServerHouses
one or more applications to be accessed over a wide user base |
64MB - 4GB |
| Directory ServerCentral Management
of network resources |
128MB - 1GB |
| Print ServerDistributes print jobs
to appropriate printers |
128MB - 512MB |
| Communication ServerManages a variety
of communications such as PBX, Voicemail, Email, and VPN |
512MB - 2GB |
| Web ServerInternet and intranet
solutions |
512MB - 2GB |
| Database ServerManages simple to
complex databases of varying sizes |
256MB - 4GB |
* Please Note: These figures reflect work done in a typical server
environment. Higher-end workstation tasks may require up to 4GB. Naturally,
a chart such as this evolves as memory needs and trends change. Over
time, developers of software and operating systems will continue to
add features and functionality to their products. This will continue
to drive the demand for more memory. More complex character sets,
like Kanji, may require more memory than the standard Roman based
(English) character sets.
System Design Computers
More questions, please contact our tech support: sdcom@sdcom.com
|