Dowager Countess telephone



This is the way the world ends
Not with a bang but a whimper.
T.S. Eliot, “The Hollow Men”

Much of the teeth-gnashing in the IT industry today appears to be about competition, but it’s more about structural change as computing enters its third age. Like the passing of the Edwardian era in “Downton Abbey,” the client-server/Web era is giving way — kicking and screaming — to the rise of distributed systems. We see this everywhere, from the infrastructure models of Google, Facebook and Amazon to the applications and programming languages powering mobile, social and cloud.

Computing eras seem to peak every 30 years. Like old rock stars, they never go away completely and are frequently brought back for gigs.

Begun in the 1950s, the first empire was the mainframe era, and its first king, the IBM 360, came to the throne in the late ’60s/early ’70s. It could run a host of applications, and provided a range of tools for programmers and operators.

This gave way to the feudal era of client-server, which creates dozens and dozens of powerful tech companies across the stack. From its beginnings in Xerox PARC, a long list of new entrants emerged in the ’70s and ’80s. The birth of the World Wide Web in the ’90s — democratized by Marc Andreessen’s Mosaic browser and its commercial child, Netscape — took the client-server architecture further than anyone expected.

When new stars are formed, they originate in the collapse of enormous clouds of gases in space, called star nurseries. We are now swirling in the computing nursery of the distributed-systems era. This is a computing architecture that challenges the hierarchies or control points of the prior era. It’s more open, distributed, and virtual than client-server. This is a flatter software-centric world: APIs, programming languages and open source are magnitudes more efficient.

Moreover, this is the era of the sharing economy, where hardware and software (e.g., Salesforce, Workday) can be rented and consumed as needed versus owned. Enormous computing resources can be rented for a tiny faction of the cost of buying a range of hardware. Platforms are everywhere.

The IT industry needed a change. It has been growing at a low-single-decimal rate for the past decade (according to Gartner).Take out social, mobile and cloud, and the rest of IT feels like the old growth economy.

The apps and the infrastructure are going to change, albeit not overnight. The flexibility and speed of creating new apps — i.e., supporting new businesses or business processes — is going to drive the inevitable shift. For some new players this is the new gold rush. Lots of new entrants (including my own company) are hustling to become the Levi Strauss, Wells Fargo Studebakers (wagon makers) and Pinkertons of this era. Companies like Amazon Web Services and Rackspace are offering to give you an entire frontier town in which to set up shop.

For the distributed-systems computing era to take off, it will need the same kinds of capabilities found in the prior wave. Innovation can only succeed if you can reduce or eliminate the tradeoffs in:

  • Speed
  • Reliability
  • Performance
  • Privacy
  • Security

Want to figure out the next generation of winners and losers? Figure out who will solve these five issues.

Many of the client-server-era companies are not going to make the transition in their current forms. It’s like Michael Jordan playing baseball. Sure, he could get into a uniform, hit a few balls, and catch a pop-up, but it was not his natural sport. IBM made the transition from mainframe to client-server by transforming itself into a services-and-software company.

If you are in IT at the dawn of this third era, you don’t want to seem like the Dowager Countess of “Downton Abbey,” who, upon seeing the telephone for the first time, asked: “Is this an instrument of communication or torture?”

Alan Cohen is the chief commercial officer of Illumio. His last two startups, Airespace and Nicira, were acquired, respectively, by Cisco and VMware. He also serves as an adviser and board member to several technology companies. Reach him @ascohen.


In a classic security model, the five points above can be reduced to three:

Confidentiality (encompassing the Privacy and Security points above)

Integrity (encompassing the Reliability point above)

Availability (encompassing the Speed and Performance points above)

This is known as the CIA triad and traditionally, you can pick two of the three, but you can't have all three.  So if you picked high integrity and high availability, you would need to compromise on the confidentiality.

It would be very interesting to understand the proposition of someone being able to address all three points.  Looking forward to it!


With respect, my first thought is: we're definitely in some kind of a tech bubble. Every single tech bubble I have experienced is marked by this kind of article, replete with mismatched metaphor and analogy, written in a formless hermetic language, and riven with wide, sweeping, unsupported assumptions.

The mainframe IBM era was a kingdom, replaced by a feudal system based on client-server, now under threat as even the Edwardian world portrayed in a somewhat inaccurate BBC drama was placed under threat by, apparently, the telephone. This latest change being, also apparently, somewhat similar to the birth of star systems. 

Goodness. Really? Well, that certainly has helped to make everything clear. 

To paraphrase _The Graduate_ I have one word for you: Decaf.

Umm, the kingdom thingy, that really is basically feudalism -- I mean in the real world.

"Take out social, mobile and cloud, and the rest of IT feels like the old growth economy."

Well, yes, and if you remove the engine, gearbox and drivetrain from an automobile you have something not dissimilar to a horse-pulled wagon. 

"Companies like Amazon Web Services and Rackspace are offering to give you an entire frontier town in which to set up shop."

Well, not really. They are actually offering you servers. To be used by clients. It's just a much more sensible pricing strategy, and there is a bit better sharing of expertise via AMIs.

I could go on, but it is really rather pointless.

I would suggest that if there is a key word for technology in 2014, it is "fragmentation". There isn't a single dominant pattern out there, there are about half a dozen patterns at least. It isn't a time for frantic declarations of fealty to some imagined new paradigm. It's time to think and reflect, much more Prufrock than Hollow Men.


Get every new post delivered to your Inbox.

Join 300,246 other followers