At first we started each instance off on a small stream of data. It was a very bespoke system at first and only one of us ahd the unique qualifications to ensure the data delivery system. Yet the bandwidth simply wasn’t there to deliver the data sufficiently. So we replaced the upstream with a third party vendor solution. While this had a greater technical cost at least both parental units could now assist in connecting each instance up to a data stream.
Teething: Original Manual Until It Hurts
Then each instance started to grow, and kept growing. They were hungry for more data and began to add new server clusters to chew through even greater amounts of data. The downstream grew in size and smell.
Yet every time an instance would add a new cluster to the data processing servers, man would the machines whine. I mean scream relentlessly in the night. It was like their fans just kept wheezing and would not stop until the powercore overheated and the damn thing finally shut down.
We tried so many strategies to get over it. We should have listened to our grandmothers and just poured brandy all over the servers.
Teething Strategies
Like many new parents we provided the first instance with a little too much care. Only the best ice chip sets or handcrafted tools to try and bring down the pain of adding to clusters the dta processing center. It was so far back in the company’s history I don’t remember all the details.
I do remember freezing wash cloths and draping them over the emerging server. By the second instance we would even add a bit of juice to the cloth to add additional staying power. By the third instance we just gave up and said, “Here have this popsicle.”
Nothing worked better. There are these tiny little slow melt popsicles. The could quiet down the screams of any instance dealing with additional servers cutting through in the dta processing center. Plus they were cheap and plentiful.
Other Strategies
Frozen fruit in one of those net pockets. Use melon, berries stained all the server casings and even the packaging that wrapped each instance.
Give instances to grandma, grab an hour of sleep.
Cry ourselves
Wonder if we should use brandy and then realized, “Who the hell still drinks Brandy?”
Cried more, slept some.
Emerging from the Pain
Eventually we got all the servers in the stack in place. After about five years they needed replacement. Strangest thing, each instance worked something out with the night crew and would recycle the broken server for a dollar gold coin. There is a lot of inflation for used servers. We do not allow that in our startup.
Eventually an instance needs to fall in the black. Yet we have a runway shorter than most Series A companies begging up and down Sandy Hill for a Series B and legacy costs that will last well into the next decade.
Featured Image by Jess Klein
LOL nice use of my illustration 🙂