On Premises
“It is said that the future is always born in pain. The history of war is the history of pain. If we are wise, what is born of that pain matures into the promise of a better world, because we learn that we can no longer afford the mistakes of the past.“ -G’kar: Ambassador, Political Exile, Prophet, Man dressed as an Alien on TV
They also say that everything comes from something, nothing comes from nothing, and that nothing goes nowhere.
As you try to politely nod and walk away from all this sooth, you are grabbed by the sleeve and also told that there’s no such thing as a bad pizza.
Everything comes from Something
The origins of the cloud
You don’t know the tune, but it sounds awfully familiar; a rising and falling of pitch and tone. It at once feels agitating and relaxing. You realize that your exhausted mind has hallucinated a symphony out of a Nagios alert.
It’s the 9th time this week. The batch processing job has hung again. The last delivery date your department got for new servers was still weeks out, and then a week after that for “expedited” racking, stacking, and provisioning.
This time you hope the provisioning team installs the right OS on the machine, you’ve gotten good at bootstrapping from 5 different flavors of linux into whatever the security team are saying is the flavor of the year, but you’d rather not.
Nevermind the networking team.
That doesn’t matter, sleep is calling. You begin prepping the usual job-massaging scripts to clear up the backlog and hope to get things going again.
You wake up not knowing when you fell asleep, or what transpired between your fingers and the terminal, nevermind if it worked or not.
Enough is enough. You call in sick that day. Word was that some upstart bookseller is making it super easy to provision whatever machine you want within seconds, not weeks. Certainly it was bullshit. The name sure as hell was. What the hell is a “Cloud” anyway?
You enter your company credit card information, log on to the console, and join a horde of others across the world engaging in what would become a tsunami of successful skunkworks projects.
Nothing comes from nothing
The fortunes of the cloud
By some conscious prank of fate it turned out the mainframe folks were right, and the future of business computing was not the microcomputer frenzy, but good old fashioned time sharing.
It would have been obvious to those who had the context. Of course you want specialists managing everything. Of course you want to drive efficiency by centralizing compute. Of course you want to control interactions with compute via established APIs.
Of course, of course, of course.
Once the obvious was known, it spread. But spread from where?
As the skunkworks became work work, enough critical mass was reached for Gartner and the like to begin prophesying yesterday’s news. Stock photographers everywhere began wondering why they were suddenly receiving royalty checks for random pictures of the sky as every C-level focused magazine in the world ran covers relaying the Good Word about the Cloud.
When the time came to count beans and break piggy banks, it wasn’t just discussion of cooling and hardware refreshes and colocation strategies echoing in boardrooms and double-duty closets, but also of freeing up large wads of cash in favor of renting everything forever.
In large part one replaced the other.
The bounty of the Cloud, the absolute reversal of the dotcom bust, was directly extracted from the inadequacy of compute infrastructure at businesses across the globe.
In those early days many pennies collected by AWS were formerly a dollar spent on self-hosting.
Nothing Goes Nowhere
Legacy still exists
What accelerated the growth of the cloud(and maintains its dominance) is specialization. Most companies are trying to sell you something. They’ll design, manufacture, and distribute a titanium wallet, then write marketing materials as to why you want to buy it.
Google designs, manufactures, and provides compute, then they write papers as to why you’ll never solve computing as well as they have.
On occasion they’ll hire your best people to keep it that way.
There are those who haven’t made the jump, who have continued to recuse themselves of the cloud. The wisdom of this depends on many factors, and it may be wise indeed, however we’ve had very few upstart booksellers or acquisition-fueled search engines disrupt an entire sector of compute in recent years, so a safe wager would be one in favor of most of these companies wasting time and energy on something not core to their business at the expense of something that is.
In other words, most of these companies are still terrible at I.T.
That never-ending staccato of Pagerduty née Nagios and turnaround times and other infrastructure woes still exist at those companies. These organizations exist as nature preserves (decreed either by local politics or regulatory bodies) for legacy (whose breeding habits are inadequately studied), and the Cloud desperately wants the money tied up in these last bastions.
There is a danger here.
The success of the Cloud has been driven by many things, but one key driver is the continued refining and elevating of the science of compute, making an offering out of it, and having customers meet you at that new height.
This is essentially what professionals of all sorts have been doing since the beginning of everything. However there are some who wish to put this age-old pattern to the test.
But why?
Would you call a plumber to fix your leak, and give them a leatherman to do the job?
Would you invite a dentist to your kitchen to crown a tooth, then hand them some tin foil and a knife?
Then why invite Google into your company and hand them your ailing datacenter?
This is what Anthos is proposing to do in bid for that last slice.
A Bad Pizza
On diminishing returns
There is an unfinished pizza in the living room. When faced with an unfinished pizza, many are inclined to go for the last slice. Upon meeting resistance, some may just tut and leave the party, others might slap a hand or two. One might dip their adhesive-infused bandaged fists into a bucket of crushed glass and call for a kumite.
The temptation of an incomplete pie chart is a lust that has been ingrained in the collective consciousness of everyone trying to make money for the last hundred years.
The fallacy here is that you will almost never be able to actually own the whole pie, and the harder you fight for that last piece, the more opportunity you’ve wasted.
It turns out there was an unopened pizza in the kitchen, safe from the spray of entrails in the living room. As you watch the melee from your safe space next to the stove, you lament that you’re not much for anchovies and pineapple.
Then, as a fifth combatant materializes and unceremoniously brains one of the would-be pizza-gladiators, you ponder that at some point in the past people weren’t much for tomato and cheese either.