Transcript Web Server Administration

Web Server Administration
Chapter 2
Preparing For Server Installation
1
Overview





Identify server categories
Evaluate server components
Plan for system disasters and reduce
their effects
Evaluate network components
Set up IP addressing
2
Identifying Server CategoriesFile Servers




File servers emphasize sending and receiving
files
A fast disk subsystem is more important than
the processor type
File servers are useful in a Web environment
when you have a large number of static
HTML pages
File servers are also appropriate for an FTP
server
3
Identifying Server CategoriesApplication Servers




Server applications can process requests from
many users at a time
Requires extensive processing power
A DBMS is a typical example
Exchange Server does more than send e-mail
messages


Its complexity puts more burden on the processor
Servers may also combine the need for fast
processors and fast disk subsystems
4
Evaluating Server Components

Processor is usually the main focus





Speed
Cores – 8 to 16 core processors
Cache
RAM
Secondary storage


Hard disk
RAID – redundant array of independent disks
5
Processor

CPU



“Brains” of a computer
An integrated circuit
Arithmetic logic unit (ALU)


Responsible for mathematical and logical ops
Control unit (CU)

Regulates instructions
6
Processors (continued)

Clock speed



Number of pulses per second generated by
the motherboard
Sets processor’s tempo
Measured in MHz (millions of cycles per
second) or GHz (billions of cycles per
second)
7
Processors (continued)

Cache




RAM built into the CPU that’s used to store
frequently accessed data and instructions
L1 cache - Level 1 (L1) cache
L2 cache - Level 2 (L2) cache
L3 cache- Level 3 (L3) cache
8
Processors (continued)


CPU performs calculations with binary
digits (bits)
32-bit processor can store 32-bit
numbers


Correlates with the amount of data a CPU
can process per clock cycle
64-bit processor can theoretically
double the amount of data the CPU can
process
9
Physical Memory: RAM & ROM



Memory chips installed on a computer
Two major categories
RAM: random access memory


Volatile- needs constant source of power
ROM: read-only memory


Nonvolatile
firmware
10
RAM


Dynamic RAM (DRAM)
Static RAM (SRAM)




Much faster and more expensive than
DRAM
Used in L2 and L3 caches, and other
memory chips
Synchronous DRAM (SDRAM)
Double Data Rate SDRAM (DDR
SDRAM)
11
ROM




Programmable ROM (PROM)
Erasable programmable ROM (EPROM)
Electrically erasable PROM (EEPROM)
Flash memory
12
Storage Devices
13
Hard Disk Drives (HDD)

Integrated Drive Electronics (IDE) &
Extended IDE (EIDE)
Small computer systems interface (SCSI)
Fibre Channel (FC)
Serial Attached SCSI (SAS)
Serial ATA (SATA)

Redundant Array of Inexp/Indep Disks (RAID)




14
Selecting a HDD





Capacity
Data Transfer Rate
Buffer Size
Average Seek Time/Access Time
Rotational Speed
15
Network Interface Card (NIC)
NIC Type
Speed
Media
Use
Standard
Ethernet
10 Mbps
Twisted pair
(sometimes fiber)
Home computers
Fast
Ethernet
100 Mbps
Twisted pair
(sometimes fiber)
Home computers
Gigabit
Ethernet
1,000 Mbps
Fiber (sometimes
twisted pair)
Servers and
LANS
10-Gigabit
Ethernet
10,000 Mbps
Fiber
Backbone
Connections and
High-end servers
ATM
25 Mbps – 622+
Mbps
Fiber
Rare due to IP
dominance
16
Evaluating Server Components

Intel Processors
http://ark.intel.com/
GT/s stands for gigatransfers per second

AMD Processors
http://www.amd.com/en-us/products/server
17
Break

Complete EX 2-1
18
Ex 2-1
1. What are the two most popular server chip
manufactures?
2. Create a comparison table to compare two servers,
one from each of the popular manufactures and
identify the following:
Processor Name/Server name
Processor chip set
Processor Speed
Processor Cores
Cache size
Word size
RAM size
RAM type
Cost
Other interesting items
19
Disaster Planning



“Insurance plan” for server
A challenge to balance costs with benefits
fault tolerance is the objective

The ability of a system to keep running even
when a component fails
20
Disaster Assessment and
Recovery

Determine




what can go wrong
what effects it will have on the org
what it will take to fix
Hardware Failure



How to replace failed hardware
How to restore software (if applicable)
Who will diagnose hardware problems
21
Disaster Assessment and
Recovery



Software disasters are more complex than
hardware disasters
Could be a combination of application
software and the operating system
You may be directed to install software that
may not work correctly


Document installation and problems in detail
Make sure that support is available
22
System Redundancy

Multiple power supplies


UPS – uninterruptible power supply
Spare hardware


on hand as appropriate for the number of devices,
reliability, and cost
Also useful are warrantees that allow for
replacement within 24 hours
23
Disk Redundancy through
RAID

RAID 1




Data is duplicated across two drives
Also called mirroring
If there are two adapters, it is called duplexing
RAID 5 (most common)



Distributes data across a minimum of 3 drives
If any one of the drives fails, data can be
recovered from the remaining drives
Hot Swap – replace drive while server is running
24
Achieving High Availability
with Multiple Servers

Clustering - Many computers act as one





Combines for computing power
Fault tolerant
High server availability
IBM – assigns different tasks to each
Microsoft – uses load balancing cluster

distribute work
25
Setting Up Backup Systems


You should maintain a series of backups so
you can restore data from a specific date
Types of backup methods




Full – all data is backed up
Differential – backs up all files that are new or
changed since last full backup
Incremental – backs up all files that are new or
changed since the last incremental backup
Test your backup system
26
Backup Considerations

Back up the operating system


Back up special application files


DBMSs and Exchange keep files open, which cannot be
backed up without special backup software modules
Back up data files


On Windows systems, you have to make sure that the
Registry is backed up
Easy and straightforward unless user has files open
Can set up separate network just for backups
27
Evaluating Network
Components

Servers/users need to be connected




Hubs
Switches (switching hubs)
Routers
WAN connection



T-carrier
DSL
Cable
28
Hubs






Connects segments of a LAN
Contains multiple ports
Processes frames of data
Received frame is amplified to transmit
Broadcasts frame to all ports
Splits bandwidth between ports
transmitting
29
Switches (switching hubs)






Connects segments of a LAN
Contains multiple ports
Processes frames of data
Received frame is amplified to transmit
Uses MAC address to determine recipient
of frame
Uses full bandwidth for each transmission
30
Routers



Connects LAN to WAN (or other LAN)
Routes packets to other network
Often integrates





4-8 port switch
Network Address Translator (NAT)
Dynamic Host Configuration Protocol (DHCP)
Domain Name Server (DNS)
Firewall
31
IP Addressing

IPv4

32 bits – binary, 8 bits separated by period
197.64.242.100
11000101010000001111001001100100
11000101 - 01000000 - 11110010 - 01100100

IPv6 (IPng IP next generation)

128 bits – hexadecimal, 16 bits separated
by colon (8 hex numbers)
3ffe:1900:4545:3:200:f8ff:fe21:67cf
32
IP v4

32 bits =>232 =


Early decisions wasted lots of addresses



4,294,967,296 addresses
Those starting with 127 used for loopback
16,277,216 addresses, no longer available
Internal structure of bits


Network Id
Host Id
33
IP Address

Network Identifier (Network ID):


left-most bits used to identify the network
Host Identifier (Host ID):

remainder of the bits are used to identify
the host on that network.
34
Network vs Host ID

227.82.157.177



network id: 227.82.0.0
host id:
0.0.157.177
OR


network id: 227.0.0.0
host id: 0.82.157.177
35
Network vs Host ID
Could be split anywhere…
227.82.157.177 … 20 bits for network id and 12 bits for host id
36
IP Addressing


Must know where the IP address is
divided
IP addressing schemes



Convention (Classful) Addressing
Subnetted Classful Addressing
Classless Addressing
37
IP “Classful” Addressing

Divide the IP addresses into 5 classes


Class A, B, C, D, E
IP address will have a network id and
host id divided on an octet boundary
38
Class
Value of w
Network ID
Host ID
Number of
networks
Number of hosts
per network
A
1-126
w
x.y.z
126
16,777,214
B
128-191
w.x
y.z
16,384
65,534
C
192-223
w.x.y
z
2,097,152
254
N/A
N/A
N/A
N/A
N/A
N/A
D
224-239
E
240-254
Reserved for
multicast
addressing
Reserved for
experimental use
39
IP Address Classes
Network
ID Bits
Number of
Host ID Number of
hosts per
Bits
networks
network
Intended Use
A
8
24
126
16,777,214
very large organizations
B
16
16
16,384
65,534
medium-to-large organizations
C
24
8
2,097,152
254
smaller organizations
D
n/a
n/a
N/A
N/A
IP multicasting
E
n/a
n/a
N/A
N/A
Reserved for “experimental use”
40
IP Address Classes
Class
First Octet in IP
address
Usable # of
Network
bits
Number of
Networks
Number of
Hosts
Class A
1 – 127
0xxx xxxx
8-1 = 7
27-2 = 126**
>16,000,000*
Class B
128 – 191
10xx xxxx
16-2 = 14
214 > 16,000
>65,000*
Class C
192 – 223
110x xxxx
24-3 = 21
221> 2,000,000
254*
** Class A num networks less 2 (all 0’s and all 1’s have spl meaning)
* Host Id of all 0’s or 1’s has special meaning
41
Class A:
109.197.63.205
109 = 0110 1101
Class B:
152…
152 = 1001 1000
1000 0000 to 1011 1111
128 to 191
Class C:
198….
1100 0110
42
IP Address
43
IP Addressing – IPv4

IP “Classful” Addressing

Set up for organizations of different sizes
44
Rationale for Classful Addressing




Simplicity and Clarity
Reasonably Flexible
Routing Ease
Reserved Addresses
45
Problems with Classful Addressing

Lack of internal Address Flexibility



Inefficient use of Address Space


Large org given big block – doesn’t match
underlying structure
No way to create internal structure
Some waste – host size > 254… >65,00
Proliferation of router table entries

20 class c tables or 1 class B tables…
46
Break

Complete EX 2-2
47
Ex 2-2
1. What are the IP classes and give a
notable fact about each.
48
Ex 2-2 (continued)
2. Given the following addresses determine the IP class,
the network id and the host id.






192.192.192.56
204.63.100.103
92.34.240.175
252.102.66.105
147.64.242.100
63.230.34.2
49
Ex 2-2 (continued)
3. What are the advantages and
disadvantages of the class IP addressing
system?
50
IP Subnet Addressing





IP address uses 3 level structure
Network id, subnet id, host id
Network id remains fixed to the class
Host id is now divided into 2 parts
A subnet mask is provided to identify
the host id and the subnet
51
Subnetting
Subnetting a class B network
Division is based on number of sub networks and effects the max
number of host in a given subnetwork.
52
Subnet mask



Because the subnet/host division could
be anywhere – must have a bit mask to
find the subnet.
Subnet mask bits correspond to bits in
IP address
Bits used for the network & subnet are
1’s and the rest are zero’s (removes the
host bits)
53
Subnet mask
54
Subnet mask
Component
Octet 1
Octet 2
Octet 3
Octet 4
IP Address
10011010
(154)
01000111
(71)
10010110
(150)
00101010
(42)
Subnet Mask
11111111
(255)
11111111
(255)
11111000
(248)
00000000
(0)
Result of
AND Masking
10011010
(154)
01000111
(71)
10010000
(144)
00000000
(0)
154.71.144.0, is the IP address of the subnet to which 154.71.150.42 belongs
55
Setting Up IP Addressing


Subnetting the
original classes
(1 subnet per
class)
For each class, 0
bits are used for
the subnet.
56
How many bits for subnet

Dividing the host portion into 2
Each bit taken for the subnet doubles the subnets
 Each bit taken for the subnet halves the number
of hosts
1
 16 bits – 1 bit for subnet = 2 subnets or 2
15 bits for hosts = 215-2 or 32,766
or
2 bits for subnet = 22 or 4 subnets
14 bits for hosts = 214-2 or 16,382

57
Advantages of Subnetting






Better match to physical Network struct
Flexibility
Invisible to public Internet
“internet within the Internet”
No need to request new IP addresses
No routing table proliferation
58
Disadvantages or Subnetting


Largest subnet determines mask
All subnets must be the same size

Unless Variable Level Subnet masking is
used
59
VLSM
Traditional subnet masking
Variable Length subnet masking
60
Break

Complete EX 2-3
61
Ex 2-3 (continued)
1. Given the following addresses and number of bits
used for subnet, determine the subnet mask, the
number of subnets and the number of hosts.
192.192.192.56, subnet - 3 bits
204.63.100.103, subnet – 4 bits
92.34.240.175, subnet – 14 bits
252.102.66.105, subnet – 6 bits
147.64.242.100, subnet - 10 bits
63.230.34.2, subnet - 8 bits
62
Ex 2-3 (continued)
2. What are the advantages and
disadvantages of the IP subnet
addressing system?
63
IP Classless Addressing - CIDR





Classless Inter-Domain Routing (CIDR)
"Supernetting“
Eliminating the notion of address
classes entirely
Extend the life of IP version 4
the Internet becomes just one giant
network that is “subnetted” into a
number of large blocks
64
CIDR




Some of these large blocks are then broken
down into smaller blocks
Which can in turn be broken down further.
This breaking down can occur multiple times.
Allows the “pie” of Internet addresses to be
split into slices of many different sizes, to suit
the needs of organizations.
65
CIDR

How to determine the network id and
host id:


Use mask – just like subnetting
Use slash notation
184.13.152.0/22
Network id – first 22 bits
Host id – last 10 bits
66
CIDR - example
CIDR (“Slash”) Notation and Its Subnet Mask Equivalent
67
Assigning IP Addresses


IANA/ICANN divides addresses into large blocks
Distributes to the four regional Internet registries
(RIRs):


APNIC, ARIN, LACNIC and RIPE NCC.
They further divide and distribute them to



lower-level national Internet registries (NIRs)
local Internet registries (LIRs)
Internet Service Providers (ISPs)
68
Assignment example
69
Private Network




IP Addresses used within a LAN that has no
direct connection to the Internet
IP address must be unique within its network
Special IP addresses are used for Private
networks
NAT is needed to translate the Private
Network IP address if connecting to the
Internet
70
Private Networks
Range Start
Address
Range End
Address
“Classful”
Address
Equivalent
Classless
Address
Equivalent
Description
10.0.0.0
10.255.255.255
Class A network
10.x.x.x
10/8
Class A private
address block.
172.31.255.255
16 contiguous
Class B networks
from 172.16.x.x
through
172.31.x.x
172.16/12
Class B private
address blocks.
192.168.255.255
256 contiguous
Class C networks
from 192.168.0.x
through
192.168.255.x
192.168/16
Class C private
address blocks.
172.16.0.0
192.168.0.0
71
Network Address Translation
(NAT)



Extend the life of IP version 4
Allows a small number of public IP
addresses to be shared by a large
number of hosts using private
addresses
Growing Concerns Over Security –
indirect access to servers etc
72
Network Address Translation
(NAT)


NAT allows an IP address from one
network to be translated into another
address on an internal network
You can also use NAT to translate a
single IP address valid on the Internet
into multiple internal addresses

Useful if your ISP gives you a single IP
address, yet you have multiple servers and
users on the internal network
73
Port-Based NAT



Ports - The port number in a message helps
identify individual connections between two
addresses (like an phone extension)
The combination of an address and port
uniquely identifies a connection
The inside address and port can be mapped
to an outside address and port by the NAT
device.
74
Port numbers


Port number assignments
Choose a port number usually >1024
75
76
Advantages of NAT






Public IP Address Sharing
Easier Expansion - easy to add new clients
Greater Local Control – Mostly private LAN
Greater Flexibility In ISP Service - Changing the
organization's Internet Service Provider (ISP) is
easier because only the public addresses change
Increased Security – Automatic firewall
(Mostly) Transparent - changes take place in one or
perhaps a few routers.
77
Disadvantages of NAT




Complexity
Problems Due to Lack of Public Addresses - Certain
functions won't work properly due to lack of a “real”
IP address in the client host machines.
Compatibility Problems With Certain Applications –
some applications like FTP require more work with
the mapping.
Problems With Security Protocols – tinkering with the
header…
78
Disadvantages of NAT


Poor Support for Client Access- “Peer-to-peer”
applications are harder to set up
Performance Reduction
 Each time a datagram transitions between the
private network and the Internet, an address
translation is required. In addition, other work
must be done as well, such as recalculating
header checksums. Each individual translation
takes little effort but when you add it up, you are
giving up some performance.
79
Summary





File servers and application servers are the
two basic types of servers
Various families of processors are available
from Intel, AMD, and others
There are many techniques for minimizing
loss due to hardware and software failures
Routers and switches allow servers to
communicate
IP addresses are divided into the network and
host portions
80