Recently, on Bill Gates’ suggestion, I picked up ‘The Grid’ by Gretchen Bakke. To borrow Bill’s words: “This book, about our aging electrical grid, fits in one of my favorite genres: “Books About Mundane Stuff That Are Actually Fascinating.”
I often use the term “utility style computing” when talking about our IT-as-a-Service product, and the parallels between the IT industry today and the way our electrical grid developed in the late 1800’s and early 1900’s are striking. To (very) briefly describe the first few chapters in the book, when electrification first became popular, we were limited to DC (direct current, as opposed to today’s more common AC or alternating current). The primary limitation this imposed was range; it was very difficult to transmit DC electricity more than a mile or so. Thus, “private” small generation plants became the defacto standard for electrical power generation. Factories, private estates, and even the occasional municipality would deploy their own stand-alone infrastructure to power their facilities.
During this time, many small power companies emerged, and due to the limited range their plants were idle much of the time (a factory only runs during the day, street lamps only at night, etc). Because there is no cost effective way to store electrical power for later use, the capital investment was huge and the return very limited. Many of the small utilities turned to selling and servicing private plants as a way to make ends meet.
This really reminds me of the way IT service for most SMB’s developed during the 90’s and 00’s. Lots of businesses with servers sitting in closets, idle 96% of the time. Yet that degree of investment was required because there was no shared infrastructure (or grid) comparable to today’s cloud services.
About 30 years pass, and by 1920, AC (alternating current) becomes the norm. Long-distance transmission and voltage conversion become cost-effective. Samuel Insull, Thomas Edison’s long time aid and business manager took control of Chicago Edison, a small generating station downtown. He came to a series of interlocking insights about the utility business:
1) The key to making money as a utility is keeping your infrastructure as close to 100% utilization 24×7 as possible.
2) The lower your prices, the more money you will make.
3) Subscription is ultimately cheaper for the consumer than independence.
Those insights were enough to propel Chicago Edison under Insull’s leadership from bit player to regional monopoly. He realized that federal and state regulation would turn utilities into natural monopolies, meaning competition in those markets would largely disappear due to high infrastructure costs. It remains to be seen if something similar will play out in the cloud space, but with major infrastructure players developing their own custom chips, and what Cory Doctrow (a personal hero) called a “war on general purpose computing,” it wouldn’t be an unsurprising outcome.