Distributed Intelligence

What is Distributed Intelligence?

Distributed Intelligence is the opposite to centrally processed intelligence.

Computer terminal equipment which is fully functional even when the central unit to which it is attached is not available is an example of a distributed intelligence system. A failure of centralised computing may be found in an airport check-in or when on the phone to the bank and we are told that there is a problem with the computer or the communication lines so that they are unable to check us in or give us the information we need.

Distributed Intelligence would in this case be attained by giving each of the parties involved a computer containing the full database, enabling the system to continue to work regardless of what the others were doing.

If this concept is applied to the security industry and in particular to access control, a fully distributed intelligence system is one where not only is the access control decision made at the point of entry but also where the entire system database is stored. This means that at no time does the system require help from a higher level unit or computer in order to make the correct access decisions speedily from its local database.

In the case of the IET networking access systems, the availability of this multiple copy database is taken one stage further. In the event that an Access Control Unit (ACU) loses part of its data due to power failure, maintenance or electrical disturbance, because every ACU on the system is a master (no slaves), any ACU will automatically restore the current database from other units on the system. This capability further enhances system reliability.

What is NOT Distributed Intelligence ?
This is when the full database is not available at each ACU.

Based on the above, let us suppose that you have a site to protect. You want the best for your client and you propose a system that has fully distributed intelligence. The site has 7 perimeter doors or barriers. The system that you have selected has controllers that each have the capacity to control 8 card readers and is ideally suited to the site layout. As a result one these controllers is designated to protect all the card readers at the perimeter access points.

This means that one ACU now controls all of the perimeter access points and you have in fact created a centrally processed system for the site perimeter! This means that you have provided your client with a system where full distributed intelligence has been removed. What you wanted was a system so designed that should any processing component fail, as few access points as possible would be affected. With this proposed system, if the ACU handling the perimeter access points fails, you have jeopardized the entire perimeter of the site.

This example begs the question of how many card readers a controller may handle and still be considered as a fully distributed intelligence system. The ideal situation would be of course, to have a controller containing the full database at each individual access point. Of course, in reality this is not very practical and would prove to be extremely expensive.

ACU’s with the ability to control a maximum of 4 doors are a good compromise, however great care should be taken when proposing a system where maximum security is required.

Be careful. Please note that some systems which purport to have distributed intelligence do so because the processing is not done by a host PC or the local ACU, but by another remote unit. These low budget systems invariably fail to provide sufficient security and reliability.