Amazon Web Services: A great power on the Internet

Icon: enlarge

Andy Jassy, ​​the future CEO of Amazon

Photo: Mike Blake / REUTERS

Andy Jassy didn’t miss anything.

He described his entry into Amazon as follows in a podcast: "I took my last final exam on the first Friday in May 1997, and started at Amazon the following Monday." The company was already growing rapidly at that time.

From eleven employees in 1995, the company had grown to 158 employees by the end of 1996.

A year later there were already 614. In the same period, sales rose from $ 15.7 million to $ 147.8 million.

Jassy, ​​who started out in marketing, quickly rose to work as co-managing director alongside Jeff Bezos, a position where he saw himself as something of a chief of staff who worked very closely with the real boss.

At the end of 2003 the idea arose to develop a technological platform for an infrastructure that was supposed to help Amazon accelerate its growth on the one hand, but which could also be rented to other companies on the other.

Three years later it became Amazon Web Services, or AWS for short.

Today, AWS are by far the largest cloud computing service in the world.

According to the Synergy Research Group, Amazon's market share in this area is over 30 percent, while the next largest competitor, Microsoft, is well behind with 20 percent.

But above all, AWS is like a money printing machine.

With a turnover of 12.7 billion dollars, the cloud division only contributed around ten percent to Amazon's turnover, but brought in half of the net profit of 7.2 billion dollars.

What makes AWS so valuable

The reason for the brilliant business result lies in the variety of services that Amazon offers in its cloud data centers.

Cloud storage space is offered under the umbrella term S3, computing capacity under the EC2 brand, and Amazon Sage Maker provides capacity for machine learning.

A list of the ten largest AWS customers who buy computing power from Amazon, compiled by the market research company Intricately, reads like a who's who of the Internet: First and foremost is Netflix.

The video streaming pioneer pays $ 19 million per month to run its services on Amazon's servers.

According to the company, Netflix uses 100,000 virtual servers in the Amazon cloud for this.

However, Apple transferred even more to Amazon in a 2019 CNBC report.

At that time, the iPhone manufacturer is said to have paid $ 30 million a month for unspecified AWS services.

It is unclear whether this number is still valid today.

However, Apple now operates a number of its own large computer systems.

Most recently, the company opened a data center in Viborg, Denmark.

Big names - from Facebook to Fortnite

But even without Apple, AWS can't complain about a lack of big names on its customer list.

In addition to Twitch, Airbnb, Facebook, Siemens, Twitter and the BBC, even the Chinese search engine Baidu uses Amazon's web services.

The popular online game "Fortnite" from Epic Games also runs in the Amazon cloud.

The fact that AWS is so popular is probably due to the flexibility that Amazon's customers buy with it.

If a company like Netflix were to purchase its own servers, it would not only have to build data centers all over the world in order to be close to its customers, but also maintain these systems, keep them up to date and expand them in line with the growth in user numbers.

At AWS, on the other hand, companies can take advantage of this as a service: They book the capacities they need when they need them.

Amazon has to ensure that the computers are running and have sufficient capacity available - and it is well paid for.

A fateful addiction

At the beginning of 2017, it became clear that the dependence of many web services on the functioning of the Amazon cloud also has its downsides. When the Amazon web services went down for hours, it was as if someone had simply switched off part of the Internet: Amazon Prime Video no longer showed any films the travel portal Expedia was not available, no more music played on Soundcloud.

SPIEGEL.TV was also unable to show any films for more than three hours.

As the reason for the failure, Amazon stated at the time that an employee simply made a mistake and accidentally took more computers offline than planned.

They recognized the problem and prevented something like this from happening again.

If that did happen, the consequences would be potentially fatal.

Since the incident at that time, the dependence of many Internet offers on AWS has rather increased than decreased,

Icon: The mirror