=[ x25-box.com ]=

_________________________________________________________________

X.25 PACKET NETWORK ACCESS STANDARD

Assume that there is a chain of motels. There is a central site, which
accepts calls by phone from customer, who want their rooms reserved at
different location all across U.S. Each of the motels are equipped
with personal computers. Now the central site wants to send
reservations to different sites, what choices are there. Certainly
leased lines are impractical, next choice will be switched long
distance phone lines to send data, which can be still become costly
per call. Instead there is another choice, public packet switched data
networks. The central site gives the network a data packet along with
the destination address. The public network will deliver the packet at
the destination site.

The same thing can be achieved by the remote motels to send their data
packets to the central site, or any other destination. To achieve this
each sending node has to establish a virtual circuit to the
destination, then send the packet(s).

Figure X.1 Packet switched Network [van]

When a user site signs up with a packet switched network (with a
common carrier), there are certain "ACCESS POINTS" to the network that
it can connect to. These access points can be at the user's corporate
office, or the users might dial up to the access points through
conventional telephone lines.
_________________________________________________________________

X.1 Virtual Circuits and Datagrams

In virtual circuit systems, a logical connection is established before
any packets can be sent. This can be achieved by first sending a
CallReq and then getting CallAcc packet back. On the other hand each
datagram is handled independently. A datagram contains the Source and
Destination address on each packet. No Call setup has to be
accomplished first. If a network provides a virtual circuit service
the packets always arrive in sequence to the user on the other end. If
a network provides a datagram service, packets may arrive out of
sequence. X.25 provides a virtual circuit service.

Other important distinction between datagrams and virtual circuit
service is that in datagrams (packets) every packet contains the
address of the destination and source devices, on the other hand in
virtual circuit only the Call-Req packet contains the full
sender/destination address, after that when the session is established
then the further packets only contain the virtual circuit number which
is assigned to the caller when he sent a Call-Req packet.

Figure X.2 Datagram VS. Virtual circuit traffic [vcdgram]
_________________________________________________________________

X.1.1 Virtual Circuits and Packet multiplexing

We are all familiar with TDMs (time division multiplexing) where more
than one R-232 ports are connected over one communication line. The
idea of packet multiplexing is the same, except here the packets from
different sources/destinations may be multiplexed on the same
communication lines. The device (may be a host) can then de-multiplex
these packets (for different programs/tasks) by looking at some type
of multiplexing ID, in the case of X.25 packets these are the assigned
virtual circuit numbers. In the following figure packets from two
sources with their own virtual circuit numbers "VC-1 & VC-2" are
multiplexed on the same line.

Figure X.3 Packet multiplexing [VCmux]
_________________________________________________________________

X.1.2 Packet Switching and routing

Access points are connected to switching nodes, and then each
switching node inside the network is connected to two or more
switching nodes. These nodes make the packets travel through the
network, by crossing one or more switching nodes till they arrive at
their destination. When the packets are delivered to the switching
node (via access points) the switching nodes depending upon the
destination of the packet make a decision which route to send the
packet.

Figure X.4 Switching-Nodes and Routing [route]

For instance in the above figure Node-1 and Node-2 have setup virtual
circuits 1 & 2 to Node-4. Node-3 has a virtual circuit established
(VC-3) with Node-5, where Node-5 is at different site than Node-4.
Then Node-4 has established VC-4 with Node-5. At this time we have
shown three switching nodes SN-1, SN-2 and SN-3 in the network. By
examining the diagram you will see that SN-1 will send VC-1 & VC-2
over route 1 and VC-3 through route 2 towards SN-2. Similarly SN-2
will send VC-1 & VC-2 packets over route-1, but divert VC-4 over
route-3 towards SN-3. Likewise SN-3 sends VC-3 over route-2 towards
SN-1, but VC-4 through route-3 for SN 2. These switching nodes are
typically PDP-11s or Micro-VAXs, they are also known as store and
forward nodes. Since each of them receive packets from one input
lines, put them into their buffers (store) on a temporary basis, and
then when they get a chance they will deliver it to the appropriate
output line (forward).

One important thing to remember is that X.25 provides virtual circuits
connection, so once the session (circuit) is established the same
route is taken till the session is terminated. On the other side in
case of datagram service, since each packet (datagram) contains
complete source and destination address every packet can be routed
independently of each other depending upon the loads on the lines
connected to the switching nodes.
_________________________________________________________________

X.2 The X.25 Interface and layering

Consider a case the you have logged from your home into the host
computer of your university via modem. Most probably you are sending
and receiving information from the host on a character by character
mode. In other words your are only using the physical layer of the OSI
without any error recovery done by any upper layers. Every time you
type a character it gets transmitted and vice-versa. You might have
also noticed on occasions, that some characters were damaged while
being sent from the host to your terminal. These characters are shown
garbled on the screen. Or at the same time your host might send you a
"parity error" message back, in which case you may have to retype the
entire line. However when you are doing any file transfers between
host and your PC at home over modem and phone lines by using (say)
Kermit. then the Kermit protocol uses a DLC layer over the physical
layer (similar to Bisync DLC), which also calculates the CRC. So if
there is any bit damaged the entire frame is re-transmitted. It is the
DLC layer which recovers from any bit damaged at the physical layer.

When a user gets connected to the public packet-switched networks,
their requirement (mostly) is that you want to interface to the
network then you send them the data, not in character form, but
instead in packet form. The set of "INTERFACE" is known as X.25
standard. The purpose of the packet mode is that in case if there is
any part of the data damaged between a customer premises and the
access point, the X.25 standard will recover the packet/frame by
re-transmitting it. Another thing to remember is that these networks
are some times also called as VANs (Value added Networks). The reason
they are called value added is that these public carriers have added
certain qualities to their networks. For instance once you have
delivered an undamaged packet to the network access point, the network
itself will route it to the proper destination, if there has been any
error during the travel of the packet within the network itself, the
network switching nodes will automatically recover it, without ever
asking the user node for re-transmission. It is very important to
remember that X.25 is an "Interface standard" to Value Added Networks.

Figure X.5 X.25 Interface to the Value Added Network [van1]

According to CCITT terminology, the user machines are referred to as
DTEs or Data-Terminal Equipment. It does not mean that it is a
terminal, instead it is one of the end (terminal) end points of data
transmission. A DCE, or Data-Circuit-terminating-Equipment, is a
network node (possibly a switching node). In X.25 standards the
network nodes do not assemble/de-assemble packets, instead these
functions should be performed by an interface (mostly software)
between the DTEs and DCEs. According to X.25 recommendations, a user
information passes through the DTE/DCE interface of the user machine,
where it is transformed into packets, address may be put, virtual
circuit is added, sequence numbers are added, type of packet is
inserted, and other user-to-network operations are performed. Then
this is sent to the DCE, from the DCE the packets travel through the
network to the destination DCE, where they are finally passed to the
X.25 interface to the destination DTE.

This Interface provides three layers. The First is the Physical layer,
Secondly a DLC, which is a subset of HDLC and finally the third layer
is the Packet layer.

Figure X.6 X.25 Recommendations between DTE and DCE [DCE]

A brief description of the X.25 layers are described as follows:

Figure X.7 The Layering Of X.25 Interface [Layers]

o Physical:

X.21 or X.21 bits to provide circuit switching, X.21 is similar to
RS-232-C

o Link:

This layer is essentially a DLC layer. A subset of HDLC is used. This
layer can use both the SLP (Single link procedure) and MLP (Multi-link
procedure). LAP-B is used for SLP. A link layer can also be used as
MLP or just like TG (transmission group) of SNA ( System Network
Architecture). This standard uses transmission over multiple links
rather than one, to achieve greater throughput. To understand more
about use of MLP please refer to X.25 PADs section. To send a packet
through MLP, a 16 bit MLP header is attached to the header. Out of
these 16 bits, 12 are used for 12 bit sequence number. Then this data
and MLP header becomes the data to the SLP frame as shown below:

Figure X.8 Multi-Link-Procedure Frame [mlp]

o Packet layer:

This layer establishes and tears down multiple virtual circuits. Two
type of virtual circuits are possible. PVC (permanent virtual
circuits) and SVC/VC (switched virtual circuits). This layer does not
Fit in ISO layer 3. However it does have certain features of OSI 3rd
layer and some features of OSI transport layer.

Address length fields:

Here the octet 4 consists of field length indicators for the called
and calling DTE addresses. Bits 4,3,2 & 1 indicate the length of the
called DTE address, and bits 8,7,6 and 5 indicate the length of the
calling DTE address. Both of these indicator fields represent number
of semi-octets(4 bits) in the address fields. Each indicator field is
binary coded and bit 1 or 5 is the low order bit of the indicator.

Address fields:

Octet 5 and the following octets consists of the called DTE address
when present, then the calling DTE address when present. Each digit of
an address is coded in 4 bits in binary coded decimal form, with bit 1
or 5 being the low order bit of the digit. Starting from the high
order digit, the address is coded in octet 5 and consecutive octets
with 2 digits per octet. In each octet, the high order digit is coded
in bits 8,7,6 and 5. The address field (calling DTE and called DTE
both is known as one address field) shall be rounded up to an integral
number of octets by inserting zeros in bits 4,3,2 and 1 of the last
octet if necessary.

Facility length field:

Bits 6,5,4,3,2 and 1 indicate the length of facility field in octets.

Facility field:

This field is present only when the DTE is using an optional user
facility requiring some indication in the Call request and incoming
call packets. Bits 8 and 7 are unassigned are set to zero. The
facility field contains an integral number of octets. The actual
length of this field depends upon the facilities which are offered by
the network. However this maximum does not exceed 62 octets. Possible
facilities are flow-control parameters (e.g., window size), packet
re-transmission period, reverse charging etc.

User data:

Following the facility field, user data may be present. The user data
field may be 16 octets or more until 128 octets.

Figure X.12 Call Set up and Clearing Phase [session]

Steps performed in setting up a virtual circuit and clearing up.

1: Host X generates a call_req packet, puts the group number, virtual
circuit number, complete source and destination address. And send this
packet to its X.25 network switching node or communication processor.
In technical term Host X & Y are called DTEs (Data terminal
equipment). And communication processors or switching nodes connected
to X & Y are called DCE (Data circuit terminating equipment). In other
words Hosts are called DTEs and Nodes are called DCEs.

2: The Network communication processor send this packet to the
destination network communication processor, to which host Y is
connected to.

3: Switching node of host Y selects an appropriate VC numbers, and
send the packet as Call Indication packet (or incoming call).

4: II host Y wants to accept this call, it generates a Call- accepted
packet, and uses the same VC numbers, as that on Call-Indication
packet, and send this packet to its Switch node.

5: After this is routed through the network, and reaches to switch
node of host X, this switch node puts the VC number, which the same,
that host X put on its Call-Req packet.

6: After the virtual circuit is established, then further data
transfer occurs, without any source and destination address contents
on the packet. Only the VC number is used.
_________________________________________________________________

X.6 Important fields of packets

Byte 1: bit 8 = Q bit, bit 7 = D bit, Bits 6-5 represent Modulo 8 or
modulo 128 is being used. 01 = Modulo 8, 10 = Modulo 128
_________________________________________________________________

X.6.1 Q bit (Qualifier bit)

Though the Q bit is not defined in the standard. But can be used to
identify that this packet, carries PAD information.
_________________________________________________________________

X.6.2 D bit (Deliver confirmation bit)

If D=0 then it produces a local acknowledgment. This Ack is done
between the DTE and either the local DCE or remote DCE,to continue the
flow control. It d=1 then the target DTE will generate the final "ACK"

Figure X.13 Local and Remote Acknowledgment [ack]
_________________________________________________________________

X.6.3 M bit (more-data bit)

Packets contain the P(s) and P(r) with send and receive sequence
numbers. These sequence numbers are used between DTE and DCE for local
control flow. The packet size defaults to 128 bytes. Optionally 16,
32. 64, 256, 512 and 1024 bytes can be allowed. The user will specify
at subscription time a particular packet size. It is possible that the
network has to divide a packet into smaller packet sizes. A typical
example can be a host with using 256 bytes packets. The other DTE may
be a simple terminal, and can only handle 64 bytes of user data. The
network then will divide the incoming packets from CPU into 4 smaller
packets. each of 64 bytes. In order to signal the destination DTE that
there is a relationship between packets, the DCE will use the
More-data-bit, to express the relationship. It will put M bit "ON" for
the first 3 packets, and M bit "OFF" on the fourth one. The P(s) and
P(r) delivered to the destination DTE will be different than the P(s)
and P(r) of the originating DTE. If the D-bit is set, then the network
has to manage the delivery of correct P(r) to the originator, after
the receiving DTE has returned a P(r) equal to the P(s)+1 from the
packet it received with D-bit on. The opposite can happen as follows.
Now the CPU is expecting 256 byte packets. Now the network can combine
subsequent packets if M bit is set, the packets are full and D-bit is
OFF. If the D-bit is set, the network should provide the correct P(r)
to the sending DTE. While a D-bit response is pending, the window of
the flow control mechanism remains shut.

Figure X.14 Use of More bit [m-bit]
_________________________________________________________________

X.6.4 A and B type packets

A packet is of type A or B if it contains:
A
B
------ ------
M = 1
Not A type
D = 0
packet is full(max length)

When a larger packet is broken down, it starts with A packets, and
then a B packet. X.25 can combine these packets to and from a large
packet. On the other sides, if a SN has to combine the packets, it
needs typically As and then a B packet to make a larger packet.

Figure X.14 A and B packets example [ab-pkt]
_________________________________________________________________

X.6.5 Source DTE & Destination DTE address length

X.25 can handle different addressing schemes (will be explained
later). This field is only 4 bit only, which can show values 0-15 This
shows the length of address field is how many 4 bits.
_________________________________________________________________

X.5.6 Source and Destination DTE address

This is variable length field.
_________________________________________________________________

X.6.7 Facility length (16 bits)

This mentions the facilities filed in octets
_________________________________________________________________

X.6.8 User Facilities

Possible User facilities are:

o Closed User group: On a Call-Req packet, the originator may specify
which user group he belongs to.

o Flow-Control Parameters: Specifying the Window size and Maximum data
field length requirements.

o Throughput requirements: Typical throughput requirements, typically
from 75-bps to 48 Kbps.

o Fast Select: The DTE is interested this message to be treated as
fast-select facility packet.
_________________________________________________________________

X.7 Virtual Circuit Assignment

A DCE can allow its DTE to setup 4095 Vs, over the network, to
different DTEs. Each packet contains 12 bits of virtual circuit
number, which is made up of 4 bits of logical group number and 8 bits
of logical channel number fields in the packet. VC 0 is always used
for "Restart" and diagnostic packets. If the DTE can have only one VC,
then VC # 1 is assigned. Then the rest of the VC #s are broken down
into 4 categories:

1: Permanent Virtual circuits 2: Oneway In-coming Calls
3:
Twoway Calls 4: Oneway Out-going Calls

Assume that
Node A and B has the following set up:

NODE A NODE B
------
------
Num. PVCs 100 200
In-coming 200
300
Two-way 100 200
Out-going 100
300

Figure X.15 Virtual Circuit Number Assignments [circuits]

Assume that nodes A and B, have no other SVC active. Node A wants to
set up a session with node B. It looks at the outgoing range, and
picks up the highest unused VC, which happens to be 500. It puts that
in the CallReq packet. When this CallReq packets reaches the
destination DCE, it assigns it lowest possible VC from the incoming
VC, which happens to be 201. This packet when arrives at the dest DTE,
if call is accepted, the DTE will use 201 as the VC for the CallAcc
packet. When this packet reaches source DCE, source DCE puts VC 500 in
it, and sends it to source DTE. This way the connected DCE and DTE
multiplex packets between each other. The two-way calls are like
over-flow area. If active in-coming calls are more than 200, then the
next incoming call will have a VC from two-way zone. Same rule applies
for the outgoing calls. However the worst case is when there is only
one VC left in the two-way range, and just by chance Node A wants to
send a CallReq packet, and at the same time the DCE has received a
CallReq destined for Node A. Then DCE might pick up the same VC
number. However when DCE detects a duplication of the same VC on two
different packets, it will issue a ClearReq to both sides, and VC will
be released. This situation is known as CLEAR COLLISION.
_________________________________________________________________

X.8 More about X.25 packet types
_________________________________________________________________

X.8.1 Reset and Restart packets

Two types of packets are used by X.25 to receiver from errors. The
Reset packet is used if the receiver and the transmitter are out of
Sync. In other words P(s) is not the same as P(r) on the receiver
side, till it is time to send the Reset, which will make the sender
and receiver nodes, start with sequence number of zero. Any data and
interrupt packets in the transit are lost, and the upper layers will
have to recover from it. For reset packet, it still uses the existing
virtual circuit number, and affects only the VC# in question.

Restart is issued when error is more serious. The VC#0 is used for
Restart packets, it clears up all the virtual circuits. To start
again, CallReq procedures have to be issued. This can happen, if the
access to the X.25 is lost, or DCE becomes down.
_________________________________________________________________

X.8.2 Interrupt packets

The interrupt packet is transmitted with higher priority than the
normal data packets. It does not contain any sequence numbers on it.
It can also contain up to 32 octets of user information. It cannot be
stopped by closed window on the receiver side. The interrupt packet
always gets end-to-end acknowledgment. Another interrupt packet cannot
be sent, until the previous one sent has not be acknowledged.
_________________________________________________________________

X.8.3 Fast Select Service

Basically X.25 is designed for virtual circuit service. In other words
before any data can be send, first a session has to be established, by
going through call set up procedure. For example first sending a
CallReq packet, etc. Let us now look at the example, that now a store
wants to use credit card usage, by connecting to the central data
bases by X.25 network. They are normally not using PVC circuits.
Rather they are using SVC. This way they have to use more than one
packet, to get the information about the card. In other words the goal
was to use a service just like datagram service. Another approach was
taken, known as the Fast-Select service of X.25. Here to verify the
amount on the card, will generate a CallReq packet, with Fast-Select
facility in this packet until 128 bytes of information can be stored,
(though the normal CallReq packet only 16 octets of data can be sent)
which may contain card number, amount etc. When this CallReq packet
arrives at the database system, the destination send a ClearReq
packet, along with 128 bytes of credit card data. So in just one
exchange a virtual circuit was established and then tore down. Fast
select facility is used for the above mentioned type transaction
oriented applications. To use this facility the calling DTE mentions
in its facility field of its Call-Request packet.

Figure X.16 Example of Fast Select Service [fast]
_________________________________________________________________

X.9 X.2 (Optional packet-switched user facilities)

The standard is X.2, the hosts which want to use X.25 network, will
negotiate the following facilities, with X.25 network people, and can
use the facilities for the period of contract. Some of the facilities
are:

1: Throughput:

This facility allows the user (per call basis), the number of bits
which should be transmitted per VC. The typical range is 75 bps to 48
Kbps

2: Flow Control Parameters:

The user (per call basis) can specify its window size and maximum
packet size.

3: In-Coming Calls Barred:

This facility will not let any incoming calls get to this DTE.

4: Out-Going Calls Barred:

No outgoing calls will be accepted, by the connected DCE

5: One-Way logical Channel Outgoing:

If all the in-coming, outgoing and mixed calls are e.g. 300. Then this
facility will set the lowest outgoing VC# to be 200. The next outgoing
will be 199, and so on. All the other VC# will be less than 200

6: Fast Select Acceptance:

If this facility is accepted, then the DCE will send Fast select calls
to the DTE. If the DTE does not have fast select implemented, then DCE
can block these calls.

7: Closed User group:

A DTE can belong to any closed user group. A DTE can belong to more
than one closed user group. The DCE will only send the incoming
message from the network, if the DTE belongs to the closed user group,
mentioned in the group# mentioned in the first byte of the X.25
packet. This way DTEs are protected from unauthorized calls.
_________________________________________________________________

X.10 Overview of X.25 layers

+--------+---------+---------+
| X.21 | HDLC | Packet |

| Level 1| Level 2 | Level 3 |

+---------------------+--------+---------+---------+
| Connection set up |
X | | X |

+---------------------+--------+---------+---------+
| Data transfer setup |
| X | X |

+---------------------+--------+---------+---------+
| Error control |
| X | X |

+---------------------+--------+---------+---------+
| Flow control |
| X | X |

+---------------------+--------+---------+---------+
| Connection clearing |
X | | X |

+---------------------+--------+---------+---------+

Figure X.17 X.25 layers functions [?]
_________________________________________________________________

X.11 Tariffing of X.25 PDN (public digital network)

The X.25 tariff structure has basic three parts.

1: The tariff tends to be distance independent within a country.

2: All of the tariffs are volume dependent

3: There are charges for attachment to the X.25 PDN, the service and
the optional facilities.

When one user attaches himself to the PDN, there is a connection
charge. It includes the cost of physically attaching the DTE to the
network. The monthly charges are composed of the elements shown in the
figure. As mentioned above first there is an attachment charge, which
depends on the speed of the connection. Optional facilities defined on
a DTE basis are also part of this monthly attachment charge. Charges
for the usage of virtual circuit services are the second part. There
is a monthly usage charge per PVC. For the SVCNC the charge is
computed as the duration of the connection, along with a fixed minimum
charge. Optional facilities for PVCs, and for VCs per call are also
added. Finally is the volume of data. It is computed as number of 64
octets (or bytes) segments that are sent. Depending upon the network,
the segment cost will vary from the time of the day. overnight
transmissions are less expensive than the day time.

+----------------+++---------------------------------------------+
|
CHARGES ||| MONTHLY CHARGES |

+----------------+++---------------------------------------------+
| Access
to X.25 ||| Speed-dependent + User Options |

+----------------+++---------------------+-----------------------+
|
||| Permanent VC charge | Switched VC charge |
| Connections
||| per PVC + options | per time connection + |
| |||
| options |

+----------------+++---------------------+-----------------------+
| Volume
||| Charge per 1000 bytes |

+----------------+++---------------------------------------------+

Figure X.18 X.25 tariffing [?]
_________________________________________________________________

X.12 Addressing Flat and Hierarchical and X.121

The address of destinations can be put in two ways 1: Flat Addressing
and 2: Hierarchical Addressing. The flat address can be thought as a
single universal counter, every time a new address is needed
regardless of location of geography, the current value of the counter
is allocated and then the counter is incremented by one. Social
security numbers are a good example of flat addressing. Because by
looking at the number one cannot judge that this belongs to New York
or Ohio etc.

Hierarchical addressing on the other sides breaks down the address
into certain sub-fields. For instance an address can be specified as:
[Hierarchical Address] = [country] [network] [host] [socket], A good
example of this type of addresses is the phone system, for instance:
513-229-3831 specifies as [area code] [end office] [phone]

CCITT has recommended a fourteen decimal digits public networks
similar to public switched networks known as X.121. Each digit can be
represented in 4 bits, and all the 14 digits can be accommodated in 7
octets, if encoded in binary form and 14 octets if in ASCII/EBCDIC
readable character form. Here the first three digits give country
code, 4th digit is the network number. The rest of the 10 digits are
local to one independent network, and they can use it, the way they
prefer, however one possible way is to assign 7 digits to an
independent host and the last three digits to a particular socket/port
(or an application program) on that host. U.S. has been assigned
country codes of 310 through 329, and Canada from 302 through 307. For
U.S. the first four digits give 200 networks and for Canada 60
networks altogether. For instance in case of U.S. networks can be
given numbers from 3200 till 3299. In other words 3299 3200 gives
distinct 200 networks. Country codes with initial digits of 8 is used
for public telex, and initial digit 9 for telephone networks.

Figure X.19 X.121 addressing Scheme [x121]

Though the above mentioned scheme can result in 1 million host, each
having 1000 addresses. Even if 10 digits local to any packet network
is not sufficient then the user data field at the end of Call-Request
packet can be used for sub-addressing to give the desired address
resolution.
_________________________________________________________________

X.13 Universal Addressing #

ISO along with the CCITT has come up with a universal addressing
scheme, which will accommodate practically every type of switched
voice and data communications needs of today's networks. This contains
a common format for addressing and network routing including every
thing from packet switched networks, to telex and to international
telephone calls. This plan is also known as "Network Layer Addressing
Plan". Following are a few important existing addressing formats:

CCITT X.121:

As mentioned above.

CCITT F.69:

The Telex network numbering plan, which use 8 decimal digits.

CCITT E.163:

Numbering plan for public-switched public telephone network, which is
12 decimal digits long.

CCITT E.164:

Possible ISDN (Integrated Services Digital Network) for combined
voice, data and video, standard format is 15 digits.

ISO DCC:

4-digits country code similar to X.121, rest is open.

ISO ICD:

Binary or digital data-device addresses that uses 4-digit
International Code Designator prefix that identifies a specific
organization's international network. Such as NATO or SITA airlines
reservations network.

ISO OSI:

ISO open System Interconnection addressing, binary data device
addressing including 6 octets (48 bits) or 7 octets (56-bits) local
networks, NSAP (network Service Access Point) Address and so on.

Figure X.20 Universal Addressing Plan [megaplan]

The ISO addressing plan can include any of the above mentioned
addressing plan. It the address is binary encoded, then it can fit
them in 20-octet field or 40 character (ASCII/EBCDIC e.g.,) fields, if
the address is composed of decimal digits. In either case, the first
two characters of every address is known as the AFI (Authority and
Format Identifier) field. This special AFI field tells the network
that the rest of the address section complies with which of the above
mentioned address format (e.g., X.121, telex, ISO OSI etc.). It also
specifies that if the address if in binary code of character decimal
form. For instance a 37 AFI tells that the address is X.121, and it is
binary encoded, the first 4 digits (2 octets) contain the country code
and the network number. [Ref. Datacomm/May 1985 pg.64]
_________________________________________________________________

X.14 X.25 Pads

As mentioned earlier that the access to the public switched networks
is via X.25 standard, which contains three layers in it. As long as
two computers are connected to each other both of them can have X.25
interface built in them to communicate over the network. However when
people wants to access the remote host via public switched networks
through terminals which have no programming capabilities then X.25
Pads are used to solve the problem.

Figure X.21 X.25 Padset up [pad1]

Typically these Pads are available in 4, 8, 24 and 48 RS-232 ports.
One of these ports can be assigned to the supervisory mode. From this
port the manager can set up certain parameters. For instance as
mentioned earlier that network use 14 digits addressing scheme, these
addresses are hard for a user to remember. So the manager can assign
addresses to certain mnemonics. Now a user can use the X.28 command
(Selection PAD):

CALL,MIAMI,RC=Y [ret]

This command means set up a virtual circuit to preset location named
MIAMI, with reverse charged. Here the address and other parameters of
the MIAMI site has already been set by the manager through supervisory
port. Once the PAD receives this command, it prepares a Call Request
packet along with the logical channel number and the group number and
delivers to the packet network. The Pad receives an X.25 packet from
the network, depending upon the virtual circuit, it puts into the
buffer of the designated RS-232 port. Then it delivers the data part
to the start/stop (character mode) to the device connected to it via
RS-232 port. It also gathers the characters sent by the attached
terminal, once (say) when the [return] key is hit, it prepares an X.25
packet and send it to the network. It keeps sliding windows per device
it is attached to.

The X.25 pads also comes with MLP (Multi-link-protocol) feature to
achieve greater through put over the network. For instance with
dual-data-links application speeds until 19,200 BPS full-duplex can be
achieved. These two data-links parameters are separately programmable,
for instance window size, time-out and retry counters and type of
network supported.

Figure X.22 MLP (Dual-data-link) application [pad2]

Because of the flexibility of the multiple data links treated
independently, each link can have a distinct network identification
code and call address. For example port1 can be connected to a remote
resource via TYMNET through one data-link, and port 2 can access the
remote host via Accunet by the second data-link.

Figure X.23 Access through different networks [pad3]

These Pads meet certain requirements of CCITT to support the following
domestic U.S. packet switching networks e.g., Accunet, CompuServe,
Telenet, TYMNET and Uninet. and also international networks as:
Datapac, Datex-P, DCS, DN1. PSS and Transpac etc.

Besides each of the ports can be accessible by certain passwords,
setup by the manager, can have adaptive speeds, idle port disconnect
timer. These PADs can also be programmed to delay the action taken
upon the arrival of a Clear-Request packet for as long as 255 seconds
in order to prevent any loss of data. If the terminal has not sent any
data for quite some time, the pad can send a "receiver ready"
supervisory frame (HDLC). This interval can be specified from
typically 1 to 255 seconds.