Let Us Consult The Tomes

Long, long ago, in the before-time, when thunder lizards and warlocks roamed the earth, (i.e., a couple of years ago), I would have been impressed by marketing literature and contract language guaranteeing 99.999% availability. I mean, who wouldn’t be impressed at 5 minutes of downtime each year?

But five-9’s doesn’t cut it anymore. And neither does a 1 followed by 2 zeroes. Fact is, daytime users don’t care if the network, server, or [insert infrastructure type here], is available at night. A lot of dedicated people work awfully hard to keep those green lights blinking, but good availability shouldn’t come as a surprise.

Users care only about whether they can do their job at least as effectively as they are accustomed to doing it. Can they place an order, access an application, download a key file, view a video conference, place a VoIP call, etc. as effectively as they normally can? Yes? Super… but availability is still irrelevant to them. No? Not good. But don’t unsheathe your trusty availability metrics. You’re likely to be brow-beaten back to the era when people had only one brow (i.e., last week, for some of us who don’t own tweezers.)

Network gurus have passed down from generation to generation the ancient work, preserved forever, on lambskin punch-card parchment, the parable of the performance-technology gap. Let us consult the tomes.

The Guru Amnon was angry, as his field of network equipment would often fail (as it tended to do in those early days,) and his work would be undone through no fault of his own.

So Amnon left his network field and headed to the Silicon Mountain, to plead for mercy from the gods.

For three days and three nights, Amnon traveled, (with an overnight layover in the ancient city of Memphis). But he finally pleaded his case before the gods.

“Please, my servers are failing, my network cable is unreliable. I try so hard to work my network fields, so the business can harvest its bounty. But I am undone by the most frustrating of things that are never my fault.”

The gods did take pity on Amnon, but felt that it would be best to give him wisdom instead. So the Gods said to Amnon: “Go, Now, and by the time of the next great Temple Convention, you will have a network that never fails – a masterwork of infinite uptime.”

True to their words, the Gods granted Amnon’s request. By the time of the Temple Conference, on the High Holy Day of Installation, the green lights on his servers glistened like sapphires, blinking on and off in a rhythm harmonious with the universe. And the Guru was pleased, for he thought that he would not have to do so much work to keep the business going.

But it was not long before business unit owner, Casiphia, came forward and said: “What is this, wise Guru Amnon? I see the lights, but my unit cannot place or process orders. I see the green lights promise a harvest of luscious pairs, but they turn to ash in our mouth when we dare to sup.”

The Network Guru looked at his perfect network, and was confused: “This network was crafted from the gods themselves, and I, the wisest of the Gurus, maintain it! How can this be?”

In his hubris, Guru Amnon realized that he had neglected that a perfect network availability does not guarantee perfect network performance. With this, Guru Amnon was enlightened.

Measuring and reporting about availability may be a comfortable way of quantifying how a network behaves and it may be all you can squeeze out of a provider. (And yes, it is necessary… but not sufficient.) But it’s no way to measure user-effectiveness. And it’s certainly no way to support the business. The business runs on applications. Those applications run across a complex array of routers, servers, and switches. So skipping a performance first approach while managing just the underlying infrastructure is a sure fire way to find yourself foiled by your own hubris.