Cloud Chief Information Officer: How the IT department becomes a cloud service provider

' One of the most fascinating aspects of cloud computing is that many, if not most, of the discussions about cloud computing are focused on cloud computing if it affects infrastructure, ' said Bernard Golden, CEO of Cloud computing consultancy Hyperstratus, Bernard Gordon. In short, most people take the time to think about what management programs their cloud should be based on, what kind of servers to host their cloud, what storage devices to use to support their virtual machines, and so on.

There is no doubt that cloud computing is a huge change in the infrastructure. But this change ignores the fact that cloud computing is made up of a flexible infrastructure that is combined with automated operations. If you install cloud computing without automation, your revolution is only half done. The next half of the revolution is the introduction of automation into day-to-day operations and the assurance that a company's cloud provides optional, resource access, application scalability and resiliency, and a common pool of resources to be provided as needed.

Implementing a cloud environment means that resource consumers and resource providers must communicate in a service interface. The service interface is a set of automated services that can be invoked without human intervention: No phone, no application label, no meeting.

In other words, a company must include all of these features in order to become a cloud provider.

It is instructive to look at what the public cloud provider offers and how it operates, as a model for the CIO (Chief information Officer) of the cloud service provider. What core competitive advantage does a cloud service provider need to operate?

First, there are some basic competitive advantages:

Consumer Self-Service. NIST (National Institute of Standards and Technology) the first element of the definition of cloud computing is that consumers of IT resources must be able to self-service and do not need to interact with another person in the resource requester. To achieve this goal, you need to use some type of web interface, which typically comes with a service catalog of preinstalled resources. This is definitely not an e-mail message to the help desk asking for a virtual machine to be created on behalf of the requester.

Applications extracted from specific infrastructure. Cloud service providers provide computing power rather than specific hardware resources. In other words, virtual machines provided through self-service services can migrate around the cloud infrastructure and do not promise that the virtual machine will be located in a specific hardware. In the most recently published book, "Visible Ops, Private Cloud" (visible running, proprietary cloud), the authors call this virtual machine migration "initiate and transfer."

Infrastructure investment is separated from application. Many CIOs play games like this: trying to get the infrastructure investment necessary by bundling infrastructure investments with specific application plans. As a cloud service provider, this means owning a common pool of resources. The application can use this resource pool, but it is not bundled with this resource pool. Therefore, allocations for infrastructure must be handled separately from the application plan. In a way, this is just the difference between bookkeeping. However, in institutions that invest in infrastructure, this is not something that needs to be prioritized. Tying this up with apps is the only way to make this investment possible. One can expect cultural and institutional changes to be necessary. In addition, it will be noted that the overall level of infrastructure spending is likely to increase significantly. While every shift in the platform (for example, from mainframe to microcomputer) results in a reduction in overall IT spending, in fact, reducing costs always leads to significant growth in application and overall IT spending. Cloud computing is no exception.

In addition to these basic scenarios, what does this mean for the responsibility of the cloud service provider? For an IT organization, the next set of meanings is more revolutionary and challenging, and mistaken will lead to a failure of the plan and forced transfer to an external cloud provider.

Support applications with highly variable workloads and resources. Traditional application resource consumption is very stable, but a new class of applications has a higher standard of workload bias. For example, you create a Facebook application. When people "Praise" your Web page, the app offers people a chance to register a free sample of your product. This will result in a user account creation as well as an order entry. A Twitter celebrity who introduces your product on Weibo will have tens of thousands of fans in 24 hours. After two days, the focus dropped to hundreds of fans a day. Your cloud must be able to host the application well in both extreme workloads, that is, to provide sufficient resources when the workload is high, and to undo those resources when demand reduction is no longer needed.

Run automatically. These resource storms are inconsistent with normal working hours; that celebrity may be blogging on the other side of the world. When workloads impact a provider's infrastructure, resources must be able to be allocated to an application that does not require human intervention or requires no one to do any manual work. In other words, the operation of the cloud must be pre-set. In this way, resources can be automatically increased or reduced by automated rules. As a cloud service provider, staff should design this system. This system should manage individual resource requests. The need for human intervention to change the application resource allocation or structure is tantamount to admitting failure.

Capacity planning. This problem has been solved in previous blogs. However, this issue is important and needs to be reiterated. As a cloud service provider, the visibility of resource requirements is limited and lasts a short time. However, this service interface implies a commitment that resources will be readily available upon receipt of an application. Of course, a policy can be fixed (as Amazon does) and limit applications to a fixed number of servers. However, if an application is as volatile as the above Facebook application, the policy must be allowed to have exceptions. More importantly, there must be enough usable capacity to meet the level of application workload needs.

High usage rate. Many demo asserts that using a private cloud is cheaper than a service from a public cloud service provider. There is a lot of controversy about this view. Many people have made important comments on both sides. Gordon said that in all of the demos he had seen, the use of more than 70% per cent of the speculation increased the reason why private clouds were cheaper. Because the cloud service provider, like any capital-intensive retailer, such as an airline, would go bad if the utilization rate did not reach the required workload. In the past, the responsibility for server utilization was primarily for the application department: if they were configured too much, the expected results would be regrettable and bad. If you run a private cloud and explicitly promise to be more cost-efficient than a public cloud, it is important to suddenly achieve high utilization. This task becomes more challenging given the changing nature of the application's workload stability. Amazon uses smart incentives, such as on-site instance pricing, to increase resource utilization when workloads are low. Such things may be necessary for the CIO of a cloud service provider.

All of these elements can be summed up in one thing: operations like a cloud service provider need to do things like a business rather than a cost center. The beauty of a private cloud can be matured only when it comes to the realization that, as a cloud service provider, the flexibility of the infrastructure is combined with operational capabilities. Every previous generation of technology upgrades drives the same way of doing business: Smart people install and configure each part of the hardware manually. The word rack and stack perfectly illustrates these practices. Cloud computing is the second generation of operational upgrades that need to be accompanied by technology upgrades.

(Responsible editor: admin)