Content Addressable File Store facts for kids

The Content Addressable File Store (CAFS) was a special computer part made by a company called International Computers Limited (ICL). It was like a super-smart hard drive that could find information very, very fast.

Imagine a huge library of books. A normal computer would have to read every single page to find a specific sentence. CAFS was like having a super-fast librarian who could instantly find the exact book and page you needed. This was important because regular computer brains (processors) couldn't search data as quickly as hard drives could deliver it. CAFS helped solve this problem by doing the searching right where the data was stored.

How CAFS Started

CAFS development began in the late 1960s at ICL's research center. Two researchers, George Coulouris and John Evans, studied how large computer databases worked. They found that if the search part was built directly into the disk, it could make finding information much, much faster.

At first, this special search ability was built right into the disk's reading head. Some early versions of CAFS were used by customers in the 1970s. For example, BT used it for their Directory Enquiries service, which helped people find phone numbers.

CAFS Becomes a Product

By 1982, CAFS became a standard feature in ICL's big mainframe computers, like the 2900 series and Series 39. To make it cheaper and faster, the search part was moved into the disk controller. This is the part that manages the hard drive.

Computers could then send a search request to the disk controller. The controller would then find the data very quickly. This feature was first added to ICL's own database language called Querymaster, which worked with the IDMS database. Later, it was also used with the Ingres database on ICL's VME systems.

ICL received a special award for CAFS in 1985. It was called the Queen's Award for Technological Achievement.

Why CAFS Was Tricky

Even though CAFS was very fast, it had some challenges.

• It needed to know exactly how the data was arranged on the disk. This made it hard to use with different database programs.
• Changing database programs to work with CAFS was often very expensive.
• Also, CAFS searched data without knowing if other parts of the computer were already using or changing that data. This could sometimes cause problems with data being correct.

SCAFS: CAFS for Smaller Computers

ICL also made a version of CAFS for its smaller computers, called the DRS minicomputer range. This version was named SCAFS, which stood for "Son of CAFS."

SCAFS worked differently. It used special software on a regular computer chip. It was sold as a "Search Accelerator" for other popular databases like Ingres, Informix, and Oracle. Each database program needed to be changed a little to work with SCAFS.

The technology was even licensed to IBM for use with their DB2 database on the RS/6000 computers.

Why CAFS Disappeared

Eventually, CAFS and SCAFS became less useful.

• Computer processors became much, much faster. They could search data almost as quickly as the disk could deliver it. This removed the main reason for CAFS.
• Computers also got a lot more memory. This meant that many medium-sized databases could be stored entirely in the computer's memory. Searching data in memory is always faster than searching it on a disk.

These changes made CAFS and SCAFS too expensive and unnecessary for most users.

See also

• Content-addressable storage