SS8pA1-A8.

qxp 7/27/05 7:44 AM Page A1

Networked Video for

HDK

Networked Video for

HBY
HBY1

W
HBY11

AT A GLANCE
•TX10

Brought to You by

Presented by

Part 3 of 4
AUGUST 2005— www.securitysales.com
AUGUST 2005 — www.securitysales.com A1
SS8pA1-A8.qxp 7/27/05 7:44 AM Page A2

Networked Video for

Part 3 of 4

Essentials of NETWORK
DESIGN & FUNCTION

By Bob Wimmer
Principal
Video Security Consultants
cctvbob@aol.com

AT A GLANCE
• Two networking standards are
802.x by the Institute of Electrical
and Electronic Engineers and the
Open Systems Interconnect seven-
layer model by the International
Organization for Standardization
• Seven-layer model defines how
everything from hardware to soft-
ware is supposed to work in a
network Determining the number, type and location of cameras and
• 802.x standards spell out how DVRs when designing CCTV systems is often only half the
networks themselves are de-
signed to function and intercom- battle.Today, security technicians also need to know what it
municate with other network takes to get video onto the network, not just to it. Find out what
processes and devices
• IP addresses identify two things:
concepts and terminology you need to learn to effectively
the device itself and the network interface with your clients’ IT managers.
to which that device belongs; sub-

netting allows for greater numbers
of IP addresses
• Media access control (MAC) ad-
dresses are assigned by manufac-
turers to every network interface
and establish connections between
two nodes
W
elcome to the third installment of Security Sales & Integration’s four-
part series — “Networked Video for D.U.M.I.E.S.” The first two parts of
this series dealt with getting video to the network. We talked about the
quality of the cameras themselves and the installation (see “Essentials of
Networked Cameras and Lenses” in the April issue). We also took a look at the different
compression methods and the process required to turn analog video into digital (see “How
to Beat the Bandwidth Blues” in the June issue).

A2 www.securitysales.com — AUGUST 2005

SS8pA1-A8.qxp 7/27/05 7:44 AM Page A3
Now we’re going to look at what
it takes to get it on the network.
While designing the analog portion of
our system is very important, all that
work will be for nothing if the
network isn’t designed and imple-
mented correctly. The first two parts
of this series were designed to give
network experts a peek into the world
of CCTV, while the last two segments
will give CCTV experts a peek into the
world of networking.
Make no mistake: Network design
and function is an exhaustive subject,
and far deeper than we can cover in a
couple of articles. However, what we
can do is gain a basic understanding
and learn some of the common ter-
minology to communicate more ef-
fectively with the network expert at
your next jobsite.
Networking World Ruled by
Standards and Protocols
Unlike the traditional CCTV industry,
the network universe is all based on pre-
determined agreements and sets of
rules called standards and protocols.
Without them, we wouldn’t have net-
works, the Internet, or much of the PC
and wireless technology we enjoy today.
There are two main sets of stan-
dards that govern networking today.
They are the 802.x standards, created
by the Institute of Electrical and Elec-
tronic Engineers (IEEE); and the Open
Systems Interconnect (OSI) seven-
layer model, created by the Interna-
tional Organization for Standardiza-
tion (IOS). We will look at each,
beginning with the OSI model.
7-Layer Model Defines Components
of Network Intercommunications
In the early days of networking (late
1960s-early ’70s), when computers
took up entire rooms, there wasn’t a
good, easy way to get information from
one to the next. At the time, those com-
puters were mostly proprietary hard-
ware/software combinations, so each
machine was essentially isolated from
the others. This dilemma was known
as islands of automation.
OSI Model Consists of 7 Layers
Application
Presentation
Session
Transport
Network
Data Link
Physical
The top and bottom layers of the Open System Interconnect’s (OSI) seven layers are fairly
easy to understand; however, things get a little fuzzy in the middle. The layer numbering is
done from the bottom up, Layer 1 being the Physical, etc. For the most part, the lower levels
are hardware-centric protocols, and the upper levels deal primarily with software.
In the late ’70s, a proposal was
made to the IOS describing a proto-
col stack of seven different layers,
each relating a different component
of intercommunication on a network.
From the network interface card and
the wires used to carry the data all
the way up to the applications, the
user created the data with each
layer dealing with a specific part of
the process.
This way, hardware and software
manufacturers could concentrate
their efforts only on the techno-
logies they were specifically respo-
nsible for and know that their prod
uct would work on just about
any machine.
The OSI model (see diagram above) is
a reference in almost every network doc-
ument used today and is still a central
part of any network education. As a
practical matter, however, the TCP/IP
protocol stack — which roughly maps
to the OSI model and that we’ll look at
later — is implemented more often in
hardware and software
While all the OSI layers together are
necessary to communicate, we will
focus primarily on the bottom four —
Physical-1, Data Link-2, Network-3,
Transport-4 — and the devices that
reside at each one.
Upper
Lower
Physical Layer Deals With Hardware
Like NICs, Hubs
At the very bottom of the OSI stack
resides the Physical layer, or Layer 1.
This layer is responsible for taking the
digital bits, changing them into elec-
tronic impulses and sending them
down the line.
You are probably very familiar with
the hardware that belongs to Layer 1.
If you were reading this article online,
the network interface card (NIC) that
connects your PC with the Internet is
a Layer-1 device.
Another very common hardware
device at Layer 1 is a hub. We’ll dis-
cuss hubs and other hardware devices
in detail in the next article.
Data Link Layer Ensures Network
Node Connections, Checks Info
The second level is called the Data
Link layer. Layer 2 provides an inter-
face, or link, between the higher lay-
ers and the network hardware. This
layer allows us to bind the higher
layer protocols (in some cases, more
than one) to our network cards,
for example.
The Data Link layer is primarily re-
sponsible for three things: It makes
sure a connection is available between
two network nodes, it encapsulates
AUGUST 2005— www.securitysales.com A3
SS8pA1-A8.qxp 7/27/05 7:44 AM Page A4
Networked Video for
the data into frames for transmission,
and it ensures that incoming data is
received correctly by performing some
error-checking routines.
Layer 2 is divided into two sublay-
ers, logical link control and media ac-
cess control. For our purposes, we will
concentrate on the media access con-
trol sublayer. This name should sound
familiar for those with any network
experience, but is better known by the
acronym MAC. Many of you have
probably heard of a MAC address.
A MAC address is a hard-coded ad-
dress assigned to every network in-
terface on any device made to attach
to a network. This address is assigned
by the manufacturer of the device,
and is ultimately the way a connec-
tion is established between two
nodes. The MAC address is unique
throughout the entire world; in fact,
the current MAC addressing scheme
should supply us with addresses
through the year 2100.
The address itself is a hexadecimal
48-bit address, broken into six octets
(eight-digit numbers, in binary). Being
48 bits long, it will give us 248 unique
addresses, or 281,474,976,710,656
possibilities.
It is important to understand the
structure of MAC addresses (see dia-
gram below). Connections between
How a Media Access Control (MAC) Address Is Structured
MAC address
00:0A:27:02:01:FC
Organizational unique identifier
(Mfg. ID)
The first part of the address, or the first three octets, is unique to the manufacturer of the
device. It is called the organizational unique identifier. This one happens to belong to Apple.
Every device with 00:0A:27 in the MAC is made by Apple. The second part of the MAC, or the
last 3 octets, is unique to each device.
A4 www.securitysales.com — AUGUST 2005
Comparing an IP Address to Computer’s Binary Interpretation
192.168.1.0
11000000.10101000.00000001.00000000
A proper IP address (top ) consists of four sets of numbers, separated by periods, or
dots. Underneath, the IP address is shown in a different way, in binary. This is the way
the computer actually sees it.
devices on a network are ultimately
made by MAC address, not IP ad-
dresses or domain names. Those
methods simply assist a device in
finding the MAC of another device.
The first part of a MAC address, or
the first three octets, is unique to the
manufacturer of the device. It is called
the organizational unique identifier.
Every company that makes network-
based devices has one, or several. The
second part of the MAC, or the last
three octets, is unique to each device.
No two devices in the world should
have the same MAC address. There
have been some cases where devices
with duplicate MACs have left facto-
ries, but they were caught and re-
called before they hit the streets.
As mentioned, the second sublayer
in Layer 2 is the logical link control
Device ID
(LLC). The LLC takes the raw data bits
from the upper layers and encapsu-
lates them in preparation for trans-
mission. It organizes the data into
frames, adding information such as
addressing, error checking, etc.
After framing and addressing is
complete, the frames are then sent to
Layer 1 to be converted into electrical
pulses and sent across the wires.
Switches are the most common
Layer-2 devices. Later, we will see that
they can also operate at higher layers,
with more intelligent processing
of packets.
Network Layer Handles Addressing
and Routing Tasks
The next level is known as the Net-
work layer. This is the layer that has
defined most of what we do on a net-
work or the Internet today.
The Network layer is primarily re-
sponsible for two things: addressing
and routing. Both of these functions
are related. This layer contains the IP
protocol, part of the TCP/IP protocol
stack. Once again, you probably have
heard the term “IP address” if you’ve
done anything with computers in the
past five to 10 years. This is the Net-
work layer’s responsibility.
An IP address is a 32-bit address
(remember, when we talk about bits
we are referring to binary counting)
that is purely logical. This means it
needs to be assigned by a user or ad-
ministrator somewhere, not set at the
factory. The fact that it is user-assign-
able means there is great flexibility in
addressing. But this also means care
SS8pA1-A8.qxp 7/27/05 7:44 AM Page A5

must be taken as an incorrectly as-

signed address will prevent you from Breaking Down 3 Classes of Network Addresses Into Binary
communicating on a network.
In order for a device to be used direct-
ly on the Internet, that address has to be Bit
Class A
78 31
unique throughout the entire world. #0

This is very important when it comes to Network Host information

information
allowing DVR access across the Internet. Class B
Bit
A great many providers, however, are #0 15 16 31
becoming more resistant to giving out
Network information Host information
unique addresses, for reasons we will
discuss below. Bit
#0
Class C 23 24
31
A proper IP address (see diagram
on page A4) consists of four sets of Network information Host
information
numbers, separated by periods, or
dots (as we in the biz like to say). That
is the way we are most familiar with The gray lines represent IP addresses in binary, from bit 0 to 32. Under Class A, the first
them. However, computers actually eight bits are titled network information. This identifies the network itself, and is
see the IP address in binary. If you are shared by all devices on that network segment. After the vertical divider line, the host
information part uniquely identifies each device.
going to get into networking with any
seriousness, you will need to be very
familiar with binary.
The current IP address format is How IP Version 4 Assigns IP Addresses
called IP version 4, or Ipv4. In the
Ipv4 scheme, there are 232, or Network class Beginning octet Number of Host addresses
4,294,967,296, possible addresses. networks per network
Unfortunately, the original method of A 1-126 126 16,777,214
allocating addresses was based on B 128-191 >16,000 65,534
classes and extremely inefficient. C 192-223 >2,000,000 254
Take a look at the diagram on the
upper right-hand side of this page. As
you look at each class, you will notice
that the dividing line moves. Class B ad- The current IP address format is called IP version 4, or Ipv4. Unfortunately, the original
method of allocating addresses was based on classes and extremely inefficient. For ex-
dresses are divided in the middle, with
ample, if a company needed 50 addresses, the smallest allocation it could get would
two octets for the network ID, and two be 254 as a Class-C network. Subnetting helps overcome these limitations.
for the device ID. Class C addresses have
the first three octets for the network,
and the last one for the device.
Moving that dividing line and chang- Dissecting an IP Address/Subnet Mask
ing classes determine how many differ-
ent networks we can create, and how Dotted decimal notation Binary notation
many hosts are on each. A host is any
device with an assigned address. IP address 166.144.40.33 10100110 10010000 00101000 00100001
If a company needed, let’s say, 50
addresses, the smallest allocation it Subnet mask: 255.255.255.0 11111111 11111111 11111111 00000000
could get would be 254 addresses. Re-
member, at the time, every address in
Indicates that this octet Indicates that this octet
the business had to be globally unique is part of an extended represents host information
if they wanted access to the Internet. network prefix
At the opposite end of the scale, the
big corporations that needed lots of Notice where all the No. 1s are in this comparison of an IP address and subnet mask in
addresses, got lots of addresses. binary. Anywhere you see a No. 1 indicates the network portion of the IP address. Any-
A single class-A network could where there is a zero shows the host portion.
give a company 16,777,214 globally

AUGUST 2005— www.securitysales.com A5

SS8pA1-A8.qxp 7/27/05 7:44 AM Page A6
Networked Video for
unique addresses. You can kind of get an
idea why people are saying that we are
now running out of addresses. The large
majority of addresses have already been
allocated. What the powers that be dis-
covered very quickly was they needed a
better way of assigning addresses, as
they were running low. This is where
subnett-ing comes in.
If you ask any network expert what
his least favorite part of networking is,
there’s a good chance he’ll tell you ad-
dressing or subnetting. That’s because
in spite of the fact that dealing with IP
addresses is fairly simple for us,
there’s a lot going on behind the
scenes to make those addresses work.
We said earlier that communication
paths between devices are actually es-
tablished through MAC addresses, but
because everyone in the world might
not want to be sharing a network with
everyone else in the world there had to
be a better way of dividing up resources.
With all the privacy issues we have,
imagine what it would be like if every-
one was on the same network!
The better way came about in the
form of IP addresses. Those sets of
numbers we are so familiar with actu-
ally do more than just identify a de-
vice. Like a MAC address, an IP ad-
dress identifies two things: the actual
device itself and the network to which
that device belongs.
Remember, the host portion of an
IP address is unique for each device
on a network, while the network por-
tion is the same on all devices that
share a network. The way we tell
which part of an address is which is
called the subnet mask. The subnet
mask is another 32-bit number that
looks similar to an IP address, but
does something entirely different.
Take a look at an IP address and
subnet mask in binary (see diagram
on page A5).
When you put the two together, no-
tice where all the No. 1s are in the sub-
net mask. Anywhere you see a No. 1
indicates the network portion of the
IP address. Anywhere there is a zero
shows the host portion. Just remem-
A6 www.securitysales.com — AUGUST 2005
IEEE 802 Project Standards
• 802.1 HLI Higher level interface (architecture)
• 802.1d Spanning tree brdge
• 802.1q Qos
• 802.2 LLC
• 802.3 CSMA/CD or ethernet
• 802.3c 10Mbps ethernet
• 802.3u 100 Mbps (fast ethernet)
• 802.3ab 1000 Mbps (gigabit ethernet)
• 802.3ae 10 gigabit ethernet
• 802.3af power over ethernet
• 802.4 Token passing bus
• 802.5 Token passing ring
• 802.6 MAN DQDB (distributed queue dual bus)
• 802.7 Broadband LANs (video, voice & data)
The Institute of Electrical and Electronic Engineers (IEEE) standard (or family of standards)
known as Project 802 was created in 1980 as a way to unite and standardize the many dif-
ferent networking technologies that were rapidly being developed at the time. As the charts
above and at right show, almost every aspect of networking is defined in these standards.
ber, if two addresses are not on the The other thing the Network layer,
same subnet, they won’t be able to specifically the IP protocol, handles is
talk to each other. Comparing the the routing of packets between different
sending address and subnet mask networks. Routers are the original Layer-
with the receiving address and subnet 3 hardware devices, but are being joined
mask, in binary, will show you if rapidly by more intelligent switches. We
there’s a difference. will discuss hardware functions and fea-
So is there any relief from all this tures in the next article.
nonsense? Actually, there is. Some
relief will come in the form of IPv6. IP Transport Layer Gets Data Packets
version 6 is 128 bits long, compared From Point A to Point B
to v4’s 32, giving us 3.4 X 1,038 possi- The last OSI level we’ll talk about is
ble addresses, a few more than ver- the Transport layer, or Layer 4. This
sion 4 to be sure. A version 6 IP ad- layer is responsible for reliably getting
dress is also hexadecimal, and looks the packets from point A to point B.
like this: fe80:0000:0000:0000:020a: This layer supports two different
95ff:fef4:a3b1. transmission methods: connection-
IPv6 implementation is rather oriented and connectionless.
spotty at the moment, and it’s kind of Connection-oriented transmis-
a catch 22. Providers don’t want to sions are handled by Transmission
spend the money to implement it as Control Protocol ( TCP). These are
there aren’t too many users yet, and point-to-point connections for guar-
there aren’t too many users because anteed reception of data. An E-mail
there aren’t too many providers that message, accessing a Web page or
support it. It is the wave of the future, downloading a file are all examples of
however, and it remains to be seen connection-based exchanges.
how soon we’ll see it in DVRs and Because there is a guarantee of data
other industry products. reception, error checking is performed.
SS8pA1-A8.qxp 7/27/05 7:44 AM Page A7
IEEE Standards
• 802.8 Fiber optic TAG
• 802.9 ISLANs Integrated digital & video networking
• 802.10 LAN/MAN Security
• 802.11 Wireless MAC & physical layers
• 802.11a 5 Ghz 52 MBs
• 802.11b 2.5 Ghz 11 Mbs Wi-Fi
• 802.11g Dual mode 802.11a & 802.11b compatible
• 802.11i Wireless security
• 802.12 100 BASE VG AnyLAN
• 802.13 Nothing superstition
• 802.14 Cable TV-based broadband
• 802.15 Working group WPANs
• 802.16 Wireless MAN Air Interface
• 802.16a Broadband wireless WIMAX
• 802.17 Resilient packet ring working group
If packets are received incorrectly or
damaged, they will be present. You can
probably imagine this wouldn’t work
too well for video. It would be a little
annoying if a sending device kept re-
transmitting parts of a video clip; it
would be unwatchable.
For video streaming, we use a con-
nectionless protocol, user datagram
protocol (UDP), which does not guaran-
tee delivery of data. In other words, if
you get it, you get it … if not, oh well.
This is the way large amounts of video
will be moved in future network-based
systems that require access from multi-
ple receivers or workstations.
UDP is also the foundation of video
multicasting, which is a one-to-many
method of streaming. It is also a cru-
cial element of networked video sys-
tems. We’ll discuss multicasting in de-
tail in the next article.
For an interesting and easy to under-
stand analogy for the OSI model, check
out www.lewistech.com/rlewis/re-
sources/jamesbondOSI2.aspx.
Project 802 Standards Spell Out
Nearly All Aspects of Networking
The IEEE standard (or family of
standards) known as Project 802 was
created in 1980 as a way to unite and
standardize the many different net-
working technologies that were rap-
idly being developed at the time.
The base standards in the group
were drawn from Ethernet technolo-
gy, which had been around since the
early ‘70s. Specifically, 802.3 was set
up to describe Ethernet itself, and
even more specifically the method
Ethernet used to avoid data colli-
sions: carrier-sense, multiple ac-
cess/collision detection (CSMA/CD).
With CSMA/CD, when an Ethernet
device wants to transmit, it senses the
line. If it is clear to send, out goes the
data. If a collision is detected, the
transmitter waits a pre-determined
period of time before it tries again.
While this is a slight oversimplifica-
tion, it’s important to understand that
the main thing that separates the dif-
ferent types of networks (Ethernet
and Token Ring, etc) is how they han-
dle collisions.
As the 802 chart shows (above), al-
most every aspect of networking is
defined in these standards. They even
had the foresight to make room for
technologies that were, at the time,
just sparkles in their inventors eyes.
While not many people refer to Eth-
ernet as 802.3, we are familiar with
one of the higher levels, 802.11, com-
monly used to refer to wireless net-
working (see, you knew these stan-
dards without even realizing it!).
Aside from Ethernet and wireless,
Project 802 defines several other net-
working technologies. Token Ring and
Token Bus are two competing tech-
nologies to Ethernet.
Token Ring was a technology pro-
moted and developed by IBM in the
‘80s. While Ethernet used the line-
sensing collision detection scheme
we talked about earlier, Token Ring
used a small piece of data called a
token, and only the node in posses-
sion of the token could transmit. It
never really caught on for widespread
use and is not found too much any-
more. While it would be possible to
do video over Token Ring, perform-
ance limitations would not make it a
real practical application.
You May Exhale Now:The Worst Is
Over! Next Up: Implementation
Protocols and standards are not
very interesting to read about. How-
ever, it is important to understand the
impact they have on networking
today. Without this basic foundation,
we would not have many of the tech-
nological advances we enjoy.
These standards go beyond just hook-
ing computers together, though. Many
of the cell phone systems today use
some or many of the networking tech-
nologies defined by these groups. The
impact those early groups of people had
on our lives today is immeasurable.
In the final part of this series in the
September issue, we will look at how
some of these technologies are imple-
mented and the hardware that holds
it all together. We’ll also look at the
how’s and why’s of getting video onto
a network successfully. ■
Robert (Bob) Wimmer is president of Video Security
Consultants and has more than 33 years of experi-
ence in CCTV. His consulting firm is noted for tech-
nical training, system design, technical support and
overall system troubleshooting.
AUGUST 2005— www.securitysales.com A7
SS8pA1-A8.qxp 7/27/05 7:44 AM Page A8

Reader Service Card No. 156