How To Choose Between Multiple Commercial Machine Clouds?

Actions
How To Choose Between Multiple Commercial Machine Clouds?
Linda Jones

Article by

Linda Jones

Feb 12, 2021

You have to be picky.

The cloud began as a commodity. Of course, there were small differences like the size of RAM or the way it measured for virtual processors, but the cloud provided a seemingly endless supply of seemingly identical machines. They were executing the same distributed packages and responding in the same way at the command line. And as soon as you snap your fingers, cloud service providers would give you root on the other. Keeping it, all the same, was the easiest way to lure developers into the safety of air-conditioned racks in the same building.

Machines as commodities mean that there are no surprises or problems. All the clouds offered the most popular operating systems in which all the bits were arranged in exactly the same way. The big problem for managers is dialing. If Ubuntu 18.04 is the same everywhere, what's the difference with Google, Microsoft or Amazon hardware? If major distributions are supported everywhere, how do you decide?

To make the choice harder - but consumption easier - space is very competitive. Developers at cloud companies come up with clever ideas, but they are quickly copied. Genius very quickly becomes average. Innovation creates disruptions that grow into everyday sets of characteristics that we take for granted.

How to choose? You can't just throw a coin. It's not scientific - even if you wear goggles and wear a lab coat while doing it. If directors ever notice that you are throwing a coin, they will understand that there is no need to wait for artificial intelligence to be strong enough to replace a purchasing manager. They can replace you with a monkey and a coin right away.

The solution is to be nice. Yes, probably any of the commodities from any of the major clouds would satisfy - or even many of the not so large clouds - but who wants to go through life just like that? Who wants to meet the average?

Being picky sounds petty, but it's really the beginning of innovation, a tip that drives real change. It is really sensitive to important differences and taking differences into account.

To help you with this process, here are 10 different picky reasons for choosing one of the major clouds. The reasons are not unambiguous, as it is usually possible to achieve the same thing using any of the competitors. But if something is possible, it doesn't mean it should be done.

 

APIs

All Cloud offers many smart and sophisticated APIs such as Google Cloud Vision, Azure's machine training service or Amazon's GameOn. There are hundreds of them, and they make it much easier to create your own code. There is no reason why you should not be able to call these APIs from any cloud or from any computer on the Internet, but sometimes you will need performance that is achieved by working on the same network or even in the same data center. If a cloud offers what you need, it might be a little faster if you do most of your computing and storage in the same place.

 

Location

All clouds have data centers scattered around the world. Microsoft Azure, for example, has 54 regions and they carefully record where data is "at rest" and which government has sovereignty. You may have a large collection of clients in one country. Perhaps the legal section has identified a special and particularly lucrative "feature" of another country's tax law. There are dozens of strange and often deceptive reasons why you might want your code to be executed in one country and not in another. Most of these different data centers are clones of each other and it makes sense to stick to the same stack around the world. It just makes things simpler. The only downside is that not all data centers are perfect clones and not all products are available in all locations.

 

Special Spice

Although computing and storage resources are mostly commodities, each of the clouds offers several products that stand out. For example, Google's Firebase database, which automatically transfers the correct data to the client, saves you a lot of worry about consistency and download code. Amazon's S3 Glacier will store data at a very low cost over the long term. Microsoft has Azure Cosmos DB, a database that supports multiple data models. There is no reason not to merge services from different clouds, but it is often a little easier and faster to buy all the computing goods from the same cloud.

Desktop integrations

Clouds are always the corners of larger corporations and there is often synergy among different departments. If your office is using Google G Suite, everyone will be more comfortable working with some of the Google Cloud Platform services and uploading their data to Google Sheets. Microsoft Excel is one of the standard ways many large organizations handle data, and it can import data from Azure or work with Azure Analysis Services. Old habits die hard, and the desktop is one of the oldest habits in the business.

 

Spot Market

If you are a true believer in free markets or simply like the excitement of "winning" the auction, Amazon offers a spot market for computing power that allows the price to adjust to supply and demand. This is a great opportunity for jobs that do not have to be completed at any particular time. Batch jobs and monthly reports can wait until the price is low enough. Just remember that demand can jump at odd times. Some of the flow services, for example, use the cloud to handle additional demand on Fridays and Saturdays at night.

 

Automatic Discounts

Google offers a pricing mechanism that doesn't require your thinking. When your specimen runs for more than 25% in a month, you start to get a quantity discount. No negotiations. No new contracts. No switching to a different plan. The price simply goes down. This is called a "continuous use discount".

 

Microsoft's acquisition

It's hard to deny the dominance Microsoft has had over many enterprise development teams over the years. If your stack originated during the .NET era and still uses many Microsoft parts, there is a good chance that you may find it easier to choose Azure. This is not necessary as other cloud vendors can run Microsoft machines and there are good containers full of Microsoft code that can be executed anywhere. But sometimes it's not as nice as home.

 

OS versions

Different versions of operating systems are intended to be commodities and are, for the most part, exactly the same. Ubuntu 18.04 is likely to behave exactly the same on any of the clouds. At least that is how it may look. But sometimes there are small differences deep into the core and sometimes those differences show up in performance during evaluation. In some clouds, this optimization is obvious. Clouds offer their own versions of operating systems and it usually makes sense to use them if you can.

 

Performance measurement

The specimens look like commodities, but there are subtle differences between the architectures and the speed of the various components. Measuring the performance of your code on different platforms and different types of machines make great sense and may reveal that some clouds offer a better price-performance ratio. Many cloud machines look like different versions and the only real difference is the number of processor cores or the amount of RAM, but sometimes there are architectural differences that can affect the performance of your code.

 

Logo

You will see this logo every day when you log in to check your copies. It's better if you like it though.

Comments (0)

You must Register or Login to post a comment

1000 Characters left

Copyright © GLBrain 2022. All rights reserved.