Amazon was once solely known as a fast growing online retailer. The company moved its systems numerous times between various servers. It started with Hewlett-Packard before moving to Rackable Systems. That move allowed the company to take full advantage of new innovations in high density rack design. This company merged with SGI. Now, things are changing again. Amazon Web Services, perhaps one of the fastest growing online companies, designed its own custom servers and switches and has been thriving in this environment for some time as a result.
Why Set Out on Its Own?
Like most massive companies in the field, Amazon Web Services didn’t formally say why it moved to its own servers, but a bit of light was shed on this at the re:Invent 2015 conference held in Las Vegas in October. At the event, the vice president of infrastructure with the company, Jerry Hunter, discussed this specific topic. While Hunter did not (can could not) provide specific design components in this very competitive industry, he did provide some insight.
First, consider how the retailer portion of the company works. The company’s virtuous cycle allows it to offer an always increasing product line to an increasing number of customers and that drive allows the company to leverage its size to demand the lowers prices. As a result this pushes Amazon even further up in volume. Now, consider Amazon Web Services component.
In this area, the company provides capacity on demand and does so for more than a million users. Every one of these individuals has different needs that change frequently. The company has leverage in this area for several reasons. First, it uses a minimalistic machinery approach with only the hardware and software the cloud requires. And, it has a very complex redundancy built into a network of machines. The focus here is on regions, availability zones, and data centers. All of this gives the company the ability to leverage its market. That’s something most other companies cannot accomplish.
Scale Makes the Difference
According to Hunter, the ability to customize like it does allows the company to get rid of the excess features other companies often rely on and provides for a more simplistic design. This improves the operational component and the reliability aspect of the company’s function. The company works with OEMs and other manufacturers to ensure capacity levels remain and that all components – memory DIMMs, racks, hard disk, networking gear, racks and so on continuously remain within reach for on-demand needs.
What does this afford the company?
Through the cycle, the company achieves a higher level of reliability and it helps to drive down costs. As this occurs, Amazon Web Services gets even better because it can better compete against often times larger rivals including on-premise data centers. And, this attracts new customers, which increase the company’s scale even further, continuously turning the cycle and making it harder for virtually any other cloud company to compete with them.
Take a look inside Amazon Web Services and how the company’s size and continuous growth allows it to offer custom servers to continually drive success.
In addition to her work with rack solutions Katrina has a video series featured on Youtube called “Ask Katrina” that aids in answering your IT needs.