gui yong nian - stock.adobe.com

Vintage rail freight system showcases 50-year-old innovation

A system from national rail operator British Rail in 1975 used distributed computing and a 400km data network to track freight

Fifty years ago, three IBM System 370 mainframes powered a pioneering scheduling system run by the UK’s national rail operator, British Rail. Called Total Operations Processing System (Tops), when it went live on 27 October 1975, the system revolutionised the control of all rail freight operations across Britain online and in real time.

It used British Rail’s own telephone network and, along with a pair of IBM System 370/168 and a System 370/158 mainframe, Ventek minicomputers with built-in punchcard machines were installed at every area freight terminal. 

In an article published in its 30 October 1973 magazine issue, Computer Weekly described the system as: “One of the most extensive and comprehensive freight management systems in the world.”

Online Tops keeps the freight train rolling (page 1)

Commenting on the 50th anniversary, Jonathan Aylen, a Tops specialist at the University of Manchester, said: “Tops covered the whole of the UK, every freight wagon and loco, every train movement and every cargo all monitored by a central computer system at Marylebone in London. The control headquarters was described as ‘space age’ for its day.” 

Marylebone housed 32 IBM 3330/33301 Control Data Drives, providing a whooping 3.2Mbytes of storage. 

The Computer Weekly article reported that Tops divided the country into 152 Tops Responsibility Areas (TRAs), each with a Ventek 9200 minicomputer system. Describing how the system operated, the Computer Weekly article noted: “The basis for Tops in the field is the punch card, one card for one wagon. As traffic is moved from one TRA to another, new cards showing the changed status are produced. The receiving officer at an Area Freight Centre (AFC) checks the cards against wagons and feeds this information into the system to update the database.”

As Aylen noted in The convergence of computing and telecommunications: Cold War to coal trains paper he co-authored in 2024, the network British Rail used was state of the art for the time, using co-axial cables for the main trunk line links, with 4Mhz analogue transmission (equivalent to 960 telephone channels). Local links were carried on 12 channel frequency division multiplex systems on balanced copper cable pairs. These gave a high-quality speech path of 4Khz channels to every major station, office complex and freight yard in the country.

But it also stretched overseas. Computer Weekly reported that Ventek minicomputers were also deployed in France at Dunkirk, and there was a Telex link in Zeebruggee, Belgium. Data from the freight centres was fed into the Marylebone computer room via a data network which spanned 400km.

Online Tops keeps the freight train rolling (page 2)

“Today we take computer control of business operations for granted. But fifty years ago, it was a revolution to know what was happening right across the business in real time, especially a system which saw up to 3,000 freight train movements daily,” Aylen said. 

Tops highlighted the actual demand for wagons, which meant that many were surplus and sold for scrap, therefore reducing maintenance costs and helping to pay for the overall system.

Bob Gwynne, rail expert, said: “This had one major unforeseen consequence: one scrapyard in South Wales concentrated on quick and easy wagon disposal. Difficult steam locos were set aside for later. This breathing space led to 213 steam locomotives being acquired for preservation, approximately two-thirds of the total steam locomotives preserved today.

“So, an unintended effect of Tops was the birth of the heritage railways as a nationwide tourism product, rather than the handful of locations that pre-date 1975.”

Online Tops keeps the freight train rolling (page 3)

Many thanks to The National Museum of Computing (TNMOC) for providing scans of the original Computer Weekly article. The full Computer Weekly archive dating back to September 1966 is held at TNMOC.

Read more articles about distributed computing

  • Building distributed systems requires effective developer teams: Technologies edge computing and containerisation call for teams capable of building distributed systems.
  • Key features of a distributed file system: Distributed file systems enable users to access file data that is spread across multiple storage servers. A DFS should uphold data integrity and be secure and scalable.

Read more on Datacentre systems management