1992 - 97   AR&T

From the point of view of the emerging world of networked digital equipment, computing chips were going to be in just about everything that consumes electricity.  And, an unknowable variety of things were going to want to be connected.  Boeing, like the world, was becoming digitized and networked.

The key to building a network that could provide a unifying utility for a sprawling corporate behemoth like Boeing was the directory.  The things I had learned on P-3 made it clear that whatever it was that defined people and resources in the connected digital world, that thing or service would control everything that is layered on top of it.  Every application and every bit of automation needed to have its existence registered, and it needed to be safely discoverable by the right other nodes on the network.  This would be required if an unknowable variety of things were to be able to be easily integrated based on whatever bit of connectivity someone might discover that they needed to accompliosh.  The control of identity was everything.  But what is identity?

It would not become clear to me until over a decade later that all identity is context dependent.  In those early days of networked computing, we did not yet have a good grip on what a identity was about, and what network directory need to do, let alone what its technical specifications would look like.  The trick was to lay the right evolutionary foundation, so that a process of discovery and change could take it where it needed to go.  Some people thought that one of the then surviving remnants of the badly misguided Open Systems Standards effort, a specification called X.500, and its close cousin X.509 certificates were the way to go.  But to me, that was deeply suspicious.  X.500 had a horrifically awful surface addressing presentation, and it was rigidly hierarchical.  Both of those would have to change or X.500 would die the same way the OSI network protocols had.  Usually, the best seed to plant to start a successful evolution is something simple.  One thing for sure, all of the early PC-LAN systems were super simple when it came to their directories, except one: Banyon Vines.

Vines was an interesting product.  It was horribly complicated and not really very usable.  But, the engineers who had built it were some of the more thoughtful folks around.  Beneath its unwieldy surface it had an elegant beauty about it.  Once again, it was Walt Pietrowski who had clued me into this.  The Banyon engineers had figured out how to build an operating system kernel that could spread computing tasks across multiple processor cores.  And their directory called "Street Talk" seemed to be highly adaptive to working with different networked data structures.  These data structures were as different as digital resources like memory addresses in a host computer and user accounts.  The Banyon engineers seemed to get that.  Their work needed to be watched. 


This point about the people who worked at Banyon  was important.  All of the directional choices for how to go about building Boeing a unified network at the functional level was going to be mostly about how various potential supplier engineering teams thought and worked together.  What the people were doing and how they went about their work was vastly more important than the current state of their shipping products.  The world was changing fast.  Being adaptive was critical for anything and anyone who was going to not only survive, but thrive.

I needed to find a new network operating system, and there was one obvious place to go sniffing around: Microsoft.