Instalment 1
A Matter of Trust
The Wolds data communications open up with the introduction of the Internet. Well before that, data connections, in terms of exclusive data links, were outrageously expensive.
The location of huge Data Centres was often confined to house as many directly connected users. The notion of placing a secure data centre and have the bulk of the users housed in a far distance place; meant the pure cost of dedicated data connections was the overriding factor.
Before the Internet, all manner of input devices from workstations to Terminals or PC’s needed to gain access to a data centre, were not automatically trusted to permit access.
This was seen as a good thing! It provided a totally secure network because any input device that was connected or wanted to gain access; was not trusted by default!
Trust was only granted to input devices so listed by the Data Centre’s Communications front end that held all security access tables of trust.
There was the marriage of the input devices needing to be trusted as well as the input device itself, granting trust only to users who’s sign on credentials was given by the input device.
This was the standard approach of security for many years. The input device only trusted certain users sign on credentials and in turn the input device was trusted only by the set of security trust tables held by the communications front end; to then be trusted by the data centre processing itself.
The 3rd level of trust was afforded to a few input devices, which trusted only certain users sign on credentials, for the maintenance of all data security trust tables; where trust itself was created, modified or changed.
At a programmers level of sign on, input devices and user credentials were still vetted by the communications’ data security trust tables of the front end.
There was a fourth and final level of trust which differed slightly on this model – these input devices and user credentials did not have to pass through the communications front end.
There were a number of input devices at the Data Centre that held the highest level of rust. These input devices were given many different names but lets just call them Prime Console Input Devices. - Sorry no acronym I am aware of.
A Prime Console did not connect to the processors via a communications front end, where all the security tables were held.
A Prime Console connected directly to the Processors and thus trust was unlimited by such an input device.
User credentials themselves were allocated trust by hard coding within the operating systems. There were few such Prime Input devices and few user sign on credentials.
This was done not so much as to guarantee trust but to ensure that no matter what; a users sign on credentials plus the Prime Console input devices were both needed and guaranteed to have limitless power over all functions of the Processors.
It was at this level that Operating Systems were able to directly effected not so much as the programs; I’m talking about being able to either terminate a rouge process or shut all processors down themselves.
Most of these data centres had remained online, without a pause, for over three decades and often far longer.
The time in which a data centre could be shut down, as far as its Operating System and power down, was NOT a short time. This would often take many days and was highly structured and the stuff of Reference Manuals was needed.
Most of the minders and keepers and Data Base Analysts of the Operating System would never see a data centre shut down during their working life. Such was the rarity, often the manufacturer of the hardware needed to be consulted to get this correct.
With todays Transactional Tracking Systems and Transactional Processing the DBA was king. Terminating a rouge process and moreover, shutting down a data centre was a long audacious task. Without today’s superior TTS and TP, every process needed to be shut down in a very specific order and left closed. If this was not done the processor would come online with its applications looking like scrambled eggs.
As an aside the value of UPS was no luxury.
I’ll give you and example of a very cut down type of total system shut-down procedure, a far far cry from a clustered bank of blade servers which carry Operating Systems, Applications and Communications Access – If this was not important, I would have left it out!
Create and Verify user authentication was a Binary Value as Hexadecimal values were not included into the Operating Systems and Front end control. Hexadecimal expressions were a neat interpretation of binary values which added an unnecessary translation at the time. The size of files was the biggest limiting value as hard drive capacities were only say 20-40 Megabyte in size.
The neat representation of hexadecimal values added an uncomfortable amount of file space and was a luxury at best. Coding was done via huge but very compactable files written in assembler! The only way to shorted tedious coding was resolved by symbolic names and all up these huge files would generate down into a very small executable files. The biggest issue for the coding was the absolute file length that was one of the system endpoint of the text editors.
>Dismount All
>>Input create user authority codes
>>Input verify user authority codes
>Broadcast message to all Prime Consoles: All online storage will be dismounted ! Close or Terminate all open users front end connections
>>Proceed? y/n/f
>>Halt All
>Input create user authority codes
>Input verify user authority codes
>>Broadcast message to all Prime Consoles: All processors will shut down in -x hours: minutes – Terminate front End Access
>>Proceed y/n/f
>>Broadcast message to all Prime Consoles: All processors will shut down in -x minutes – Terminate front End Access
>> Input create user authority codes
>> Input verify user authority codes
>>Proceed y/n/f
>Broadcast message to all Prime Consoles: All processors will shut down in -x minutes – All front End Access will be deigned!
>> Input create user authority codes
>> Input verify user authority codes
>>Proceed y/n/f
>Broadcast message to all Prime Consoles: All processors will halt and power down in -x minutes
>> Input create user authority codes
>> Input verify user authority codes
>>Proceed y/n/f
>Halt UPS
>> Input create user authority codes
>> Input verify user authority codes
>>Broadcast message to all Prime Consoles: UPS Power terminated!
>>Proceed y/n/f
>>Broadcast message to all Prime Consoles: All processes will halt and power down immediately
>>Proceed y/n/f
>Offline _ _ _
One other reason why shutting down such data centres took so long was to ensure that every single user in the data centre knew, not only in advance, but also to warn everyone that someone was going to pull the plug – Indeed the issue of trust had to be earned in both the logical and physical attributes or every input device.
Circa 2001 Enter current day of say, the year 2001 – as we really don’t want to get in the way of Y2K, 2038 and other system range endpoints! Welcome the PC and the Internet.
…(second Instalment to follow 1 of 6)
