Fusion-io chairman and chief executive officer, David Flynn (left) with Fusion-io chief scientist, Steve Wozniak (right)
While Steve Wozniak is widely known as the co-founder of Apple, in recent years he has become known for his involvement with Utah-based flash memory vendor, Fusion-io.
At the recent Fusion-io media event in Sydney, it was plain to see that Wozniak is the public face of the company and a media spokesman.
However, his official role with the company is chief scientist, though this position has had a tendency to be overlooked by the public when they come face-to-face with the man.
When probed on what key contributions Wozniak has made to Fusion-io in his chief scientist role, Fusion-io chairman and chief executive officer, David Flynn, says it has been the man’s design philosophy to "simplify" and "use software to manage complexity."
“Steve’s major impact on our design has been a relentless focus on what can you do more with software and how can you simplify the hardware,” he said.
Out the window
Flynn points out that when Wozniak was in high school, what he did for a hobby was to take the schematic design of the chips that would make up mainframes back in those days.
“He would try to figure how to build that with fewer chips, as he understood that’s what determines cost, reliability and performance,” Flynn said.
“He has always had this design philosophy of getting rid of needless complexity.”
Flynn adds that the interesting thing about this is that Wozniak’s mindset has not changed since the days he was building the Apple II and being involved in the actual electrical engineering.
Wozniak was there when the personal computer spelled the demise of the mainframe, and Flynn says that Flash accelerators, similarly, will spell the demise of the storage mainframe.
“If you think about it, Steve has been through this before,” he said.
“He has been through this process of disruption and miniaturisation.”
Flynn shares an anecdote from that time, when Wozniak coined the phrase, “don’t trust a computer that you can’t pick up and throw out the window.”
“We say the same thing at Fusion-io, ‘don’t trust any storage that you can’t pick up and throw out the window,’ he said.
The fundamental principle behind Wozniak’s philosophy is that small, when it competes with large, always wins.
Flynn says that we are already seeing that same disruptive potential now that storage is moving away from the era of mechanical disk drives.
“Those drives were inspired by Thomas Edison’s phonograph, a spinning platter and a needle from 150 years ago, so you could say that storage technologies in today’s datacentre are using a 150 year old storage technique,” he said.
It is only now that manufacturers have been able to miniaturise the technology.
“We are changing from a world needing to move atoms to a world where you can move electrons,” Flynn said.
“If you can move electrons instead of whole atoms, you win.”
Flynn admits that the transition has taken a long time due to the challenges in miniaturising storage and moving electrons as opposed to atoms.
Despite the delay, he says that Wozniak will be presiding over the transition of storage in the same way he did in the transition of mainframe to personal computers.
Flynn makes it clear that Wozniak is not involved in “connecting a wire into a pin on a chip” at Fusion-io, as that is “not what he does anymore.”
“What he does more valuable than that is keep people focused on what really matters, and that is the transformative potential,” he said.
“He is too valuable as a spokesperson around the world to have him spend time focusing on schematic designs and electronic engineering.”
For Fusion-io, the real value in having Wozniak on board is to inspire the design philosophy of software defined storage, where the storage system no longer requires “a big proprietary piece of technology” and instead can be miniaturised.
The other innovation is using software for a lot of the functions that in the past needed specialised hardware.
“The reason Fusion IO has been so successful is that we took a counter intuitive approach in how to use solid state memory to get the maximum benefit,” Flynn said.
“We could have turned flash into a pretend disk drive and gone into existing storage area networks, but that would have been a terrible mistake.”