As I See It: Cloud Cover
May 14, 2012 Victor Rozek
Herb Grosch was the second scientist ever hired by IBM. And he was a good hire. His resume resembled an achievement highlight reel. It included doing calculations for the Manhattan Project, and helping develop the Whirlwind computer at MIT–the first system that actually operated in real time and used video displays for output. He was also the first to formalize the relationship between cost and performance in what has become known as Grosch’s Law: “economy is as the square root of the speed.”
But perhaps his most fascinating insight dates back to the 1950s–more than a half-century ago–when he postulated that the entire world would operate on dumb terminals powered by about 15 large data centers. He was wrong about the dumb terminals, but the rest looks suspiciously like cloud computing.
The concept of cloud computing is nothing new. We used to call it time-sharing when an enterprise with excess capacity sold system time (often at night) to smaller concerns with CPU-hungry programs to run. But in the context of today’s technology, it’s a game changer. From the moment online retailer Amazon noticed that its data centers were frequently operating at a fraction of capacity and launched Amazon Web Services, the traditional data center was headed for the scrapheap. Like Ma Bell, which was hacked into Baby Bells, the in-house mainframe–once the flagship of IT operations–has been splintered into various kinds of commodity servers so plentiful they can be measured in acres.
The data center has gone virtual, and with it so have many of IT’s traditional roles, responsibilities, and relationships. For one thing, hardware has lost the last of its symbolic stature. In the early years, IT operations focused chiefly on keeping the hardware running and available. But as hardware became more reliable, it required less attention and correspondingly declined in importance. Focus shifted elsewhere. Like a modern car with 1,300 computer chips that do heaven-knows-what, we no longer know or care how it works, only that it does. When every 14-year-old has more computing power in his hand than was available in his father’s mainframe, the mystery and majesty of hardware is lost. Hardware became less of a wonder and more of a tool. Software, in all its stunning variety and usefulness, came to dominate the industry and the imaginations of end users.
In that sense, cloud computing is timely. Migrate to the cloud and there is no need for the trophy data center, the chilled glass house boasting rows of enterprise systems, monuments to processing prowess. No need for a data center at all. Managers can cobble together any combination of servers and storage (with free software thrown in) from a vast resource pool available over the network. IT’s unending ordeal has been made manageable: companies can increase capacity and add capabilities on the fly, without a huge investment in infrastructure and employee training. It’s pay-for-play computing. No worries about when to upgrade, which vendor to choose, or what to purchase. With cloud computing, the Great Technology Chase–that perpetual thorn in IT management’s paw–has slowed to a shuffle. At the very least, cloud computing has shifted IT’s focus on outcome rather than process; on what can be accomplished, rather than how and by whom.
Managers, who once were required to know a great deal about hardware, can now turn their attention to more pressing matters: how to keep their networks running while ensuring their employees are not spending half their time on Facebook (incidentally, another variation of cloud computing).
Vendor relationships will also diminish in importance as they, too, become virtual. Reliance on personal contact will shift to mega-data centers in remote locations no one will ever visit, run by people no one will ever meet. Whether the cloud provides utility computing, web services, managed services, or even SaaS applications, for IT managers who are, in my experience, often enamored of control, moving operations to the cloud is the ultimate act of delegation, entrusting the observable to the invisible.
On the bright side, if data center size is important as evidence of management stature, then in the cloud, everybody has a big one. When you’re playing with the likes of Microsoft, Google, Amazon, IBM, Verizon, you’re playing in the resource major leagues. And if you’re competing against them, you’re opposing the Borg. If you manage to show signs of becoming successful, you will be assimilated. Down the road, when the players sort themselves out, Grosch’s prediction of 15 global service providers does not seem farfetched.
But giving up control in return for ease and a bill every month requires a surprising level of trust.
Surprising, because trust in all institutions is at an all-time low. Ron Fournier and Sophie Quinton recently published a revealing article in the National Journal called In Nothing We Trust. It’s about the disillusioned people of Muncie, Indiana. Their jobs are gone, the banksters took their homes, their local government is bankrupt and helpless, state and national governments ignore them. Churches offer stale reassurances, insurance companies gouge them, the educational system is failing their kids. They see corporations make billions and pay no taxes. Seven out of 10 people believe the nation has wandered radically off course. They are angry, and they are desperate because the institutions that once buttressed the middle class have either failed them, or turned against them.
Muncie is far from unique. Reportedly, only 19 percent of us have confidence in big business. And yet, cloud computing is really the creation of a new institution, a storehouse for the nation’s data, controlled by the biggest of the big. We can only conclude that people trust technology much more than they trust each other. The flaw in that preference is self-evident: people make and manage the technology. If information is power, then putting so much power in the hands of so few is an open invitation to abuse.
If cloud computing has an Achilles heel, it is a growing concern about security and privacy. At present, small companies are better served by the cloud because they can obtain sophisticated services at pedestrian prices. Fortune 500 companies have already invested considerable sums in IT infrastructure, and their proprietary information is as tightly held as a piece of the true cross. Nonetheless, bit by bit, non-mission critical applications are migrating to the cloud. Once the security concerns are resolved (a process that may begin with uploading only encrypted data), other applications will follow.
In 414 BCE, The Birds, a comedy by the Greek playwright Aristophanes, had its debut. The main character, Pisthetaerus, is tired of quarreling Athenians and capricious gods. He conspires with the birds to build a city in the sky which they call (I swear I’m not making this up), “cloudcuckooland.” Their purpose is to capture the smoke from sacrifices offered to the gods, thereby gaining control of all communication between what we would call the 1 percent above and the 99 percent below. Everything important would be trapped in the cloud. The theory was that the gods needed people they could manipulate for their amusement, and men needed the favor of the gods. Control communication, and you control them both.
That’s the problem with cloud computing. There’s no way of knowing if Pisthetaerus is going to turn out to be a good guy, or just another opportunist.