“Everybody lies.” [Dr. Greg House]
“If you tell the truth, you don’t have to remember anything.” [Mark Twain]
I’m going to start a series on this blog – which I’ll eventually turn into a white paper – called “everybody lies: backup.” This is a brief introduction to the series.
There are two reasons that exaggeration and outright misstatement are so common by backup vendors. The first is a desire to sell. Look – everyone has that desire – and buyers are rightfully aware of that desire and tend to be skeptical about statements made by vendors – which in turn leads to “turning the volume up” even further in an attempt to sell – and the cycle continues. This of course applies not only to backup vendors but quite frankly to much of the way things work throughout the world’s economy.
The second reason is more specific to backup. When people talk about running Windows Server 2008 R2, they don’t typically ask how fast it is on an absolute basis. They may ask how fast it is versus Windows Server 2003, and they may ask how fast it is versus RedHat Enterprise Linux on some very specific set of tasks. But there’s a general understanding that the speed of Windows Server 2008 R2 is based on the server upon which it is running. The characteristics of that server – CPUs, memory, disk, networking – any external storage associated with that server – and of course the application load on top of the operating system – are fundamental factors in determing how fast it is.
Most backup vendors sell software. So when they’re asked about performance, they make assumptions about the underlying hardware to put their software in the best possible light. That’s the reason that very intensive operations like deduplication are so difficult to size. Other vendors of secondary storage (vendors like Data Domain, Exagrid, and the like) are in an even more isolated position – they have to make assumptions about not only the backup server but the backup software itself.
So a trivial question like “How fast can I backup?” is very, very difficult to answer. And thus the most optimistic possible answer is given – or even answers that stretch the bounds of the truth beyond human recognition.
This series is a tribute to truth that has been stretched to a degree in which it is no longer recognizable. And thus the title: “everybody lies.”