11 Reasons That Make You Hate the Cloud
11 Reasons That Make You Hate the Cloud: For software developers and IT professionals, the pay-as-you-go cloud service model sounds like a dream come true. But then the bill comes due.
When the first price lists from cloud companies arrived, everyone reacted like a kid in a small candy store with a $100 bill. There is no reason to devote hours to thinking about the pros and cons when you can just grab a shovel and fill a large laundry bag. Many cloud prices are well below a single penny. How can you go wrong when the price approaches zero?
Those days are long gone. Many of us have already survived dozens of difficult meetings with bean counters who received the monthly cloud bill and set off a DEFCON alarm. The prices may be low but they add up. So we sit around the table and wonder why. Is there some way to get around it? Is there something we can do? Or is there some secret lever we can pull? Some secret phone call to transfer?
Netflix may be the best example of a company that thrives in the cloud. Every Friday and Saturday night, demand starts to spike as people want to relax. Netflix spins up thousands of new instances to handle the load and then gives them back after a few hours when people go to bed. A week contains 168 hours but Netflix only pays for their machines during the few hours their customers need them.
However, not all workloads have such distinct peaks and valleys. If you are going to run your instances 24 hours a day, seven days a week, well, the advantages of the cloud start to fade because the price can be steep even after discounts.
Here are 11 reasons why the rest of us hate the cloud model.
Splitting Costs Is Difficult:
If you buy a server, it is entirely yours and therefore you pay 100% of the cost. If you want to share a server, well, you have to split the resources. Splitting costs equally only works at a restaurant, and even then, the people who did not order alcohol leave annoyed.
Trying to be precise about cloud costs is surprisingly difficult for a commodity product. Should work done at peak hours cost the same as background computation that can be interrupted and deferred? What if utilities charge less for electricity at night or perhaps on sunny days when the solar farm is full of electrons? What if one customer wants a service level agreement while another is willing to be more relaxed about downtime?
Modern Architectures Encourage Consumption:
Mathematicians love to contemplate the Banach-Tarski paradox, a strange theorem known as the “pea and the sun paradox”. It is a simple algorithmic method by which a pea can be cut into a finite number of pieces and then reassembled to become the size of the sun. Of course it would not work in the real world, but it all makes sense in the world of lower topology.
Sometimes the cloud seems to encourage the same game. The hot architectural approach to development is to split your single large application into dozens or perhaps hundreds of different services, each running multiple times as containers manipulated by Kubernetes. The good news is that when the load spikes, Kubernetes steps in and starts buying new instances to satisfy everyone. The bad news is that no human can keep track of how many containers or instances are running simultaneously. Our knowledge of this is literally fuzzy – at least until the bill arrives. What was once a nice small application running on a single server now generates bills as big as the Sun.
Free Can Be Expensive:
We all love free. Giving developers the opportunity to try your product for nothing is the best form of marketing. But at some point, someone needs to pay and it is difficult to manage that massive step from free to paid. It may be just pennies, but the increase on a percentage basis is limitless!
All the free fun makes it difficult for developers to predict real costs. There are plenty of stories about companies that were blindsided by the slings and arrows of outrageous fortune that come from the viral fascination of digital masses. Everyone cheers when content goes viral, but then the bill arrives at the end of the month and it is much larger. It seemed cheap when it was free.
Good Products Are Sticky Sprays:
The cloud may be all about freedom but somehow we get locked in anyway. For example, good developers try to develop their applications on the smallest, cheapest instance. It works fine for a while but then the load spikes when something goes viral or a big shopping holiday arrives. The devops team will boost the RAM or virtual CPUs, and everything will be fine again, but unless the devops team is on top of things, the configurations will stay at the higher-priced settings.
The ratchet effect is more pronounced in the disk space allocated per instance. It is often easy to add more space but hard to reduce it, even if the disk space is unused. There are 23 steps in this special guide to shrinking your volume.
Sometimes Location Matters:
Data centers in the cloud do not have fixed addresses and hardly seem connected to planet Earth. They just hover in some lower regions.
Sometimes, however, location does make a difference to the price. Alibaba, for example, charges more for instances running on Chinese territory. Amazon charges $0.0056 per hour for a t3.nano machine in Ohio but $0.0062 in Northern California. The Azure price index shows that prices in the North Central US can be half those in the South Central US.
Is it because real estate is more expensive in some places? Are taxes higher? Electricity? Perhaps we do not want to spend time understanding this. The cloud is supposed to free us from such earthly concerns.
Discounts Require Commitment:
One of the big selling points of the cloud is that you can run machines whenever you want and use them for as long as you need. Everything is about freedom. But you will pay a steep price for this flexibility. The standard price list is very steep and you will need to make serious commitments to bring down your price. Some cloud companies do not require a literal commitment because they apply volume discounts automatically, but the effect is the same. As the T-shirts say, “Freedom isn’t free”.
Moving Is Harder Than It Looks:
When setting up a cloud machine for the first time, moving to it is very easy. Networks are fast and there are often no charges for bringing data into the cloud. However, getting the data out is not that easy. It is not that it is forbidden. It is just that the mechanism is not well-oiled.
This makes it difficult to take advantage of price differences by moving your account to cheaper data centers. If your workload is heavy on computation and light on data access, it is not that hard to spin up new machines that are cheaply priced. But if you want to move data, you pay for the flow through the network. Often the meters count the bytes flowing and send you a “data egress” bill.
Sometimes the Software Is Just Open Source:
Cloud companies are full of innovation but sometimes they wrap a clever name around open-source projects and resell them as a “service”. This is not unfair. It is not just servers that cost money. Maintaining, upgrading, securing, and keeping software running smoothly takes time and skill. Some open-source projects resent cloud companies making money this way and others see them as potential partners.
The hard question for development teams is whether the paid service is a good deal. Often it is much cheaper than paying some team members to cram more obscure knowledge into their brains. But if someone already understands the software, setting up your own machine may be much cheaper. This is especially true if your usage pattern hits the expensive parts of the cloud provider’s pricing equation.
Death by a Thousand Line Items:
Cloud companies are not obscure secret societies that send you a magic bill every month without explanation. Not at all. You get plenty of detail about the thousands or millions or even billions of events that add up to that bill. You can slice and dice it with your reporting software all day long. But even after a long weekend with spreadsheets and report generators, most of us are left scratching our heads saying, “Good lord, that’s more than I thought”.
The details are all there but that does not help make the cognitive leap in interpreting what. You have to do it yourself and when you start diving into the endless logs, you realize that the data gets in your way. The truth does not set you free. It just gives you millions of data points to analyze.
Nobody Really Knows the Right Price:
The price list looks reassuring because numbers mean science, especially when there are so many precise numbers. But the prices reflect a small slice of a huge basket of changing commodity costs. The price we pay bundles together the costs of real estate, electricity, computers, engineering staff, and federal, state, and local taxes. Oh, and the Friday cake too. All these costs change periodically, sometimes significantly. Yet the business model is to hide all of this and present a nice fixed price per minute or even per second. How can one price be right?
We Asked for This:
Humans love to throw their hands up in frustration and blame someone or something else. Nevertheless, the cloud is the sum of every IT department’s wishes. We asked for billing transparency. We wanted to be volatile and buy on a whim. We needed the freedom not to commit. And we also wanted to get what we want with a few clicks. Then the bill arrives and the comprehensive schedule transparency does not make it easy to pay.
This is the fundamental paradox of cloud computing economics. The world has never had so much price transparency and so much deep digital insight into any part of the economy before. But when the bill comes, we still sit there like a sailor at the end of a beach vacation wondering where all the money went. We know but we don’t know.
