A Free Trial That Lets You Build Big!
Start building with 50+ products and up to 12 months usage for Elastic Compute Service
Building a private cloud is not a quick project. Starts with understanding the enterprise environment's expectations and definitions of Yu Yun, and then frames the models that have been created. Be sure to cover the entire organization, all of its processes, and the technologies that build cloud computing. Here are ten steps to conceive, build, and maintain a private cloud.
1. Determine what you are going to get from the cloud
The cloud journey is a huge it trend. The problem is that the terminology cloud computing has a different meaning for everyone. To start this journey, businesses need pragmatic cloud-computing goals. Many companies find themselves looking to cloud computing as a commitment to virtualization, like integration of physical hardware in the datacenter, energy saving and cost savings. Others find that their virtualization process is going into the next phase, and that standardization and automation are part of their IT processes.
But a handful of companies are preparing for organizational change, dealing with tough "people's problems (arranges problems)" like islands, service replicas, security and service management. These problems are often not technical in themselves, but they are across organizational boundaries and will remain in the political arena for a long time.
There are many misconceptions about the cloud, especially about how much it is defined. One of the most common misconceptions is that the cloud is based entirely on virtualization. While virtualization often takes a heavy role in private cloud deployments, private clouds can be just the shared infrastructure. Google's Gmail or Microsoft's SkyDrive, for example, are both public cloud services, but not overly reliant on virtualization.
Instead, a lot of physical hardware is used behind the scenes. The private cloud is also true for your organization, where shared services are created to replace many different duplicate services, and the use of virtualization is evaluated only as part of the service implementation. For example, a shared file Server service might replace a file server in many departments, which could be implemented on a physical server because of the compatibility between VMware Vmotion and Microsoft Cluster Service.
2. Realistic Expectations
What you should expect is that there will be no real self service it in your organization. IT departments have spent years wrapping processes and processes around creating and managing servers, and for good reason. Typically these processes are responsible for system monitoring, sizing and dependencies, documenting system design and responsibilities, processing licenses, and more.
Allowing anyone to provide a server or service without an approval mechanism is appropriate for a specific lab or development environment, but the production IT environment is a fast way to clutter, sprawl, and downtime. However, it is reasonable to expect to provide more process automation and standardization by using workflow tools and approval mechanisms, which have become part of Embotics's V-commander or Enstratus NX products.
Cloud Tour technology challenges less, and more human challenges, with process decomposition and refactoring, small tasks automation, and standardization implementation. The IT department is clumsy, slow to show off the user's needs, probably not rethinking itself and work from the right place exam. Similarly, IT departments are overworked and may not have enough time to track cloud solutions, although the cloud can save time.
The order in which you manage your IT work is important because it supports them in the face of complaints about focusing on cloud computing delays in other jobs. The motto "money-making process will cost" time for IT staff; it takes a prepaid time investment and saves the time that follows.
Finally, to expect all levels of management, including human resources, to support the transition to the cloud. It is not just that all the organizations in the enterprise see that the directors who are improving themselves are delaying it work, and that the main work of the IT staff will be automated and that they will be the targets of dismissal. They may be aggressively damaging the process. Individual problem plans, from the beginning, people communicate, they are valuable, their efforts will liberate them, to do more interesting, productive work.
3. Understanding Workloads and services
Moving forward like a private cloud model is tough, especially if you don't understand the services your organization relies on. The document is the key; without it explaining its relationship to the system, the service level agreement is unknown, and there is a lot of false assumptions. The need for people to use these services should be documented so that new cloud services can be built to meet these requirements. This is especially important when repeating services are in the centralized organization. Another reason is that a department builds its own infrastructure, rather than using shared services, and finds out why, and allows them to identify and avoid disputes.
The documentation also applies to the standardization itself, as the standard does not explain all requirements and system design requirements and will quickly get an exemption clause.
Performance information is also critical for moving to shared infrastructure and cloud-based solutions. A year or more of historical data, the highest possible clarity is highly practical and helpful for determining capacity requirements and system size.
4. Step into the path of virtualization
Although a private cloud does not necessarily require virtualization, this is a generic model. Virtualization typically drives specific knowledge and behavior within an enterprise. For example, most virtualization requires centralized storage. This same centralized storage will be an artifact of building a private cloud, so the knowledge gained in implementing virtualization is useful for private clouds.
Similarly, virtualization often breaks down data center networks. At the very least, it can transform the static flow pattern into dynamic mode. The shift to shared computing model and cloud based computing maintains this trend and increases reliance on the network, often increasing bandwidth requirements. The conversation begins with your virtualization administrator, storage Administrator, and network administrator, and the result is that the virtualization initiative will be the key to your development of cloud computing, especially when planning to service remote offices and mobile users.
5. Understanding Automation and standardization
Automation is one of the key goals for an enterprise to move to a private cloud. However, without standardization, automation is tricky. For example, the establishment of operating system and server standards allows you to assume file location, file system size, and authentication mechanisms. Based on these assumptions, you can script application software and middleware installation, such as Web servers, application servers, and firewall rules. The installation is easier to repeat, anyone can iterate over the deployment, or it can be handled as soon as the disaster occurs.
Standardization is very difficult for an enterprise that has not been able to practice it. But once standardization is adopted, time savings are obvious.
Imagine an enterprise without the standardization of operating systems, operating system versions, or component processes. Each server is different, and each operational requirement requires special attention. The time of the patch process or software installation is different, and the success rate is shaken because every master
Machines are changing. This usually has two results: the time that employees perform daily tasks on these servers is unbelievable, and many routine tasks like security bug fixes are skipped because they are too difficult and unpredictable. standardization and automation of one or two operating systems, and application deployment process productivity greater than it productivity gains.
Once you have automated most of the transformation, you can deliver it from service portals and service classifications. Although the enterprise is not likely to be 100% completely self service driven, but many processes can be automated according to the workflow; the only interaction is the approval process. IT department
The door can focus on more important issues, how to better support, monitor applications and services. It also improves the lives of application managers and developers by providing a consistent and repeatable platform. This means that it operators can build more useful, reusable streams
process, handling incidents and detecting system alerts, not every server is an exception. This may also open the door to automated response alerts.
(Responsible editor: The good of the Legacy)
Start building with 50+ products and up to 12 months usage for Elastic Compute Service