Builder · Lesson 20

How do network standards make systems interoperate?

A network standard gives different vendors a shared contract. X.25 shows how that contract once defined reliable packet-switched access across public data networks.

A shared contract

Interoperability starts with an agreement.

A standard describes behavior precisely enough that independent systems can connect: message formats, states, timing, addressing, error handling, and the meaning of each field. It does not require every device to use the same hardware or code.

Standards can be published by formal organizations, developed by an industry group, or adopted through widespread practice. Their value is the boundary they make predictable.

Compatibility is a promise between implementations.A standard does not move a packet by itself. It lets two implementations agree on what to send, what to accept, and how to react when something goes wrong.

Name the layers of the idea

Standard, protocol, implementation

01

Standard

The documented contract: formats, procedures, constraints, and expected behavior.

02

Protocol

The communication rules used by participants to exchange messages and change state.

03

Implementation

The hardware or software that applies those rules on a real network.

04

Interoperability

The observable result: independent implementations can communicate as intended.

A few landmarks

Different standards answer different questions.

01

Ethernet · IEEE 802.3

Defines widely used wired local-network behavior, including frame formats and access to the medium.

02

Wi-Fi · IEEE 802.11

Defines wireless local-network communication, radio access procedures, and frame behavior.

03

HDLC · ISO/IEC 13239

Describes bit-oriented framing and link control for point-to-point and other data links.

04

Internet protocols · IETF RFCs

Specify the interoperable rules of the Internet family, including IP, TCP, DNS, and HTTP.

05

Frame Relay

Provided a simpler virtual-circuit WAN service after X.25, with less recovery work inside the network.

06

ATM

Used fixed-size cells and virtual circuits to combine traffic classes with predictable handling.

07

MPLS

Uses labels inside provider networks to select forwarding paths, separate VPN contexts, and engineer traffic treatment.

08

Packet switching

Shares links among independently addressed data units instead of reserving a permanent circuit for every conversation.

Do not compare standards only by age.Ask what boundary each one defines: a local medium, a link, a provider service, or an end-to-end exchange.

Historical case study

X.25 turned a public network into a packet service.

X.25 was a family of ITU-T recommendations for connecting customer equipment to a public packet-switched network. It was designed for links that could be noisy or unreliable, so the network offered substantial error detection, acknowledgements, and flow control.

The customer side is the ETTD (equipement terminal de traitement de donnees, or DTE). The provider side is the ETCD (equipement de terminaison de circuit de donnees, or DCE). The ETTD requests access and exchanges X.25 packets; the ETCD presents the network interface and provides access to the provider's packet-switched service.

ETTD and ETCD define the boundary.In a simplified view, the ETTD is the customer device and the ETCD is the operator's access equipment: ETTD <-- X.25 interface --> ETCD.

A connection could use a virtual circuit: the network associated packets with a logical connection and forwarded them through an agreed path or state. The packets still shared the provider infrastructure, but the customer experienced a managed conversation rather than an isolated physical circuit.

X.25 is not the same as the modern Internet.It is a useful historical example of a provider-managed packet network, with stronger reliability and connection state inside the network than the best-effort IP layer normally provides.

Read the units on the wire

A LAPB frame carries an X.25 packet.

X.25 is easier to understand when its two nested structures are separated. LAPB, derived from HDLC, frames one hop and supplies link-level sequencing and error detection. Its information field carries the X.25 packet used by the packet layer.

Diagram of a LAPB frame containing an X.25 packet with GFI, logical channel number, packet type, and user data
01

LAPB frame

Flag, address, control, information, and frame check sequence protect the local data-link exchange.

02

GFI

The General Format Identifier describes packet-header format and related options.

03

LCGN + LCN

The logical channel group and channel number identify the virtual-circuit context at the interface.

04

Packet type

The packet-type field distinguishes call setup, data transfer, clearing, flow control, and other procedures.

The header does not contain the whole route.The logical channel identifies a managed context at the relevant interface; the provider's internal switching state determines how packets progress through the network.

OSI transport standard

X.224 defines a connection-oriented transport exchange.

X.224, also published as ISO 8073, defines the OSI transport protocol. It provides procedures for opening a transport connection, transferring data, and closing the connection. The protocol exchanges TPDUs (Transport Protocol Data Units), not X.25 packets or LAPB frames.

In the historical stack discussed here, X.224 belongs above the X.25 packet layer. X.25 moves packets through the provider network, while X.224 organizes an end-to-end transport conversation between transport entities.

X.224 transport exchange showing CR connection request, CC connection confirm, DT data transfer, and DR disconnect request
01

CR

Connection Request starts a transport connection attempt.

02

CC

Connection Confirm accepts the transport connection.

03

DT

Data Transfer carries transport user data.

04

DR

Disconnect Request begins an orderly release.

Do not confuse the layers.LAPB carries frames on a link, X.25 carries packets through the packet network, and X.224 carries the transport conversation between endpoints.

Read the history correctly

Old standards explain design tradeoffs.

X.25 added recovery where its links and services needed it. Later technologies often moved more responsibility to endpoints and used faster, simpler forwarding in the network. Comparing them teaches a general lesson: reliability, state, cost, delay, and simplicity can be placed at different boundaries.

When studying any network standard, ask four questions: who publishes it, which boundary it defines, what failures it handles, and which parts are left to an implementation.

Knowledge check

Test your standards model

Answer twelve questions. The answer order changes each time.

0 / 12