Audioexam.com - For a Limited Time only, download all 25 of our Audio Guides for $99!!!!!
Over 24 Hours of content!! 

SERVER+ STUDY GUIDE

This study guide is organized to address the objectives listed by CompTIA. Do NOT solely rely on this study guide for exam preparation. More readings and hands on will be needed to ensure a high passing score in the exam

1. Describe Physical Drives and Logical Drives and describe their functionality.

A “Physical” Drive is the hardware drive. The first physical drive will have a letter of C. If you add another physical drive, the next available drive letter will be assigned to the new drive. In theory you can have up to Z drive.

A “Logical” drive is a partition. You segment your physical drive into multiple logical drives. Each logical drive will occupy one drive letter.

You store information on the logical drives. However, when the corresponding physical drive is removed, the logical drive will be gone too.

2. Identify and describe the differences, strengths and specifications of the following SCSI solutions

a. SCSI-1, 2, & 3

SCSI 1 has a 8-bit parallel interface between the host adapter and the device. The speed is 5 MHz, with a max data transfer rate of 5 MBps. SCSI 2 has wider data paths with a higher transfer rate of up to 10mbps. SCSI 3 doubles the SCSI clock speed and achieve a transfer rate of 160 MBps. The number of devices that can be supported is increased as well.

The WIDE bus doubles the original 8-bit bus width to 16 bits. As a result, the peak transfer rates has been increased to 10 MBps. Also, max 16 devices can be supported per bus.

Fast SCSI-2 doubles the clock rate from 5MHz to 10MHz to increase the throughput. It has a 8-bit Bus and a max transfer rate of 10 MBps. Ultra has a max bus speed of 20MHz, still using 8-bit bus, but has a 20 MBps transfer rate. Ultra Wide uses a 16-bit bus with a 40 MBps Transfer rate.

Ultra 2 is also called FAST40, and has increased the bus clock to 40MHz under the condition that Differential Signaling must be used. Ultra 160 has a clock speed of 80 MHz and a max transfer rate of 160 MBps. The number of devices supported per bus has been increased to 16 as well. Ultra 320 is the latest technology with clock speed and transfer rate nearly doubled as well.

d. SCSI connectors, cables, termination

e. SCSI IDs and LUNs

f. Single Ended Devices

• Each device is given a distinct address or SCSI ID, ranging from 0 to 7 for a narrow (8-bit) bus or 0 to 15 for a wide (16-bit) bus.
• Devices that request I/O processes are called initiators.
• Devices that perform operations requested by initiators are called targets.
• Each SCSI address is a target.
• Each target can accommodate up to 8 other devices - known as logical units
• Each target is assigned a logical unit number LUN.
• Commands to the SCSI controller identify devices according to their LUNs
• The SCSI bus requires termination at both ends - leave a resistor connected on the circuit board - resistors are shipped with each SCSI unit as it ships
• In a typical setting, the host adapter is one end of the chain, and the final device on the SCSI cable is the other.
• The bus must contain two and only two terminators
• If a device on the bus is single-ended only, the entire bus will switch to single-ended mode

Regarding SCSI cables:

• SCSI-1 External A cable is the 50 pin Centronics connector cable. (C50M) 85 mil (2.16 mm) spacing
• SCSI-2 External A cable is the 50 pin High Density connector cable. (MiniD50M) 50 mil (1.27mm) Spacing
• SCSI-3 External P cable is the 68 pin High Density connector cable. (MiniD68M) 50 mil (1.27mm) Spacing
• SCSI-1, SCSI-2 Internal A cable is the 50 pin connector Ribbon cable. 100 mil (2.54mm) spacing
• SCSI-3 Internal P cable is the 68 pin High Density ribbon Cable. 50 mil (1.27mm) Spacing

g. Low Voltage Differential (LVD)

• Short for Low Voltage Differential
• Not defined in the original SCSI standards
• Provides better reliability and common-mode noise immunity.
• Reduces EMI and power requirements.
• Provides compatibility for SE and HVD devices

h. High Voltage Differential (HVD)

• Short for High Voltage Differential
• Not compatible with single ended devices.
• Isolated from the reference ground.
• Uses twisted-pair wiring with two wires to drive one signal
• allows for a 25 meters bus length

i. BUS lengths

3. Identify and describe the differences, strengths and specifications of IDE

a. ATA 33

b. ATA 66

c. ATA 100

d. Ultra DMA

e. Cabling and connectors

f. Master/slave/cable select (CSEL)

g. Jumper settings

ATA (AT Attachment) and IDE (Integrated Drive Electronics) are identical: disk drive implementation designed for integrating the controller onto the drive itself to reduce interface costs. ATA-2 is a compatible extension of ATA (IDE), with performance enhancing features such as fast PIO and DMA modes. ATA 33 is often called Ultra DMA, and can support multiword DMA mode 3 running at 33 MBps. ATA 66 has a max throughput of 66 MBps, while ATA 100 pushes the data transfer rate to 100 MBps.

DMA shorts for direct memory access. It is a technique for transferring data from main memory to a device without passing it through the CPU. Computers that have DMA channels can transfer data to and from devices much more quickly. Ultra DMA is an enhanced version of the DMA technology.

Each IDE channel can support either one or two devices. To differentiate between devices on a channel, you designate devices to be either a master or a slave. This is typically done by setting the jumpers on the drives. To avoid the potential troubles in setting the jumpers, cable select (CSEL) allows the cable connectors to determine which device is master and which is slave. To use CSEL,  both drives must first have the jumpers set to CSEL, and a special 3 connector cable with one connected to the motherboard is needed.

4. Define the features, capabilities, and implementation of the following RAID levels and explain how each relates to fault tolerance or high availability: (non-proprietary)

a. RAID 0

b. RAID 1

c. RAID 5

d. RAID 0/1

5. Describe the benefits of Hardware RAID over Software RAID

6. Describe hot swap, hot plug, and redundant capability of different components

7. Describe fail over and hot spare

RAID shorts for Redundant Array of Inexpensive Disks. Several drives are grouped together for a variety of purposes, the most obvious being redundancy - if one of the drives should fail. By grouping several drives together, they can be configured to appear to the computer system as one large drive. The grouping of drives together with various RAID levels can yield some amazing performance features as well.

The RAID subsystem can tolerate the failure of a single drive and continue working. If it is configured with hot spare, the RAID subsystem can begin recreating the data that was from the failed drive to the spare drive. With hot swap, the defective drive can be removed and replaced by a functioning unit.

RAID0 and RAID1 require at least 2 drives. RAID0+1 (also known as RAID10) needs a minimum of 4 drives. All other RAID levels requires at least 3 drives. Note that RAID0 has no redundancy at all.

RAID 0 offers the good performance with no redundancy. RAID 1 has two drives that duplicate the storage of data. RAID 5 uses striping with parity to allow for overlapped read and write operations. It uses parity information to reconstruct data. RAID 0+1 has an array of RAID-1 array of drives. It has the highest performance and redundancy at a very high cost. 

Hardware based RAID is much better in terms of performance, as dedicated processor is used for the RAID operation. Software based RAID is subject to OS support, and will consume CPU time.

8. Describe clustering, scalability, and high availability

To build a cluster, you have a group of servers attached to a common control unit and have them back each other up if one fails. With Clustering, you can achieve the following:

• Load sharing among the servers – allows for high scalability, as you can always add more servers into the cluster to share the increased work load
• Fault Tolerant – provides High Availability - 99.9% of the time the system is working fine for the end users.

Some cluster technology uses high speed fiber connections to connect identical servers together with the need of a shared drive pool.

9. Identify the basic attributes, purpose, function and advantages of Fibre Channel

Fibre Channel technology was developed to address I/O bandwidth problems associated with the deployment of centralized data storage architectures in enterprise server networks. It provides very high-speed data transfer between server and storage components. The combination of Fibre Channel and RAID technologies provides the speed, flexibility, and data-storage capacity necessary.

A Fibre Channel storage subsystem employs long-wave fiber-optic cable between cascaded switches and servers of up to 10 kilometers from the shared storage array. It delivers up to 100 megabytes per second of bandwidth for half-duplex configurations and up to 200 MB/sec for full uplex configurations.

Fibre Channel Arbitrated Loop, known as FC-AL, is the most prominent Fibre Channel standard that uses optical fiber to connect SCSI devices. It has the potential to replace SCSI on very high performance server systems.

10. Explain the differences, strengths and specifications of CPU

For example, a P133 equals 66 x 2. 2 is the internal clock multiplier, 66 is the bus speed. There are many variety of bus speeds available, ranging from 66mhz to 133mhz. Usually a system with a higher bus speed will require memory of higher speed. Also, cooling is important as a fast CPU may get overheated easily. Heat sink and cooling fan is required, with case fan to increase the overall air circulation inside the case.

Nowadays the CPU usually has a locked multiplier. Bus speed support is solely depending on the motherboard. The flexibility in adjusting bus speed has made “overclocking” a popular topic nowadays. 

More information about different CPUs can be found at:

www.tomshardware.com

www.anandtech.om

11. Describe Multiprocessing functionality

• Multiprocessing involves the coordinated processing of two or more computer processors.
• SMP shorts for Symmetric Multiprocessing - processors share both memory and I/O bus
• MPP shorts for Massively Parallel Processing – every single processor has its own memory and bus, and has a much high scalability than SMP – we refer to this as a  "share nothing" architecture
• Intel CPUs use SMP
• In SMP, for every CPU added there is approximately a 70% increase I performance

12. Define the following cache memory and explain how it works in a server

a. Processor cache

• SRAM
• Of small amount typically due to technical difficulty in fitting into the die
• Between CPU and main memory
• VERY FAST
• Level 1 cache is built into the CPU
• Level 2 cache is on board for the 586n family, but is included in the chip since the 686 generation
• Level 2 cache is usually slower and cheaper
• Due to the lower cost and technical difficulty, Level 2 cache size is usually larger than Level 1 cache 

b. RAID cache

Cache memory acting as buffer to increase performance for the RAID subsystem.

c. Write back vs write thru

• With Write-back Cache, the changes to data stored in the L1 cache will not be copied to the main memory unless strictly necessary. Since it involves much less write operations to the main memory, the performance is usually better. 
• With Write-through Cache, data is written to the main memory and the L1 cache all at once.

13. Describe the characteristics of the following types of memory and server memory requirements:

Divide the memory into multiple sections for CPU access and reduce the delay involved for the wait states. Slower memory may use this technology to “catch up”.

Shorts for  "error checking and correcting", it checks data for errors and corrects them immediately.

Shorts for Extended data out, this is a RAM chip that will increase the read performance. Most of the 586 series processors required this kind of memory as a minimum requirement. For the 686 generation, faster SDRAM is needed instead.

d. Unbuffered vs buffered vs registered

• Unbuffered memory - chipset controller works directly with the memory without any “middleman”.
• Buffered - uses a buffer to isolate the memory from the controller for the purpose of reducing the load brought to the chipset by the large amount of RAM
• Registered - controlled delay on heavy loaded memories for the clock cycle

Often known as HCL, Microsoft provides a list of officially tested and certified for compatibility hardware as a reference. It is always safe to purchase hardware listed on the HCL. 

14. Identify the basic attributes, purpose, function and performance of the following System Bus Architectures

Data are traveling on the bus without the need for CPU involvement. This will reduce the workload of the CPU.

Allows you to add devices to or remove devices from a system while the system is running, without the need of shutting down the system at all.

Use a two-level hierarchical bus - Primary level with 32bit PCI and better performance, and the Secondary level with weaker performance.

Dual PCI buses with all devices appear in IRQ.

Shorts for Intelligent I/O, it is developed by the I2O SIG to eliminate I/O bottlenecks with the assistance from special co-processors that can handle the miscellaneous works like interrupt handling, data transfer and buffering.

PCI with hot-plug capabilities.

h. PCI Interrupts

Newer motherboard will have more PCI slots and fewer ISA slots. A PCI bus has its own interrupt system to handle requests. Sometimes these are mapped to the regular interrupts to serve other cards.

Enhanced ISA, an improved version of ISA bus, is never gaining popularity.

Shorts for Accelerated Graphic Port, it is now the standard in the next generation computer platforms. With a fast, wide and direct bus between the graphic processor and the memory,  AGP greatly enhances the 3D display power. For servers, however, AGP is of little use.

15. Explain how the System Bus affects the performance of a server

A faster system bus increases the overall system performance. No matter how fast the CPU is, the bus can introduce all different kinds of overhead, which eventually becomes the bottleneck. 

16. Identify the basic attributes, purpose, and function of the following management Protocols

Shorts for Simple Network Management Protocol, it allows for centralized network management. Devices on the network which have SNMP agents running will work with the central SNMP management station for performance measurement, statistic collection, resource monitoring and software installation etc.

b. DMI

Stands for Desktop Management Interface. It was created by the Desktop Management Task Force to automate system management. Management Interface allows DMI-enabled applications to directly access, manage, and control desktop systems. DMI Service Provider collects information from products, manages the information in the Management Information Format database, and passes the information to management applications via the Management Interface. The Component interface allows components to be seen and managed by numerous applications that call the DMI. The Management Information Format defines the standard manageable attributes of products.

17. Identify the basic attributes, purpose, and function of the following types of servers

For a server to act as a gateway, it means the server will have at least 2 NICs, and will perform some sort of protocol translation functions.

For a server to act as a Router, it means the server will have at least 2 NICs, and will perform some sort of routing functions. TCP/IP must be installed, and routing table must be set up either manually by hand or by using a routing protocol like RIP or OSPF.

A server can act as a bridge that connects different and dissimilar LAN segments. Again, you need to have at least 2 NICs for this purpose.

For a server to act as a Firewall, it means the server will have at least 2 NICs, and will perform some sort of packet inspection and filtering functions. TCP/IP must be installed, and a firewall software must be running. Traffics must go through this firewall in order to be checked.

You usually set up a server to act as a Proxy Server if multiple computers must share a single connection to the internet. In this case, the proxy server will have one NIC to connect to the internet, with another connected to the internal network. The proxy clients can use the private IP addresses, while the Proxy Server will use port mapping to identify traffics. Additionally, Proxy Server can be configured to filter traffic.

A server that hosts a database server, such as SQL Server or Oracle.

The computing model which has the client and server share the processing load.

A Server oriented computing model which have almost all of the computing process running on the server.

A server for network messaging.

A server for downloading and uploading files.

A server for connecting to the IBM mainframe network. SNA is the IBM protocol very popular among the “big computers”.

A server with modems for remote clients to dial in and connect to the LAN.

A server that shares its disk and printer resources for the network clients to use.

A server that shares its fax modem for faxing or receiving faxes.

A server that provides Host Name to IP resolution for clients.

A server that provides NetBIOS Name to IP resolution for clients.

A server that provides dynamic IP address assignment for clients.

A server that hosts a web site.

18. Identify the function of the following models of Application Servers

a. Dedicated Application

b. Distributed Application

c. Peer to peer Application

19. Define Backup, Restore, and Disaster Recovery Concepts

There are several types of backup and restore. Full backup backups everything, and can be very time consuming. Incremental backup only backups the changes since yesterday backup. Differential backup also backups the difference - not the difference since yesterday, but the difference since the last full backup. Incremental backup has the shortest backup time, while differential backup has the shortest restore time. 

Proactive measures are always recommended over reactive measures towards disasters. RAID and cluster provides fault tolerance, and regular backups provide contingency. Backup media should be taken offsite everyday. UPS should be deployed just to be ready in case of power outrage.

20. Describe Security Concepts (Importance of physically securing a server)

A.     Servers should be locked in a server room

B.     Client computers can be floppy-less to prevent users from copying important data

C.    Smart card can be used for log on authentication

21. Describe Shutdown Procedures across the following network operating systems

a. Novell NetWare

b. Microsoft Windows NT/2000

c. UNIX/Linux

d. IBM OS/2

Basically all NOSs will do the following: when you are to shut down the system, you should alert the users and urge them to disconnect. If you do not want to wait, you may continue to shut down, but the users may have data lost.  You always follow the formal shut down procedure. NEVER simply switch off or reset the computers.

22. Describe the basic concepts for the following aspects of rack mounting servers

In situations where multiple PC Servers are used, you use KVM switches to reduce hardware costs by eliminating the need for a keyboard, monitor, and mouse dedicated to each server. KVM allows for one centralized point of control for multiple servers.

Different types of cable racks can be used to ensure the tidiness of cables.

d. Security

This is about carefully laying the equipments and physically securing the server hardware.

23. Describe when BIOS/Firmware upgrade should be performed

When there is device incompatibility, you may want to download the latest BIOS and patch the BIOS by “flashing” it. Due to the constant introduction of new devices, BIOS may have to be frequently updated.

Firmware is software that has been written onto read-only memory. It is a combination of software and hardware. ROMs, PROMs and EPROMs that have data or programs recorded on them are firmware.

Before you modify anything, make sure there is a standby server that can be used in case you corrupt your server.

24. Describe the importance and use of maintenance logs and service logs (documentation)

Logs allow you to trace back and find out what had happened to the system. Documentation allows you to document the scenario you are facing, and also allows you to document the way to solve this problem. This can be of great value for future reference. 

25. Identify the major features of the following Server Operating Systems

In Version 5, it has a JAVA GUI console. It runs IP natively. It relies on the famous NDS Novell Directory Service as the network logical infrastructure. Admijnistration can be done on the server via the Java console or from the clients via NWADMIN.

A complete GUI based NOS. In 4.0, the domain model has serious limitations (40000 user per domain), and the configuration of trust can become very complicated. Windows 2000 supports IP natively, and has a much better domain support. Active Directory is the Microsoft only directory structure which involves the concept of forest, tree, domain and OUs. 

UNIX is known for its rock solid stability and reliability. Linux can be said as a descendent of Unix, and is based on the spirit of “open standard”. Source codes of Linux are freely distributable, and there are many different “distributions” available in the market. In general all Unix and Linux are command line based, but there are variations of the GUI X Windows available to come to the rescue.

Once a threat to the Microsoft Windows, its market share has been declined to an extent that it is nearly “dead”.

26. Define the impact of domains and directory services on server hardware design

Domain controllers are important computers in the network responsible for authentication and maintaining the Active Directory. You want to be sure that your hardware platform can run reliably, and at the same time providing the horse power to meet the demand of the network. Dual or quad processors are recommended. The NICs should be running 100Mbps full duplex to take advantage of the 100BaseT network commonly found nowadays.

27. Describe the following common network interface protocols

Running at 10mbps, 10BaseT is a good example of modern Ethernet implementation. CSMA/CD is used for collision detection. Cat 3 cable with RJ45 is used for implementing a 10BaseT network.  A 10/100 Hub is enough for this speed.

The following also fall into the Ethernet category: 10Base2, 10Base5

Running at 100mbps, 100BaseT is a good example of Fast Ethernet implementation. CSMA/CD is used for collision detection. Cat 5 cable with RJ45 is used for implementing a 100BaseT network. Many office buildings are pre-wired for 100BaseT connectivity. For maximum performance, a 100 switch or switching hub should be used with the 100BaseT network.

It is the latest standard of Ethernet, with speed up to 1000Mbps. It supports a distance of up to 100 meters by using four pair CAT 5 cable.

Use logical token passing, it is physically a bus topology. Developed by IBM, Token Ring can run at 4 or 16 Mbps.

28. Describe Network Interface Cards as they relate to the System Bus

NIC running at 10Mbps can afford to use ISA bus. For high speed NIC like 100BaseT and 100BaseFX, the faster PCI bus must be used.

29. Describe Adaptive Fault Tolerance

Often known as AFT, it provides automatic adapter redundancy via the concept of adapter teams - with two to four adapters per team, should the primary adaptor fails the second will take over the control automatically. This technology is supported by NT + SP3, Netware 4.X and Windows 2000.

30. Describe Adapter Load Balancing

Offers the benefits of AFT, plus increasing server bandwidth to max 800Mbp.

31. Describe Adapter Teaming

When you install two or more NIC’s in a server, you can group them logically and assign them one virtual IP. They will appear to the O/S as one single entity.

This study guide is developed in Year 2001 by Yu Chak Tin Michael.

His personal web site is located at:  http://michaelyu.freeservers.com. You may also email him at:   ycthk@i-cable.com

 Copyright 2000 http://www.CERTguide.com/