I recently saw the movie Apollo 13, again, for the millionth time.
Aside from the obvious, "they lived to tell the tale" storyline, something else stood out; the complexity of it all.
Millions of parts.
Thousands of people.
Lots of unpredictable events.
And all of it held together by a highly orchestrated command-and-control system.
For the slide rule and mechanical meters era, it was an amazing triumph.
Watching the launch of the last NASA Space Shuttle, again the notion of complexity came to mind.
Yes, Mission Control was smaller and more streamlined.
And most of the words were scripted.
But it was still millions of parts, thousands of people and unpredictable events.
It was still "complex".
Thinking back at the IT tasks before me, it has been a long while since those first Ethernet networks were snaked below raised flooring.
Can anyone say "vampire taps?" Yet in the days since our industry has migrated from the sneaker net of passing tapes about to LANs and WANs and beyond, things don't seem to be any simpler.
We still have to flash our systems with operating environments, install applications, enable communications, arrange for storage, manage privileges and investigate when parts of our towers of bit babble go belly up.
Wasn't it all supposed to get simpler? On second thought, maybe complexity isn't really the issue.
In some cases we can ignore the intricacies, because there will always be those who can fix it for us.
We already do this for other technical systems such as our automobiles.
We need them to just work and couldn't care less about the details.
It's easy to forget that most of the world relates to IT this way.
We can try to hide complexity behind staff or scripts.
But the bean counters say nay to the extra bodies and trying to hide the peas under the mashed potatoes only works when someone is around to know how to peer under the potatoes and deal with those peas.
So we still need a solution.
We still have to have engineers to handle the Rube Goldberg contraptions that are our IT resources.
This is true regardless of whether we own the servers, services, and applications, or if we rent them, or if we pay to use them by the byte, second, or MIP.
We're getting pretty good at engineering solutions to manage complexity for us.
VMware's vCenter, IBM's BigFix and Hitachi's Virtual Storage Platform are good examples of this.
And it will soon get a whole lot better because of a little thing called "critical mass.
" Prediction - Significant complexity of owning and operating IT infrastructure will become a "solved problem.
" 40% chance in one year.
80% chance in five years.
Let me be clear here...
a "solved problem" is not one where a cleaver utility will performs some of the steps for us.
Going that far is like buying an instruction manual or hiring an assistant.
What I'm talking about is far more encompassing and represents a significant shift in how IT capabilities are sold and serviced.
And it would reach beyond our current efforts to define, architect and build and sell service automation products.
The word to watch for here is orchestration.
The point in time to watch for is when enough of the automation APIs have been defined, enough of the underpinnings have been built, and enough of these are combined to allow IT department to forget the "how" and deal only with the "why.
" We'll call it an SLA but we'll be thinking "so long to technical complexity" and hello to the Jetson's world of pushing a few buttons and creating all the underpinnings of a business in moments.
From that point on everyone will wonder just why so many were paid so well to play with IT toys.
Hmmm...
I wonder if I can still get into the IBEW?
Aside from the obvious, "they lived to tell the tale" storyline, something else stood out; the complexity of it all.
Millions of parts.
Thousands of people.
Lots of unpredictable events.
And all of it held together by a highly orchestrated command-and-control system.
For the slide rule and mechanical meters era, it was an amazing triumph.
Watching the launch of the last NASA Space Shuttle, again the notion of complexity came to mind.
Yes, Mission Control was smaller and more streamlined.
And most of the words were scripted.
But it was still millions of parts, thousands of people and unpredictable events.
It was still "complex".
Thinking back at the IT tasks before me, it has been a long while since those first Ethernet networks were snaked below raised flooring.
Can anyone say "vampire taps?" Yet in the days since our industry has migrated from the sneaker net of passing tapes about to LANs and WANs and beyond, things don't seem to be any simpler.
We still have to flash our systems with operating environments, install applications, enable communications, arrange for storage, manage privileges and investigate when parts of our towers of bit babble go belly up.
Wasn't it all supposed to get simpler? On second thought, maybe complexity isn't really the issue.
In some cases we can ignore the intricacies, because there will always be those who can fix it for us.
We already do this for other technical systems such as our automobiles.
We need them to just work and couldn't care less about the details.
It's easy to forget that most of the world relates to IT this way.
We can try to hide complexity behind staff or scripts.
But the bean counters say nay to the extra bodies and trying to hide the peas under the mashed potatoes only works when someone is around to know how to peer under the potatoes and deal with those peas.
So we still need a solution.
We still have to have engineers to handle the Rube Goldberg contraptions that are our IT resources.
This is true regardless of whether we own the servers, services, and applications, or if we rent them, or if we pay to use them by the byte, second, or MIP.
We're getting pretty good at engineering solutions to manage complexity for us.
VMware's vCenter, IBM's BigFix and Hitachi's Virtual Storage Platform are good examples of this.
And it will soon get a whole lot better because of a little thing called "critical mass.
" Prediction - Significant complexity of owning and operating IT infrastructure will become a "solved problem.
" 40% chance in one year.
80% chance in five years.
Let me be clear here...
a "solved problem" is not one where a cleaver utility will performs some of the steps for us.
Going that far is like buying an instruction manual or hiring an assistant.
What I'm talking about is far more encompassing and represents a significant shift in how IT capabilities are sold and serviced.
And it would reach beyond our current efforts to define, architect and build and sell service automation products.
The word to watch for here is orchestration.
The point in time to watch for is when enough of the automation APIs have been defined, enough of the underpinnings have been built, and enough of these are combined to allow IT department to forget the "how" and deal only with the "why.
" We'll call it an SLA but we'll be thinking "so long to technical complexity" and hello to the Jetson's world of pushing a few buttons and creating all the underpinnings of a business in moments.
From that point on everyone will wonder just why so many were paid so well to play with IT toys.
Hmmm...
I wonder if I can still get into the IBEW?