I recently asked a new hire how they ended up considering WANdisco. “The intriguing, memorable name” was the paraphrased answer. Frankly, when I first heard the name myself some years ago I didn’t think much of it. It’s certainly different, but struck me as too different, too weird for a company peddling Subversion support contracts.
That’s before I understood what WANdisco is about.
WANdisco lived at the edge of my radar as a bit player in the SCM space during my time as Director of Product Technology at Perforce Software. There were always little companies scraping out a living around open source software, so we lumped WANdisco in with CollabNet and called it a day. Wrong!
Turns out WANdisco founders David Richards, Jim Campigli and Dr. Yeturu Aahlad had a much more ambitious plan from the start. The core technology, a patented WAN-capable Paxos implementation, is a key enabler for the evolution of software into globally distributed, highly available systems: in other words Wide Area Network distributed computing, or WANdisco. Much like the name “Microsoft” was about the revolution of software for microcomputers in the 1980’s, the name WANdisco is about the new revolution of WAN-based distributed computing.
What’s so important about the WAN?
We’ve all heard about the globalization of the world economy. Every globally relevant company is now highly dependent on highly available software, and that software needs to be equally global. However, most systems that these companies rely on were architected with a single machine in mind. These machines were accessed over a LAN (local area network) by mostly co-located teams.
All that changed, starting in the 1990’s with widespread adoption of outsourcing. The WAN computing revolution had begun in earnest.
But there was a problem. The WAN wasn’t like a bigger LAN. It’s a different environment all together. Single machine systems like those involving a central server perform pitifully for remote workers. And it’s not easy to update these systems because the WAN is a high failure environment that punishes single machine systems with their inevitable single points of failure. It was time for the science of distributed computing to come to the rescue of this new requirement of wide area network distributed computing.