Progress-- You Say You Want a Revolution

By: Greg O'Connor

Greg O'Connor is the CEO of AppZero and a guest contributor for the the Business Making Progress blog. Pioneering the Virtual Application Appliance approach to simplifying application-lifecycle management, Greg O'Connor is responsible for translating Appzero's vision into strategic business objectives and financial results.

Here’s a revolution: AppZero protects your software from your customers.

As a proud Progress alumni approaching Revolution, I am struck by the irony of how very much I could have used the technology I now bring to my fellow software executives, who are struggling to balance revolution and cost.

It all began in a little log cabin … Progress and Sonic 2000

In 2000, I had the opportunity to gather some of the best and brightest people as I co-founded Sonic Software with Bill Cullen, Sonic CTO (and product brain)  At the time, we saw a market-making opportunity to take the AppServer world standards (formal/XML or market driven/Java) and apply them to the EAI market.

Thanks to the afore-mentioned bright folks, help from Roy Schulte (a Gartner fellow at the time), a bit of luck, and an excellent Messaging product, we created and evangelized the ESB market category.  The first ESB to market -- Sonic XQ (Xml Queue) -- was shipped February 2002.  10 eventful years later, I’m here to share what Bill Cullen, AppZero CTO, and I are up to – and how our Sonic experience shaped it.

If you sell software, you’ll appreciate this observation

Growing Sonic Software, we faced two universal hurdles that significantly impacted our business – and that of pretty much everyone who sells software:

1. Winning or losing – labor-intensive demos, proof of concepts (POC), evaluations, and trials had a huge impact on our growth rate
2. Installs that did not go perfectly, resulted in fire drills, lost business,  and a sharp dip in customer confidence

(These facts of software life are some of the acute pain points we solve here at AppZero.)

At Sonic, we were often faced with a 5 day evaluation for a prospect:  1 day to setup our software on their environment, 3 days to do the work they requested, and 1 day show off the results.  (IBM, Tibco, BEA and Webmethods   had more or less the same timeframe but seemed to send a heck of a lot more people to the evaluation.)

When the 1^st day did not go as planned, we always lost.  Always.  Every single time. No exception.

A cautionary tale:  If you sell software, you are guilty until you prove yourself innocent

Oh, and here’s how I learned that an imperfect install can still bite you long after you have successfully fought to win a customer (in this case a market icon).  A full year after having won the business and implemented our product at the New York Mercantile Exchange, I received a call from the CIO.  He had some new concerns, “Sonic messaging system appears to slow down under load”.

Arrgggh.  How could this be possible?  Sonic was ahead of its time with elastic scaling, continuous availability, and best in class through put.  This could not be correct.  As it turned out, it wasn’t.

But determining and fixing the “root cause”  took 6 labor-intensive weeks filled with tons of anxious phone calls, numerous pointing fingers (with chewed fingernails), and a couple of flights to NYC by our top troubleshooter .  Life got very unpleasant before it returned to good.

The culprit? A bug in the Java Virtual Machine (JVM) and Java Runtime Environment  (JRE) that would not do garbage collect (free memory) under load.  Now, long before that fateful phone call, we at Sonic knew all about this issue.  We had documented it, changed our install and packaging to make an easy fix.

(Cue scary music) But then the customer got involved.

Someone, somewhere along their line had installed their company’s “certified” version of the JVM/JRE thereby putting our product and reputation at risk.

“It wasn’t my fault” just doesn’t matter.  It took a long time, involving many smart people to find the 2 files that needed to be changed so that all the oil futures in the world could once again flow over the Sonic messaging system.

Morale of the story: Once a customer has your software, things happen.

If I had a time machine, I would bring the AppZero product to my(then)self

AppZero not only solves the PoC puzzle for software vendors, but protects their Windows and Linux server applications from customers.  We make it possible for applications to be pre-installed, pre-configured and then provisioned onto a physical or virtual OS -- in minutes, perhaps over lunch.

This capability effectively changes the math around POCs in a big way: we reduce the install, setup and configuration time to zero.  If I had been able to use AppZero at Sonic, I would have freed up a whole day to actually do the customer requests on every single PoC.   What would a 33% increase in productive time have meant?  I’m going to guess a higher win rate against the competition, faster company growth, bigger promotions, and more time spent with the wife and kids.

And if I had had AppZero at Sonic, our very cool software would have been safely isolated from the customer’s operating environment instead of deeply enmeshed in it.  Innocent from the start.  Hey, how’s this for a new tag line? “AppZero -- protect your software from your customers.”  (www.appzero.com)

For more fun stories and a look at another market-making technology, stop by our booth at Revolution.

Greg O’Connor, AppZero CEO grego@appzero.com