By submitting your email address, you agree to receive emails regarding relevant topic offers from TechTarget and its partners. You can withdraw your consent at any time. Contact TechTarget at 275 Grove Street, Newton, MA.
| Goodbye transactional computing; today, the file is the next big thing.
The transactional computing era is over, my friends. It had a fantastic run, generated gazillions of dollars, created countless jobs and industries, and was intellectually stimulating--for a while. It's why we work in IT. The principles that revolutionized the way business was done have served us well. For 50 years, commercial computing has been predicated on the same basic philosophies, architectures and structure, all of which are based around the transaction. The transaction is a small block I/O that has known value, and thus can be supported with known costs. Protecting and scaling our ability to transact is how all of the equipment in our data centers ultimately came to be. Ninety-nine percent of every computer gizmo in your world is a variant of a modern transactional system. This era won't end, it will just be dwarfed: first by data, then by dollars.
There was a premature celebration of pending upheaval as the distributed computing era set upon us, but it wasn't a revolution. It didn't kill the transactional behemoth, merely added to it. It sparked a class war and enabled the hippies to claim some power back from the "man," but it didn't change things. It did push the concept of computer automation to a new level, sometimes so well that we didn't need humans. Progress sometimes hurts, we found out. This era delivered the all-important file. People talk to computers with files, while computers talk to disks in blocks. The file is bigger, has the potential for self-awareness and needs to be passed around like a Frisbee to be useful. The file led to the network, which enabled usage. Usage creates new opportunities and problems. It's the circle of IT life.
The distributed computing era created as many problems as it solved, but every revolution does that. (Watch what you wish for my Canadian separatist brethren.) The distributed era killed the minicomputer in the same way the automobile killed the buggy business. The file was a new thing, one that we tried our best to make fit into our comfortable, existing world of the transaction. But we haven't truly done it justice. Instead, we built in hardware that was just smaller, dumber, cheaper versions of the big stuff we previously built. And we wondered why our lives kept getting more and more complicated.
There's a new era upon us, like it or not. The Internet computing era is dragging data into the "cloud" right along with it. You already partake in it, whether you realize it or not, and so does your company. This era will rain more files--and bigger files--down on you than you can ever imagine, causing scale dilemmas that will be almost incomprehensible. Think it's hard to deal with distributed computing issues? Just wait until 90% of your data exists in the ether and your applications execute "somewhere." This era won't go away quietly. Like K-Fed, this era is taking over for now.