X.25 is an analog packet switching network. It can be considered Slow Packet Switching. The transfer speeds are typically 56 kbps to 2.08 Mbps. There was a world-wide set of Public X.25 Networks and it is possible for an organization to have its own private X.25 network.

You don't see X.25 networks too much anymore. As a matter of fact, a google search will show very few if any still in service. So why learn about X.25? Many other current protocols were developed on the sound principles of X.25 such as Frame Relay, MPLS and others. X.25 introduced the concept of virtual circuits and permanent connections in a packet based network.

X.25 is over 30 years old and an established technology. It was developed to guarantee end to end delivery of data in an analog telecom world. At that time, WANs were unreliable and you needed a very robust protocol to guarantee that the data would arrive as expected.

There are many multi-vendor solutions and dissimilar technologies in an organization are allowed to access the X.25 network. In Canada, the main X.25 network was called Datapac which is a public offering of X.25. You pay either a flat rate or by the packet.

X.25 is used to connect LANs together. Due to its slow transfer speed, it is used for:

  • Host terminal emulations: low data
  • Client/Server applications such as E-mail: small files, bandwidth
  • File Server: large amount of data & real-time traffic (doesn't work well)
  • Databases: usually large databases but queries are small inbound and medium size outbound.

X.25 has a high protocol overhead compared to other networks. This reduces the transfer speed and bandwidth utilization - means its not as efficient.

Overhead Example:

Truck A represents X.25, it has a heavy empty weight of 5 tons (overhead). The bridge (medium) only allows 6 tons of weight, this means that Truck A can only carry 1 Ton of cargo (Data). Truck B is a smaller truck and weighs 3 tons empty, this means that it can carry up to 3 tons of cargo (data) across the bridge (medium). Truck B makes better use of its weight when crossing the bridge - utilizes its bandwidth better (it is more efficient).

X.25 OSI Layers

X.25 consists of 3 layers:

  • X.25.3 - Network Layer:
    • PLP (Packet Layer Protocol) or
    • SNDCF (Subnetwork Dependant Convergence Function)

  • X.25.2 - Datalink:
    • LAPB (Link Access Procedure Balanced)
    • HDLC (High Level Data Link Control)

  • X.25.1 - Physical with 4 different types:
    • X.21 - Sync Digital Interface 9.6kbps - unbal, 64Kbps - bal
    • X.21bis - Leased Line Analog Interface
    • V.24 - RS232 Leased Lines
    • V.35 - RS232 Duplex operation over Leased Lines

In actual fact, pure X.25 only defines the DTE to DCE connection but when we talk about the X.25 Packet Switching Network we talk about the above 3 layers.

X.25.1 is the Physical layer and is uses 4 flavours of medium (similar to the multiple Ethernet flavours: 10BaseT, Thinnet and Thicknet). The X.25 packet is carried on serial data lines.

X.25.2 uses HDLC & LAPB for the Data Link layer. LAPB is considered a subset of HDLC. Both are similar to IEEE-802.2 LLC (Logical Link Control) and provides 2 way communications. The B in LAPB stands for balanced communications and is another way of saying Full-Duplex - both sides communicating at the same time. The X.25 packet is carried within the LAPB frame's info field similar to how the LLC packet is carried within the MAC frame's info field.

X.25.3 is known as SNDCF or Subnetwork Dependant Convergence Function. X.25 uses IP network addresses and is one of the reasons for the high overhead.

X.25 connects to the network using a DCE modem or DSU/CSU (Data Service Unit/Channel Service Unit). X.25 allows 4096 logical channels to be connected on 1 physical connection. The Packet Assembler/Disassembler (PAD) connects the DSU/CSU to the DTEs (user devices) which can be terminals or LANs.

The X.3 standard governs the operation of the PAD and the X.28 standard governs the operation of the PAD to terminal connection. The X.29 standard defines the End to End communications from DTE to DTE through the X.25 Network.

X.25 High overhead

X.25 has a high overhead because it provides extensive error checking. Each device in the X.25 network acknowledges every packet sent. This slows down the transfer of information and uses up available bandwidth. When X.25 was first introduced, the quality of the analog phone lines required this extensive error checking but now with digital lines available it is not necessary.

There are 2 types of connections used with X.25:

  • PVC - Permanent Virtual Circuits: These are leased lines and require no call connect/disconnect
  • VC - Virtual Circuits: These are like dial-up lines (switched circuits). They require a call connect/disconnect procedure and end to end communication through the network.

VC (virtual circuits) have handshaking very similar to how modems connect as covered earlier. X.25 is often used because an X.25 network is considered 1 Hop.

Normal Routing (Non X.25)
Between San Francisco to Boston

Dashed Route takes 5 hops (Salt Lake City, Denver, Chicago, Detroit, Boston)
Dotted Route takes 8 hops (Los Angeles, El Paso, Houston, New Orleans, Alanta, Washington, New York, Boston)

With X.25, the complete Network across the United States would look like only 1 Hop. X.25 Packet Networks takes care of the routing path. To connect any point in our example of the USA together would be only 1 hop.

X.25 Packet Formats There are three X.25 packet formats:

  • Call Request - Call connection/disconnection
  • Control Packet - Data control
  • Data Packet - Information transfer

Call Request Frame

The Call Request Frame is used to initiate and setup the call from one X.25 service to another. After the call is established, the Call Request frame is not used.

Logical Channels

There are 4096 Logical Channels available on a single physical connection to an X.25 network. The Logical Channels are divided into Groups and Channels. There can be 16 groups (4 bits) of 256 channels (8 bits). 16 x 256 = 4096. The Logical Channel Numbers (LCN) are used to identify the connections to the Network.

Type Field:

For Call Request the Type Field is always equal to 0000 1011. Note the last bit is called the Control bit and is set to 1.

Length of Calling/Called Address:

The Length of Calling/Called Address allows different sizes of addresses for other protocols. The standard protocol is IP with an address length of 32 bits.

Calling/Called Address:

Only used during Call Connect until the Virtual Channel is established then the LCN (Logical Channel Numbers) are used to identify the connections.

Facilities Length

Indicates the length of the Facilities Field

Facilities Field

Indicates the types of facilities available. Facilities depend on the Network provider and can include information such as Charges or Call Forwarding.

Once a Call is established, The Calling/Called Addresses and Facilities Fields are not required and the Control Packet Format and Data Packet are used.

Control Frame

The Control Frame is used to control the communications during the call.

X.25 Type field table for all 3 packet types

Data Frame

The Data Frame is used to transfer data between destination and source.


Optional - Q is used to distinguish between data and control information.


D controls the type of Acknowledgment:

  • 0 - Network Control: DTE to DCE
  • 1 - End to End Acknowledgement: DTE to DTE


It is used by the sliding window. The modulo allows multiple unacknowledged packets on the network for faster response. The modulo can be Modulo 8 or Modulo 128 (127 packets out on the network)


The Piggyback field is used for Acknowledgement with the Modulus (sliding window)


The Sequence field holds a unique packet number which identifies the packet.


The More field indicates that more data is coming.

If this page has helped you, please consider donating $1.00 to support the cost of hosting this site, thanks.

Return to

TelecomWorld 101

Copyright July 2013 Eugene Blanchard