The price of computing power per 3 years will fall by 50%

"Editor's note" Ford's T-car manufacturing pioneered the use of modern assembly line manufacturing, greatly reducing the cost of the entire vehicle, the result in the 20 years in 1920-1940, the United States to the 200 car manufacturers blood to only 8. Similarly, cloud computing, which follows the "Bezos law", will also unleash such storms in the data center market. According to Bezos law, the price of computing power per 3 years will fall by 50%, and over time, the future of the enterprise data center will become more and more uneconomical, the end result is a comprehensive enterprise to cloud computing.

The following is the original text:

In the past few decades, the development of the IT industry has followed "Moore's Law", and in the era of cloud computing, Amazon CEO Bezos a "Bezos law", which is "the price of computing power per 3 years will fall by 50%".

Greg O ' Connor, Appzero chief executive of the recent "Bezos law", once again wrote in Gigaom that the total cost of infrastructure ownership of cloud computing is tcio with the constant innovation of cloud computing services providers and the decreasing cost of various costs, including hardware. Cost of infrastructure will be far more than the enterprise to build the data center, and this gap will continue to pull up, and ultimately promote the enterprise to a full range of cloud computing.

Connor points out that cloud computing is very similar to the Revolution Ford gave to the enterprise industry, Ford's T-car manufacturing pioneered the use of modern assembly line manufacturing, greatly reducing the cost of the entire vehicle, the result in the 20 years in 1920-1940, the United States 200 car manufacturers washed to only 8, a sharp reduction of 96 %。

The low cost of innovation can lead to a revolution in the corporate world, and similarly, cloud computing, which follows the "Bezos law", will set off such storms in the data center market.

In contrast to Moore's Law, which focuses on the growth of CPU performance, the Bezos law considers Tcio, which is larger in scope, not only for processor performance and cost changes in data centers. Tcio is also more persuasive for companies that pay attention to costs.

Why is Tcio more critical?

The IBM-acquired SoftLayer team commissioned McKinsey to study the Tcio, and to conduct an understanding and in-depth analysis of Tcio:

As can be seen from the above figure, the cloud service providers in many areas to significantly reduce the cost of cloud computing, hardware procurement and upgrades are only one part.

Energy costs account for 31% of the monthly operating costs of the data center, according to an article in the Amazon James Hamilton blog about the overall cost of the data center, excluding data center construction costs and staffing costs. (And this is 2010 data)

Cost reduction is not just hardware

In addition, there are plenty of reports about the ultra-low energy efficiency of Google and Facebook's green Data Center (PUE), such as "hot on Facebook's green data Center", where Facebook's data center's PUE reached an astonishing 1.07. According to the efficiency index of the enterprise data center, the average pue value of the common enterprise's data center was 2014 1.7 (below), and the energy efficiency was much lower than that of the large cloud computing service providers.

More than just Tcio, cloud computing has many competitive advantages over the company's own data centers, such as continued scalability, innovation, and competition and price transparency.

Assuming that Fortune 5000 companies have an average of 7 data centers, or about 35,000 data centers, according to Bezos law, by 2030 these companies will save 90% of their computing costs if they all turn to cloud computing. According to Gartner, the market for cloud computing will be as high as $131 billion trillion in 2017 (the company that is responsible for transforming the data center into an OPS center will make a lot of money).

Original link: Cloud Computing's Bezos law: Data Center will end (Zebian/Wei)

Free Subscription "CSDN cloud Computing (left) and csdn large data (right)" micro-letter public number, real-time grasp of first-hand cloud news, to understand the latest big data progress!

CSDN publishes related cloud computing information, such as virtualization, Docker, OpenStack, Cloudstack, and data centers, sharing Hadoop, Spark, Nosql/newsql, HBase, Impala, memory calculations, stream computing, Machine learning and intelligent algorithms and other related large data views, providing cloud computing and large data technology, platform, practice and industry information services.