After I had worked on the CLARiiON RAID algorithms for a few years, I had the itch to move on to something different. The CLARiiON write cache provided that opportunity. I was still writing code but I was also operating as project lead. It was a good experience that had me interfacing with many of the different groups involved with rolling a new product out: diag & test, QA, manufacturing, HW design, marketing, education, documentation, sales, etc.
This activity brought me right up to 1994. I had started working on the CLARiiON microcode (FLARE) in 1987. During the stretch run of trying to ship the write cache I was starting to think ahead. What did I want to work on next?
There was one person in particular that helped me with that decision.
President Bill Clinton.
Bill Clinton was elected President of the United States in November 1992. One of his campaign promises was the ability for new fathers to take a three month (unpaid) leave of absence when a child was born. True to form, the Family and Medical Leave of 1993 was signed into law in February of 1993. Medical benefits would be covered by employers during a leave.
Being that my wife and I were hoping to have child #2 sometime in 1994, we started socking money away to allow me to stay home for twelve weeks. And when she became pregnant in the summer of 1993, I let my boss know that I was going to take advantage of the new law. We started training a successor to take over the project leadership of the write cache.
In March of 1994, the FLARE write cache team was right in the middle of what can only be called a grind. We were working all hours trying to shake the kinks out of the code. One of my co-workers gave me the news: “Hey your wife just called”. And on March 9th of 1994 I walked right out of the middle of that project and did not physically return to work until June. My son was born….
and I’ve always loved him for it!
Perspective
My separation from Data General (my employer) during that time was full and complete. I didn’t call in, didn’t talk to anybody, and to be quite honest I didn’t even wonder how it was going. I had better things going on. And let me tell you that my wife and I got a lot more sleep and spent a lot of quality time with my daughter, who had just turned two years old.
As the calendar turned to June I faced the reality of my return to work. I found that I couldn’t stomach the thought of returning to what I had been doing previously. When I did return, I found (of course!) that the team had done quite well without me. I needed to figure out what to do next. My time off had given me some perspective about what I needed and what I liked to do. What I learned:
- I enjoy the process of building software
- I need to be continually learning
- I need variety
So I made a decision to pursue the blank sheet of paper.
A Clean Slate
When I was recruited to the FLARE team my potential boss told me “some people go their whole career without getting to draw on a blank sheet of paper. The RAID algorithms are a blank sheet of paper”. What he meant by that is that at a big company the opportunity to build something from scratch can be hard to find. I enjoyed building the write cache but it wasn’t quite a blank sheet of paper; there were constraints involved based on previously shipping versions of FLARE. So I made up my mind: I’m going to find another blank piece of paper and do something completely different.
Fortunately for me a new group had recently formed within Data General.
CLARiiON Systems Engineering
Data General invested quite heavily in attaching CLARiiONs to a variety of servers and operating systems, including HP, Sun, IBM, SGI, and Microsoft (among others). A new group of software engineers had been formed to take responsibility for any CLARiiON-related software that ran outside of the box. This basically included device drivers, failover software, kernel extensions, and storage management.
I decided to join this group because I felt that there would be an explosion of new work. And quite possibly a blank sheet of paper.
I was right.
The Beginnings of Navisphere
When I joined the Systems Engineering team in July of 1994 DG had already hired a set of contractors to design a graphical user interface (GUI) for CLARiiON. It was eventually called ArrayGUIde. I was given the task of creating CLARiiON’s first command line interface (CLI). It was called CLARCLI. The GUI and CLI shared a common client library called PollLIB. The contractors did a good job. ArrayGUIde and CLARCLI shipped for several years.
But ArrayGUIde wasn’t extensible and it wasn’t scalable. CLARiiON was moving towards building highly scalable, high-disk count, building-block oriented storage systems that were based on a new technology known as Fiber Channel. ArrayGUIde couldn’t be extended to manage both the old SCSI-based systems and the new fiber-channel disk arrays.
So my boss called me into his office and said “Hey Steve, how’d you like to build a brand new storage management architecture that will manage the entire family of CLARiiON products. You’d have a blank sheet of paper.”
The answer was easy.
“I’m all over it”.
We started building Navisphere in the summer of 1995.
Steve
