Cloud computing drives the development and application of large scale data centers

From the history of computational evolution, we can see that the IT industry will make a transformative change about every 15 years or so. Different technical forces collide with each other and eventually produce a revolutionary force that can change the whole industrial ecology. Every computing age will have its own technology. At present, the impact of cloud computing on IT industry is enormous and far-reaching, cloud computing because of its flexibility and scalability, is promoting it rapid evolution, making the current it ecosystem more challenging, if the enterprise can not adapt to such changes will be eliminated.

Cloud computing facilitates large scale expansion of data center

Mortenson, a company that specializes in data centers, finds that small and medium-sized enterprises face significant challenges in setting up data centers-such as scalability challenges and high-density challenges. In addition, the building of data centers is becoming more and more energy efficient, for example, the site will consider setting up a data center in some cold climates, using modular or containerized infrastructure at the time of construction, something that is not affordable to a common enterprise and requires a professional cloud data center provider.

As a result, today's SMEs are beginning to enjoy renting the resources in the cloud rather than building their own data centers to reduce costs. "Outsourcing needed data centers" has become a trend, requiring cloud data center providers to be large enough to respond to this new market demand. Ron Vokoun, head of data center construction at Mortenson, said the data center used to occupy only one or a few rooms, but now the smallest data center has reached 20,000 square feet, and the project is twice times or 4 times times larger than ever. As data centers grow in size, data center management also faces problems.

Mark Thiele, the executive vice president of the switch company responsible for data center technology, describes the management operations experience of a large data center in the cloud computing age. The company is managing a very large data center called Supernap. Built in Las Vegas, the SUPERNAP has a total area of 407,000 square feet and accommodates 7000 boxes of 250 MW of power generating capacity, which supports up to 1500 watts of power density per square foot. With a power consumption of 100 megawatts per day, the switch uses high-density T-SCIF (thermal separator facilities), complete separation of control systems, and hot and cold channel settings. Supernap can automatically choose the most effective cooling scheme under four different switches. Mark Thiele said Supernap is about to become a bigger "super-dupernap" that will expand to 2 million square feet in the near future.

Mark Thiele said they had a clear position on themselves, and they positioned themselves as "hosts and data connectors (host and interconnector of Clouds)" in the context of cloud computing, meaning that its size would continue to expand, This provides services to cloud vendors, not just end-users. The scale effect of this pooling of resources can help those cloud providers reduce economic costs and, more importantly, greatly simplify the complexity of cloud environment connections between different vendors.

It is believed that in the near future, cloud computing data center will become a huge, cost-effective and efficient (1.24 PUE) computing environment, making traditional enterprise data center and it is like a heap of scrap metal.

Case study: See how the EV manufacturer Smith uses the cloud architecture

So, from a user's point of view, what factors are driving them to the cloud data center resource providers of these third parties? At the end of the meeting, Smith described the direct impact of cloud computing on their company's business.

Michael Peacock is a software developer for Smith, a British electric vehicle manufacturer. Instead of small and medium-sized vehicles such as golf, the company produces large commercial trucks weighing up to 8-13 tonnes. In order to monitor the daily operation of these cars, Smith installed a large number of monitoring sensors in the car, these sensors will send real-time vehicle speed, power consumption, engine speed and other data to the background of the data center server. Michael Peacock says the data will be written to 4,000 MySQL databases per second, reaching 1.5 billion times a day, straining the company's server equipment.

So Smith launched a new project to migrate data across the cloud to new large-scale data centers to support different data changes that the software monitors every day. Today, the IT architecture of the traditional manufacturing enterprise, founded in 1920, has become quite trendy: not only is the cloud computing platform used, but NoSQL is used to deal with these massive amounts of data.

Also, Smith has made some optimizations for the new IT architecture:

First: Use a queue based task submission architecture to store these queues in a data center of a cloud service provider to respond to sudden surges in traffic and unforeseen data-processing requirements.

Second: Maximize the performance of SAN storage facilities to enhance storage architecture to cope with the need for data overload.

Third: Redesign the application's database with a streamlined schema to improve application performance.

Four: Improve the performance of data analysis by reducing the processing time of queries by centralizing the data in the background for the batch processing.

It is the business

It can be seen that, although Smith is a long history of traditional manufacturing enterprises, but its IT application has been in the forefront of the times. This, in part, illustrates the trend that it is moving from a business-supported role in the enterprise into a business process itself, and that traditional application design and infrastructure are not enough to support today's needs. This, in turn, has prompted a change in the computing environment, moving it from the inside of the enterprise to the cloud, to meet the needs of future computational development.