Abekas A60 Parts list manual
A
Carllnn
C;"mflany
digital
disk
recorder
A60
A60
Ethernet
Manual
PRELIMINARY
A60
Ethernet
Manual
Rev
1.2
7-DEC-87
Copyright
(C)
1987
Abekas
Video
Systems,
Inc.
This
manual
describes
the
installation
and
use
of
the
A60
as
an
Ethernet
node
supporting
file
transfer
and
remote
login
using
some
of
the
TCP/IP
family
of
protocols.
It
also
includes
an
application
note
describing
some
of
the
mechanisms
behind
the
file
transfers.
Abekas
Video
Systems,
Inc.
101
Galveston
Drive
Redwood
City,
CA
94063
(415)
369-5111
uucp
Email:
••.
!pyramid!abekas!a60mail
1.
2.
3.
4.
CONTENTS
Introduction
to
Ethernet
1.1
Ethernet
••.•
and
TCP/IP.
1.2
TCP/IP
.•••••.
1.3
Telnet
and
FTP.
1.4
Unix
'r'
Commands.
1.5
The
Abekas
A60
..•.
TCP/IP
Application
Notes
••••.
2.1
Typical
File
Transfer
••
2.2
Layered
Model
•••.
2.3
Physical
layer.
2.4
2.5
2.6
2.7
Transceivers
••..
Different
Ethernet
Standards.
SQE
•••.•...•••..•••
Data
Link
Layer
••••..
Ethernet
Addresses.
Network
Layer
••..••••
IP
••••••••..••.•••..
Internet
Addresses
•.
ICMP.
GGP.
ARP
••
RARP
••••••••
Transport
Layer
..
TCP
•••••••••.
UDP
••••••••••
Upper
Levels
.•
Telnet
••
FTP
••••
..........
TFTP
••
rep
••
.............
Installing
an
A60
on
a
Unix
Network
•.
/
etc/hosts
.•.....•..•.•..••....
Setting
the
A60
Internet
Address.
Setting
the
A60
Hostname
..
/etc/ethers
....•....•
Implementation
Notes.
Ethernet
Address
.•.
Address
Resolution
.•.
IP
••.•••.•......•.
i
2
2
3
4
5
5
6
6
7
8
8
9
10
11
11
11
11
12
13
13
13
14
14
14
16
16
17
17
18
18
20
20
22
22
22
24
24
24
25
10.
11.
Appendix
••••••••••.•••••••••••
10.1
10.2
10.3
10.4
10.5
10.6
Complete
TCP/IP
Packet
••••••••
FTP
Implementation.
Defaults
••••••••
Opening
Message
.•..
Commands
and
Responses
••
File
Names
•••••••.••••••••••..
rlogin
Implementation
••
rsh
Implementation.
rcp
Implementation.
Bibliography
••••••..•.•••
iii
59
59
60
60
60
60
62
62
62
63
64
A60
Ethernet
Manual
1.
Introduction
to
Ethernet
and
TCP/IP
1.1
Ethernet
Ethernet
is
a
Local
Area
Network
(LAN)
Standard
originally
developed
at
Xerox
Palo
Alto
Research
Center.
2
Ethernet
interconnects
a
group
of
computers
(referred
to
as
hosts
or
nodes)
with
a
single
50-ohm
coaxial
cable
with
terminations
.at
both
ends.
Data
is
passed
serially
at
10MHz
in
the
form
of
packets,
that
is
in
chunks
anywhere
from
46
up
to
1500
bytes
or
characters.
Each
packet
carries
addressing
information
to
show
its'
source
and
destination.
Unlike
the
Public
switched
telephone
system
or
a
video
routing
matrix
the
single
cable
is
shared
by
all
the
devices
on
the
network
so
there
are
a
set
of
rules
to
determine
when
a
node
can
access
the
cable.
The
technique
used
is
referred
to
as
Carrier
Sense
Multiple
Access
with
Collision
Detection
(CSMA-CD).
Before
transmitting
a
node
listens
to
confirm
that
nobody
else
is
transmitting,
then,
as
it
transmits
it
continues
to
monitor
the
cable
in
case
another
node
started
transmitting
at
the
same·time.
If
two
devices
transmit
simultaneously
it
is
referred
to
as
a
collision
and
both
devices
have
to
stop
immediately
and
wait
a
random
amount
of
time
before
attempting
to
transmit
again.
Different
manufacturers
have
adopted
the
low-level
Ethernet
hardware
and
packet
specifications
and
built
their
own
networks
on
top
of
it.
Xerox
XNS, IBM-SNA, HP-NS
and
DEC-DECnet
are
all
networking
systems
that
allow
users
to
share
resources
and
files
and
can
run
over
Ethernet.
In
the
area
of
Personal
Computers
3com
Corporation
and
Novell
are
supplying
File
server
systems
based
on
Ethernet.
Small
scale
Office
LAN's
are
mostly
based
on
cheapernet
which
uses
thin
RG58
50
Ohm
cable
and
BNC
connectors,
i~
this
case
the
coaxial
cable
is
'T'eed
directly
onto
the
Ethernet
Interface
in
the
computer.
Higher
level
applications
use
better
quality
thick
yellow
cable
and
external
transceivers
that
can
attach
to
the
cable
with
a
spike-like
tap.
PRELIMINARY
Ethernet
A60
Ethernet
Manual
3
1.2
TCP/IP
The
TCP/IP
protocol
family
is
emerging
as
a
useful
common
standard
for
network
interconnection.
The
strength
of
TCP/IP
has
been
that
it
is
not
tied
to
any
particular
manufacturer,
it
is
the
result
of
extensive
research
since
the
70's
by
the
Advanced
Research
Projects
Agency
(ARPA)
community.
with
backing
from
the
DOD
the
emphasis
for
these
protocols
has
been
to
interconnect
different
types
of
computers
running
different
operating
systems.
TCP/IP
is
now
available
as
an
add
on
to
most
computer
systems
either
in
the
form
of
and
interface
board
with
built
in
software
such
as
the
Excelan
Ethernet
Controllers
or
as
an
extra
software
package
running
along
side
a
native
Ethernet
implementation.
(The
Biblography
for
this
manual
lists
some
of
the
companies
offering
TCP/IP
packages)
One
reason
for
the
spread
of
TCP/IP
amongst
the
Computer
Graphics
Community
has
been
its
inclusion
in
the
Berkeley
Versions
of
the
UNIX
operating
system
(referred
to
as
4.2
BSD
UNIX
as
opposed
to
[he
AT&T
Unix
V),
most
graphics
engines
and
the
Workstations
that
control
them
use
Unix
as
it
is
a
popular
operating
system
for
software
development.
TCP
and
IP
are
acronyms
for
'Transmission
Control
Protocol'
and
'Inter-network
Protocol'
just
two
of
the
layers
in
the
suite
of
communications
protocols
that
are
required
to
allow
transfer
of
data
from
one
computer
to
another.
IP
is
the
layer
immediately
on
top
of
Ethernet
that
adds
Network
addressing
information
to
the
packet.
These
Internet
addresses
allow
IP
packets
to
be
transferred
to
other
networks
not
just
Ethernet,
it
is
similar
to
the
way
that
Containerized
freight
can
be
carried
equally
well
by
road
rail
or
sea.
TCP
provides
an
error
free
bidirectional
communications
channel
above
which
other
utilities
such
as
a
remote
login
or
file
transfer
can
be
built.
TCP
works
by
giving
each
packet
a
sequence
number
so
that
a
message
or
file
can
be
reassembled
even
if
the
packets
arrive
in
the
wrong
order.
PRELIMINARY
TCP/IP
A60
Ethernet
Manual
4
TCP
also
has
an
acknowledgement
mechanism
whereby
the
receiver
replys
with
the
latest
complete
sequence
number
it
has
assembled,
so
that
should
a
packet
get
lost
or
delayed
in
the
network
the
sender
will
retransmit
the
missing
packet
if
it
hasn't
been
acknowledged
within
a
reasonable
timeout
period.
The
third
TCP
mechanism
is
the
window
which
limits
the
amount
of
unacknowledged
data
the
sender
can
send
out,
so
that
it
can't
get
too
far
ahead
if
the
receiver
is
missing
a
packet
from
back
at
the
start
of
the
message.
For
the
majority
of
File
transfers
or
TCP
connections
there
is
no
data
lost,
all
the
packets
arrive
in
the
correct
order,
the
power
of
the
TCP/IP
protocols
lies
in
the
fact
that
they
are
not
restricted
to
running
on
a
single
local
area
network.
The
ARPA
Internet
for
example
combines
over
a
100
different
networks
and and
includes
satellite
links
across
to
research
facilities
in
Europe.
When
packets
are
passing
across
several
networks
through
'gateways'
which
provide
an
interface
from
one
type
of
network
to
~nother
there
is
more
chance
of
a
packet
getting
lost.
There
is
no
~uarantee
that
all
that
packets
will
take
the
same
route
to
the
destination,
this
is
possible
since
each
packet
carries
separate
addressing
information.
It
is
up
to
the
gateways
to
decide
what
the
most
efficient
route
is
and
if
during
the
life
of
the
connection
one
of
the
intervening
gateways
or
networks
goes
down
the
TCP/IP
protocol
is
robust
enough
to
be
able
to
replace
any
unacknowledged
lost
data
by
retransmission.
That
is
assuming
an
alternative
route
can
be
found.
1.3
Telnet
and
FTP
On
top
of
the
guaranteed
delivery
TCP
connections
the
A60
supports
file
transfer
and
remote
control.
Remote
control
is
achieved
by
allowing
the
remote
user
to
'login'
as
if
the
A60
were
another
computer
and
type
commands
interactively.
There
are
two
ways
of
doing
each,
firstly
the
official
ARPA
file
transfer
and
remote
login
utilities
called
FTP
(File
Transfer
Protocol)
and
Telnet
which
are
specifically
intended
to
work
between
different
Computer
architectures
and
Operating
Systems.
pRELIMINARY
Telnet
and
FTP
A60
Ethernet
Manual
5
1.4
Unix
'r'
Commands
The
other
alternative
is
the
native
Unix
utilities
'rcp'
(Remote
CoPy)
'rlogin'
(Remote
Login)
and
'rsh'
(remote
shell)
these
will
be
popular
with
Unix
users
since
they
offer
a
less
verbose
user
interface
-
file
transfers
are
achieved
by
cryptic
one
line
commands
rather
than
FTP
which
normally
produces
a
secondary
prompt
and
requires
at
least
three
commands
to
transfer
one
file.
1.5
The
Abekas
A60
The
A60
can
be
viewed
as
a
Video
Server
-
permitting
all
the
rendering
engines
and
computers
in
a
graphics
lab
to
share
the
ability
to
tryout
animation
sequences
and
layoff
rendered
images
without
the
preroll
and
lineup
problems
associated
with
single
frame
VTRs.
It
is
a
powerful
sharable
resource
that
treats
all
the
frames
or
fields
on
a
disk
as
separate
files
that
can
be
copied
to
or
from
the
A60.
The
remote
control
provided
across
Ethernet
is
intended
to
be
human
readable
in
that
the
commands
for
playing
or
setting
up
segments
are
executed
by
merely
typing
"PLAY"
or
"DEFSEG
0.20
1.30"
so
a
user
sitting
at
a
workstation
can
easily
preview
frames
or
animations
without
the
need
for
a
remote
control
panels
beside
every
workstation.
As
an
Ethernet
device
the
A60
appears
to
be
just
another
node
that
files
can
be
transferred
to
in
the
same way
they
would
be
moved
from
one
computer
system
to
another,
there
is
no
need
for
a
separate
VTR
controller
or
special
software.
Installation
on
Ethernet
is
just
a
question
of
the
plugging
the
A60
into
a
Transceiver
which
is
a
small
box
that
provides
the
interface
to
the
Ethernet
Coax.
The
System
Manager
then
specifies
an
Internet
address
for
the
machine
which
has
to
be
entered
on
the
A60
control
panel.
The
other
computers
on
the
network
can
then
use
this
Internet
address
to
access
the
A60.
PRELIMINARY
The
Abekas
A60
A60
Ethernet
Manual
6
2.
TCP/IP
Application
Notes
This
section
gives
a
brief
outline
of
the
functions
of
each
of
the
protocols
used
in
the
A60
and
the
way
they
interact.
2.1
Typical
File
Transfer
Take
for
example
the
opening
of
an
FTP
connection
on
a
Unix
system.
To
invoke
the
FTP
program
the
user
types
:
unix%
ftp
a60
Which
causes
a
control
connection
to
be
opened
to
the
host
specified.
The
hostname
is
the
name
used
to
refer
to
the
A60
on
the
users'
machine.
Somewhere
there
will
be
a
file
(/etc/hosts
on
unix)
which
gives
the
mapping
between
the
name
(or
some
other
optional
alias)
and
the
Internet
Address
that
has
been
assigned
for
the
A60.
A
typical
entry
in
/etc/hosts
has
the
following
form
:
192.5.200.9
a60
The
FTP
program
will
first
find
the
Internet
address
of
the
remote
host
by
referring
to
the
/etc/hosts
file.
Then
it
has
to
find
an
Ethernet
address
that
corresponds
to
this
Internet
address.
This
Ethernet
address
is
the
address
of
a
device
on
the
local
ethernet,
either
the
A60
itself
or
a
Gateway
through
which
the
A60
can
be
reached.
In
some
cases
the
host
computer
may
have
retained
this
information
from
a
previous
transaction
but
for
the
first
transfer
to
an
unknown
remote
host
the
local
host
has
to
resolve
the
Internet
-
Ethernet
address
mapping.
To
do
this
it
uses
the
Address
Resolution
Protocol
(ARP).
ARP
involves
sending
a
broadcast
packet
to
all
the
hosts
on
the
network.
only
the
host
with
the
required
Internet
address
or
a
Gateway
that
can
reach
it
will
reply
with
a
ARP
reply
packet
supplying
the
requested
Ethernet
address.
)RELIMINARY
Typical
File
Transfer
A60
Ethernet
Manual
7
Once
the
local
host
knows
how
to
reach
the
A60
over
Ethernet
it
can
open
up
a TCP
connection
to
the
FTP
port
on
the
A60.
The
A60
which
has
been
waiting
for
a
connection
to
be
made
to
the
FTP
port
responds
with
an
opening
message
and
the
FTP
program
passes
the
Users
name
to
the
A60.
Since
the
A60
does
not
do
any
username
checking
the
user
then
specifies
the
filename
to
transfer
A
data
connection
is
opened
from
the
A60
end
to
the
port
chosen
by
the
FTP
program
on
the
users'
machine
and
the
data
is
transferred.
The
data
connection
is
closed
once
the
data
has
been
transferred,
the
control
connection
is
closed
when
the
user
issues
the
'quit'
command.
2.2
Layered
Model
The
ISO
standards
authority
has
proposed
an
Open
Systems
Interconnection
model
consisting
of
seven
layers.
The
reason
for
dividing
the
Protocols
into
layers
is
so
that
different
protocols
at
the
same
level
can
be
used
interchangeably
to
provide
the
same
function
for
the
layers
above.
RELIMINARY
f
~-1
Presentation
.\
Application
Session
I
1
Transport
i
t
Network
Data
Link
Physical
OSI
diagram
Layered
Model
A60
Ethernet
Manual
8
The
TCP/IP
family
of
protocols
can
be
applied
to
several
low
level
networks
architectures,
in
this
case
the
two
lower
levels
-
that
is
the
Physical
layer
which
defines
the
connectors
and
voltages,
and
the
Data
link
layer
which
defines
the
way
data
is
passed
between
two
pieces
of
equipment
on
the
same
local
network
are
both
defined
by
the
IEEE
802.3
standard
(which
in
turn
is
based
on
the
Xerox
Ethernet
V2).
Another
comparable
Network
standards
is
X.25
which
is
used
for
public
packet
switched
services.
Family
Tree
2.3
Physical
layer
Ethernet
was
developed
at
Xerox
PARC
and
is
based
on
the
concept
of
a
baseband
Carrier
Sense
Multiple
Access
with
Collision
Detection.
(CSMA-CD)
Transceivers
The
host
computer
will
be
coupled
onto
the
Ethernet
Coax
by
a
device
called
a
transceiver.
In
some
instances
(particularly
in
the
case
of
'Cheapernet')
the
transceiver
can
be
part
of
the
Ethernet
interface
board
in
the
host
computer.
The
Remote
pRELIMINARY
Transceivers
A60'Ethernet
Manual
9
transceivers
can
be
up
to
50m
away
from
the
computer
and
typically
come
with
a
plug
in
module
to
allow
either
a
spike
tap,
N-series
or
BNC
connectors
to
interface
to
the
cable.
Cheapernet
uses
thin
RG58
50-ohm
coax
and
BNC
style
connectors
rather
than
the
high
grade
10mm
thick
Yellow
coax
that
permits
the
non-intrustive
"vampire
tap"
transceivers
to
be
spiked
into
it
and
removed
without
having
to
take
the
network
down.
The
thick
coax
can
be
used
for
networks
up
100
nodes
on
500m
of
cable
whereas
Cheapernet
is
limited
to
30
nodes
on
185
m
of
cable.
Only
two
repeaters
are
allowed
on
a
local
network
because
of
the
propagation
delays
through
them.
The
transceivers
should
be
placed
at
multiples
of
2.5metres
on
the
cable.
The
following
is
an
non-exhaustive
list
of
Transceivers
currently
available.
Inmac
Transceivers:
part
number
8043
HP
part
number
30241A
LAN
802.3
MAU
(Medium
Attachment
Unit)
3com
Ethernet/IEEE
802.3
Transceivers:
3C107
Transceiver
with
AMP
piercing
tap
3C108
Transceiver
with
N
series
connectors
3C109
Tranceiver
with
BNC
connectors
DEC
Ethernet
bits:
DESTA
-
Thin
ethernet
transceiver
DELNI -
Ethernet
in
a
box
The
DELNI
has
eight
0-15
connectors
for
connecting
up
to
eight
ethernet
devices
without
using
transceivers
or
coaxial
cable.
It
also
has
a
ninth
connector
for
connecting
to
a
regular
transceiver
(and
thus
to
a
larger
net),
or
to
another
DELNI
(for
a
hierarchical
DELNI
network).
Different
Ethernet
Standards
There
are
three
different
Ethernet
standards
:
Ethernet
V1,
V2
and
the
Newer
IEEE
802.3
standard.
The
are
electrical
differences
between
the
different
standards
mainly
concerning
the
grounding
requirements
for
the
transceiver.
For
Ethernet
V1
and
V2
the
connector
shell,
cable
shield
and
pin
PRELIMINARY
Different
Ethernet
Standards
A60
Eth~rnet
Manual
10
one
of
the
connector
are
all
connected
together
whereas
for
IEEE
802.3
the
cable
shield
is
separate
from
pin
1
ground.
The
grounding
should
be
achieved
at
the
host
end
of
the
transceiver
cable.
outer
Shield
1 9
Collision
-
Collision
+ 2 • •
10
Transmit
-
Transmit
+ 3 • •
•
11
Ground
*Inner
Shield
•
4 • •
12
Receive
-
Receive
+ 5 • •
13
Power
+12v
Ground
6 • •
• •
14
Ground
*Vcc
7 •
15
No
Connect
Logic
8
Transceiver
(AUI)
cable
pinout
*
Notes
(Pin
4
Electrically
Isolated
from
Outer
shield
for
802.3)
(Pin
7
appears
to
be
a
no
connect
on
most
machines)
Male
connector
at
transceiver,
Female
at
main
chassis.
SQE
Some
transceivers
support
SQE
or
Signal
Quality
Error
(sometimes
called
heartbeat)
in
which
case
they
will
simUlate
a
collision
in
the
gap
at
the
end
of
every
transmission
from
the
host.
This
provides
a
way
of
assuring
that
the
collision
detection
circuitry
in
the
host
interface
is
functioning
correctly.
Ethernet
VI
does
not
provide
for
SQE.
IRELIMINARY
SQE
A60
Ethernet
Manual
11
2.4
Data
Link
Layer
The
Data
Link
layer
provides
for
source
and
destination
addresses
and
a
trailing
CRC
allows
the
integrity
of
the
whole
packet
to
be
checked.
Ethernet
and
IEEE
specifications
differ
in
a
few
minor
respects,
The
IEEE
spec
states
that
the
field
immediately
following
the
source
address
is
an
optional
length,
however
the
Ethernet
receiving
hardware
is
capable
of
determining
the
packet
length
by
other
means.
The
original
Xerox
Ethernet
spec
defines
this
as
a
type
field
which
is
used
to
resolve
the
next
protocol
above
in
the
hierarchy.
The
types
used
for
the
ARPA
family
of
protocols
are
intentionally
chosen
to
be
illegal
lengths
Ceg ,
larger
than
the
maximum
permissible)
so
as
to
remove
any
confusion
between
the
IEEE
and
Xerox
implementation
of
Ethernet.
Ethernet
Addresses
Ethernet
addresses
are
six
bytes
conventionally
written
as
six
hex
numbers
separated
by
colons.
Ethernet
addresses
are
intended
to
be
unique
for
any
piece
of
equipment
that
conforms
to
the
IEEE
802.3
standard.
For
instance
Abekas
Equipment
will
start
00:00:76:XX:XX:XX
Using
the
Ethernet
Broadcast
packet
mechanism
it
is
possible
for
the
local
host
to
send
a
packet
to
all
the
devices
on
the
local
network
rather
than
to
one
specific
address.
This
is
address
FF:FF:FF:FF:FF:FF.
2.5
Network
Layer
The
Network
Layer
provides
a
packet
delivery
system
between
two
hosts.
IP
The
Internet
Protocol
adds
to
a
packet
the
information
required
to
pass
it
from
one
host
to
another
across
different
networks.
The
intention
is
that
a
gateway
that
is
passing
the
packet
from
one
network
to
another
need
only
examine
the
IP
header
to
determine
the
routing
information.
~RELIMINARY
IP
A60
Ethernet
Manual
'
12
The
Internet
Protocol
also
provides
a
mechanism
for
breaking
packets
into
smaller
fragments
for
passing
them
over
a
network
with
restricted
packet
size.
Internet
Addresses
Internet
address
is
a
four
byte
number
conventionally
expressed
as
four
decimal
numbers
(0
••
255)
separated
by
dots.
eg
192.9.200.5
The
Internet
Address
can
be
subdivided
into
two
parts,
the
first
is
referred
to
as
the
Network
number
and
the
second
as
the
local
address.
All
the
machines
on
the
same
local
network
should
have
the
same
Network
Number
but
different
Local
Addresses.
If
a
host
is
asked
to
send
a
packet
to
a
remote
host
with
a
different
Network
number
it
will
assume
that
the
remote
host
is
on
a
different
network
and
attempt
to
find
a
gateway
to
it.
Network
Numbers
for
Internet
sites
are
assigned
by
the
ARPA
authorities
the
local
addresses
are
chosen
by
the
local
system
administrator.
There
are
many
Ethernet
networks
that
are
not
actually
part
of
the
Internet
and
only
use
a
default
Network
number.
There
are
three
classes
of
internet
address
based
on
the
the
Network
number.
For
a
class
A
address
the
first
byte
of
the
internet
address
will
be
less
than
128,
the
local
address
is
then
formed
by
the
lower
three
bytes
allowing
for
4096
separate
hosts
on
the
one
network.
A
class
B
address
will
have
a
first
byte
in
the
range
128
to
191
and
in
this
case
the
lower
two
bytes
form
the
local
address.
Class
C
addresses
the
first
byte
will
have
a
value
greater
than
191
only
the
last
byte
identifies
the
individual
machine
or
internet
node.
Internet
addresses
are
intended
to
be
more
wide
ranging
than
the
native
addressing
scheme
for
the
the
local
network,
it
is
possible
to
connect
to
an
Internet
host
across
several
different
types
of
network
eg
Local
Area
Ethernet,
long
haul
X.25
packet
switched
services
and
Local
area
Token
ring
networks.
fRELIMINARY
Internet
Addresses
A60
Ethernet
Manual
13
ICMP
Internet
Control
Message
Protocol
Is
almost
an
integral
part
of
IP,
it
is
intended
to
handle
error
reporting
from
Gateways
to
hosts
or
hosts
to
hosts.
It
gives
the
originating
host
more
information
about
the
reason
a
packet
has
been
rejected
or
cannot
be
delivered.
ICMP
also
provides
an
echo
system
used
by
'ping'
programs
to
help
isolate
inter-network
problems.
A
Selection
of
the
ICMP
Massages
:
GGP
Echo
Request
Echo
reply
Redirect
:
Use
alternative
route
Time
Exceeded
:
Packet
died
of
old
age
Parameter
problem
:
Something
wrong
with
the
IP
Header
Destination
Unreachable,
either:
Network
Host
Protocol
Port
Gateway
to
Gateway
Protocol
Handles
communication
between
Gateways
for
control
purposes.
Allows
them
to
exchange
routing
information
and
keep
up
to
date
on
the
availability
of
neighboring
Gateways.
ARP
The
Address
Resolution
Protocol
uses
the
Ethernet
Broadcast
Mechanism
to
allow
a
host
to
resolve
Internet
to
Ethernet
address
mappings
by
asking
all
the
hosts
on
the
local
network
if
any
of
them
claim
to
be
the
required
Internet
address.
If
the
hosts
support
ARP
they
will
decode
the
packet,
it
contains
both
the
Internet
and
Ethernet
Addresses
of
the
local
host
and
the
Internet
address
of
the
host
it
is
trying
to
reach.
Only
the
requested
host
or
a
Gateway
that
can
reach
it
on
another
network
reply
directly
to
the
requesting
machine.
.PRELIMINARY
ARP
A60
Ethernet
Manual
14
RARP
The
Reverse
Address
Resolution
Protocol
allows
an
Ethernet
host
to
determine
its
own
Internet
Address
by
broadcasting
a
request
to
the
net.
2.6
Transport
Layer
The
fourth
OSI
layer
is
the
Transport
Layer,
which
is
concered
with
creating
and
maintaining
logical
connections
between
individual
processes
on
different
hosts.
TCP
The
transport
layer
adds
a
source
and
destination
port
number
to
the
packet
addresses.
This
allows
the
packet
traffic
to
be
routed
to
several
different
processes
or
users
within
a
particular
machine.
A
TCP
connection
is
characterized
by
the
combination
of
local
and
remote
Internet
addresses
and
the
local
and
remote
port
numbers.
So
for
instance
several
people
may
be
logged
in
on
a
machine
though
the
port
assigned
to
the
rlogin
service.
If
two
of
~he
users
originate
from
the
same
machine
they
will
have
to
be
on
different
ports
on
their
local
machine.
The
TCP
software
can
then
uniquely
identify
which
connection
a
packet
belongs
to.
Some
of
the
lower
port
numbers
(normally
less
than
1024)
are
reserved
for
system
functions,
these
are
'well
known'
port
numbers
that
are
published
for
other
computers
wishing
to
use
a
particular
service.
Opening
connections
to
these
ports
is
normally
a
privileged
operating
system
function.
Commonly
used
TCP
port
numbers
FTP 21
Telnet
23
Rsh
514
Rlogin
513
The
TCP
packet
header
also
carries
Sequence
and
Acknowledge
numbers.
The
sequence
number
represents
the
position
of
the
first
rRE
LIMI
NARY
TCP
A60
Ethernet
Manual
15
byte
of
this
packet
in
the
transmitted
data
stream.
The
Acknowledge
represents
the
byte
after
the
last
fully
reassembled
byte
of
the
data
stream
received.
In
other
words
the
Receiving
TCP
can
buffer
several
packets
that
may
be
out
of
sequence
but
the
Acknowledge
number
will
only
increase
once
the
incoming
data
is
complete
up
to
that
byte
number.
Sequence
numbers
are
fixed
when
the
connection
is
opened,
packets
are
exchanged
with
the
SYN
flag
set
to
indicate
the
initial
sequence
number.
The
TCP
header
also
includes
a window
to
indicate
to
the
other
end
how much
buffer
space
is
currently
available.
This
acts
as
a
method
of
flow
control
since
the
transmitter
should
not
continue
if
more
than
a
windows'
worth
of
data
remains
unacknowledged.
The
transmitting
side
of
TCP
is
responsible
for
resending
a
data
if
it
is
not
acknowledged
within
a
timout
period.
There
is
a
TCP
option
that
allows
the
maximum
number
of
data
bytes
in
a
packet
to
be
specified,
the
default
is
512.
TCP
connections
can
be
opened
either
Actively
or
Passively.
In
the
~ctive
case
the
remote
port
and
host
address
are
fully
specified,
whereas
a
passive
open
will
only
specify
the
port
number
at
the
local
end
and
wait
for
an
incoming
attempt
to
connect.
On
closing
the
connection
it
is
necessary
for
both
sides
to
exchange
and
acknowledge
packets
with
the
FIN
flag
set.
pRELIMINARY
TCP
A60
Ethernet
Manual
16
1
Typical
Data
Packet
UDP
The
User
Datagram
Protocol
is
a
simple
extension
to
IP
that
adds
only
a
source
and
destination
port
number
and
a
checksum.
A
Datagram
is
a
standalone
packet
with
neither
guaranteed
delivery
nor
special
sequence.
TFTP
and
Sun
Microsystems'
RPC
and
NFS
are
among
the
protocols
built
on
top
of
UDP.
2.7
Upper
Levels
For
the
TCP/IP
protocols
the
top
three
as!
reference
layers
tend
to
be
merged
into
one
program
at
the
highest
Application
layer.
The
OSI
model
also
provides
for
a
Session
layer
which
handles
user
validation
and
mapping
host
names
to
network
addresses,
and
a
Presentation
Layer
which
handles
machine
differences
like
byte
swapping
and
terminal
standardization.
ik>RELIMINARY
Upper
Levels
A60
Ethernet
Manual
17
Telnet
Telnet
is
a
remote
login
program
based
on
the
concept
of
a
virtual
Terminal.
The
virtual
Terminal
has
a
set
of
default
conditions
that
can
be
changed
by
'negotiation'
and
mutual
agreement
between
the
local
and
remote
host.
For
instance
echo
is
by
default
local
and
the
local
host
is
expected
to
buffer
lines
of
text
until
<Return>
is
pressed.
The
Telnet
connection
has
an
escape
mechanism
where
hex
FF
is
the
escape
character,
FF
occurring
in
the
data
stream
is
transmitted
as
FF
FF.
The
escape
character
preceds
an
option
negotiation
which
contains
a
code
to
indicate
WILL,
WON'T,
DO
or
DON'T
and
the
particular
option
code.
Before
an
option
will
be
implemented
on
both
sides
both
sides
have
to
positively
agree
to
do
it.
The
escape
sequence
is
also
used
to
implement
'out
of
band'
signals
such
as
Abort
Output,
Interrupt
Process
and
Erase.
FTP
The
File
Transfer
Protocol
uses
a
Telnet
connection
for
User
authentication
and
control.
The
control
is
achieved
using
a
pommand
and
Response
Dialog
which
mayor
may
not
be
visible
to
the
user.
Commands
are
of
the
form
"USER
Simon"
and
"STOR
pic.rgb".
FTP
responses
are
preceded
by
a
three
digit
code
which
allows
a
machine
to
assess
the
required
action.
Some
of
the
meanings
are
listed
below
1xx
is
a
positive
preliminary
reply
2xx
is
a
positive
completion
reply
5xx
is
a
Permanent
negative
completion
reply
xOx
is
a
Syntax
error
x2x
refers
to
a
connection
x5x
refers
to
the
file
system
PRELIMINARY
FTP
Other manuals for A60
2
Table of contents
Other Abekas DVR manuals
