Cloud computing and data center interdependent development are inseparable

Cloud computing is from the data center. We've always dreamed of a world where just about anyone can turn the extra computing power into virtualized resources and sell it to others, a completely distributed cloud computing model. However, the reality now is that cloud computing follows a centralized factory model: resources are provided by central "factories" (Amazon's, Rackspace, Microsoft, Google, and other large data centers) and distributed to consumers over the Internet.

In a sense, large-scale data centers are the foundation for cloud computing. Before today's incredible large modular data center appeared, selling computing resources to others was simply a nightmare. For most customers, not only do they cost prohibitively expensive to provide them, they may not be able to do so.

Scalability, availability, resilience, and security are factors that must be incorporated into data center design in one form or another, from the structure of the building where the data center resides to the software used on each server. Otherwise, any promise given by cloud providers will not be realized.

Cloud computing provider's commitment

As for the key commitments given by cloud computing vendors, especially in infrastructure as a service, we can see that most of these commitments depend on issues that arose prior to the establishment of the data center. For our customers, the promise of uptime seems simple: I can access and use resources when the service is up and running, and I will not be able to access resources if something does not work. And for service providers, the problem is trickier: the service is not functioning properly because of a server or disk failure, or a major blackout. Many lower-level problems, such as power or network connectivity, are associated with larger problems, such as the location of a data center.

In many countries, you can not access two separate external power suppliers at the same time. This fact may force companies to invest more money to buy generators or other backup energy, which undoubtedly increased operating costs. On the network side, the same problem is faced with the inability to access multiple vendors and enterprises may be forced to deploy their own infrastructure or even "accept their own accord" on the fact that they may not be able to reach a certain level of availability.

Extensibility is another interesting issue. The (almost) limitless scalability offered by cloud computing service providers means they have to continually increase their available capacity in order to be able to cope with any increased demand. Every day, infrastructure providers need to add hundreds or thousands of servers to their data centers, not only to replace failed devices but also to meet future needs. Delivering scalable resources to infrastructure end-users will result in situations where you simply do not know how many resources a single user may need (which is why most cloud service providers virtualize the number of servers available to a single account The reason for the restrictions).

Even simple things like Amazon's Availability Zone with multiple data centers in close proximity to each other with low-latency connections can introduce design and management complexity. In the case of Amazon, uptime is a measure of multiple free zones, not individual zones, and these zones must be sufficiently dispersed so that external issues affecting an available zone will not affect other available zones while , The distance is not too far to ensure low latency connection. At the same time, they represent a new data center concept: multiple data centers can operate as a single data center.

Cloud computing and data center keep pace

As deployment rates increase, cloud computing will continue to drive the evolution of data centers, from architecture to software and control processes. Broader deployments will not only force data center operators to rethink their internal operations, they also need to be able to adapt to new emerging needs. As usage increases, there is also an increase in energy consumption, and as the computing resources provided by vendors become more diversified, the environment will become more varied and management will be more complex. Simple measures (such as the use of outside air to replace air-conditioning equipment) will help businesses save millions of dollars.

If cloud computing brings the commercialization of computing resources, the data center will need to be optimized to allow businesses to survive by selling those resources. As cloud computing will depend on the data center for the foreseeable future, the evolution of both technologies is undoubtedly linked, and anyone looking at one of them should pay close attention to the other.