business data communications chapter eight servers in the enterprise

39
Business Data Communications Chapter Eight Servers in the Enterprise

Upload: theodore-lester

Post on 03-Jan-2016

219 views

Category:

Documents


2 download

TRANSCRIPT

Business Data Communications

Chapter Eight

Servers in the Enterprise

Primary Learning Objectives

Differentiate between clients and servers Describe five server physical components Identify major server types Understand the concept of well-known ports Explain server clustering Define “system area network”

Clients and Servers

Networks commonly make use of both clients and servers:

Clients request networked resources from servers Servers manage and distribute networked resources

to clients Clients and servers both have operating

systems that differentiate them from each other

Modern client operating systems are network-aware

Server operating systems are specialized

Clients and Servers

Client and Server Software

Server Physical Components

A server has a variety of physical components

Five physical components particularly important to a server’s performance are:

Physical case Memory Processor Drive interface Network Interface Card

Physical Case

A client desktop’s physical case is usually kept to a minimal size as it occupies and shares space in a user’s work environment

The case occupies a “footprint” Servers are typically secured and locked away,

with infrequent physical access by staff: Therefore, a server’s footprint can be larger than a

client’s The server’s physical case is often more rugged The server’s physical case is referred to as its chassis

Physical Case

A larger server physical chassis: Allows technicians more room for working within

the chassis Facilitates upgrading the server Provides more bays for expansion Offers more space for additional components Allows for a larger and more robust power supply

A server may require a redundant power supply, as well as its own internal power surge protector

Physical Case

Intel Server Chassis SC5250-E:

Has four fixed-drive bays and supports up to five hot-swap SCSI or four hot-swap SATA hard-drive bays.

This chassis allows adding more storage without having to take the server offline.

Expansion cards can be easily added due to the tool-free access enabled by thumb screws on side panels.

Memory

Memory is also called RAM, for random access memory

RAM “sticks” plug into slots housed on a server computer’s motherboard

Servers generally support much higher capacities of RAM than do clients

A server’s RAM slots are often able to support higher capacity RAM sticks than a client’s

Memory

Processor

A server’s processor is also called a CPU, for central processing unit

The CPU is the workhorse of the server and is critical for processing of instructions and data

Server motherboards often support multiple CPUs

If a server has multiple CPUs, its server operating system (SOS) must be able to support the number configured

If multiple CPUs are used, the SOS must be capable of managing the CPUs that reference each other’s cache

Processor

An INTEL server motherboard, model S845WD1-E

Drive Interface

Servers usually support more drives than does a typical client device

Drives include: Floppy Tape CD Hard drive

Hard drives, in particular, are a key server component

Drives require an interface to the server’s CPU

Drive Interface

Two common drive interfaces are: IDE – Integrated Drive Electronics

Is used mostly by clients Requires a controller

SCSI – Small Computer System Interface Is used mostly by servers Requires a controller The SCSI controller is often an adapter card that

plugs into a motherboard’s expansion slot The SCSI adapter card uses a parallel bus to

communicate with devices connected to the parallel bus

The SCSI adapter can support up to 15 internal and/or external devices

Drive Interface

Network Interface Card – NIC

Critical server NIC considerations include: Throughput capacity Half- or full-duplex communications Direct memory access (DMA) Bus mastering Buffers On-board processor

Major Types of Servers

If a resource needs to be shared or managed, a server can likely be configured to do it

Many types of servers are possible, including: File and application Database Web, e-mail, and FTP Domain Name System (DNS) Dynamic Host Configuration Protocol (DHCP)

File and ApplicationServers

Having duplicate files and applications on multiple devices raises numerous questions, including:

Who has the most current version of the file or application?

How can multiple copies be adequately secured? How many copies have been distributed? Who maintains these multiple copies?

File and application servers allow for centralized management, control, and distribution

Users are typically authenticated before being given access to networked file and application resources

File and ApplicationServers

Must be fast and reliable Require sufficient storage to accommodate user

needs Are often implemented in software File servers transmit the entire file from the server

to the client: For small files this may be suitable For large files this may result in a server bottleneck

Application servers control applications, using two components:

The client front-end The server back-end

File and ApplicationServers

Licensing is an additional legal and ethical consideration when sharing files and applications across the network

Three forms of licensing are: Individual Metered Site

For a file or application to be distributed across a network, it must be network-compatible

Not all applications, nor other resources, can be shared

Database Servers

Provide management access control software that makes database records available to users across the network

In most implementations, perform much of the processing, with only the result being transmitted to the client

As with application servers, require a client front-end process and a server back-end process

Allow for distributed databases

Database Server

Web, E-mail, and FTP Servers

A Web server runs application layer TCP/IP protocols

A client accessing a Web server utilizes a browser

Web pages hosted on a Web server are expressed using HTML – HyperText Markup Language

Two popular Web server programs are: Apache – an open system program Internet Information Server (IIS) – a Microsoft

program

Web, E-mail, and FTP Servers – Well-known Ports

Requests come to a Web server through “ports”

These ports are logical designations that represent a particular type of requested service

Port values are also called “well-known ports” The most common port used by a client Web browser

is well-known port 80 Well-known ports are primary targets by hackers Many application layer server requests have been

standardized to use particular well-known ports

Web, E-mail, and FTP Servers – A few Well-known Ports

Web, E-mail, and FTP Servers

A client request is initiated through a Uniform Resource Locator, or URL

For most users of a Web server the URL takes the form of a World Wide Web (WWW) address:

www.woodbury.edu A URL has four parts:

The protocol (HTTP or FTP, for example) The server’s IP or Domain Name System address An optional port number The directory or file being requested

Web, E-mail, and FTP Servers

Popular e-mail Web server protocols include: SMTP – Simple Mail Transfer Protocol POP3 – Post Office Protocol version 3 IMAP – Internet Message Access Protocol

A single physical mail server might be running multiple logical mail services

FTP provides for transfer of files from a server to a client, and vice-versa

Whether Web, e-mail, or FTP, a client component and a server component must both be configured

Domain Name System (DNS) Servers

DNS servers are critical in a TCP/IP network DNS servers resolve logical IP network layer

addresses to logical application layer domain name addresses

The DNS system is designed around a hierarchy of domains:

There are top- and second-level domains A domain can have subdomains A domain name is resolved when the DNS server

matches a host’s IP address to a domain name address

Domain Name System (DNS) Servers

Clients must be configured to know who their DNS server is.A redundant DNS server can be specified.

Domain Name System (DNS) Hierarchy

Dynamic Host Configuration Protocol (DHCP) Servers Clients in a TCP/IP network must be

configured to know their logical network layer IP address

This IP address can be manually configured or automatically configured using software:

In a small enterprise, manual configuration is often practical

In a large enterprise, with hundreds or thousands of clients, manual configuration is not practical

DHCP servers are used primarily to automate a client IP addressing configuration

Dynamic Host Configuration Protocol (DHCP) Servers

This Windows client has been configured so that DHCP has been automatically enabled.

Proxy Servers

Are used primarily to provide increased security:

A “proxy” literally means authorization to act on behalf of another

A proxy server: Intercepts an internal networked client’s request Attaches its own IP address to that request Hands the request out to the external world, so that the

external world sees the address of the proxy, not the address of the originating client

Reverses the above process when requested resources are returned

Maintains an address list of which internal client is making what request

Is often used in conjunction with a firewall

Proxy Servers

Server Clusters

Connect multiple physical servers into a single logical group:

To clients accessing the server cluster, the servers in the cluster appear to be one single server

Servers in the cluster can share their workload, resulting in load balancing of network traffic

Should one of the servers in the cluster fail, other servers in the cluster can take over its duties

This is called failover capability Servers in a cluster must have a server operating

system that supports clustering Server clusters are highly scalable

Server Clusters

System Area Networks

Are a local network designed for high-speed interconnection in cluster environments

Almost exclusively use a “switched fabric” technology:

A switched fabric refers to the way that physical ports within a switch are linked, allowing the ports to communicate and transfer data amongst themselves

Fibre Channel, a switching fabric implementation, is particularly associated with System Area Networks

In summary

Servers play a critical role in the enterprise Important server components typically include:

case, memory, processors, drive interface, and Network Interface Cards

Common types of servers include: file, application, database, Web, e-mail, FTP, DNS, and DCHP

Server clustering allows the grouping of multiple physical servers into a single logical group

System Area Networks are one way in which server clusters are being implemented