Build Your own private cloud

Private Cloud is the latest attempt by data centers to improve efficiency and enhance hardware and software flexibility. The idea is to create a pool of computing resources, storage, and network resources and manage them from a central software control point, rather than manually implementing the relevant regulations for all infrastructure, monitoring and reconfiguring management.

Predictably, the public cloud service provider itself is the most experienced in designing and running shared infrastructure. Amazon, Google, Microsoft, Salesforce will not use the old traditional approach to static infrastructure dedicated to individual applications to manage their own data centers, and in a vain attempt to use millions of users on the Internet to provide reliable and cost-effective services. The pioneers of these cloud domains run public clouds in large scale and have already made the basic model of private clouds.

Reduce the size of the cloud

Some readers will soon point out that public cloud service providers often offer a large number of large-scale services that may not apply to some typical features of a wide variety of small and medium-sized enterprises. So does the public cloud model apply to all businesses? is private cloud "just virtualization plus a little bit of automation for a marketing phase?"

In fact, the public cloud model can be used regardless of whether server virtualization is valid, even if the specific size and deployment implementation Details may not be the case. Virtualization technology has become so pervasive that we tend to think that pooling resources, especially those distributed across a physical host group, has become a fact of life. In fact, this integration represents a profound change in the way we manage our computing infrastructure.

This difference is particularly acute when virtualization reaches critical mass. At a certain point, such as reaching hundreds of physical hosts and thousands of virtual machines, or servicing many different "tenants" need to meet their immediate needs, virtualization management is essentially a different proposition.

At this point, the enterprise will need private cloud software to allocate resources, and let the related cost return to the business line of stakeholders acceptable scope. If possible, organizations need to take full advantage of the flexibility of virtualized infrastructures and allow these business stakeholders to allocate resources themselves. In addition, when merging data and storage networks, organizations also need to implement storage and network resources, as well as virtualization of servers.

All of the above and other features have been implemented by software to varying degrees from Eucalyptus,vmware,openstack, Microsoft, Citrix, and other cloud service providers in different ways. However, these different proprietary cloud software stacks are not intended to replace each of the data center's infrastructure software. They provide a framework to help administrators manage the Virtual data center infrastructure overall.

Elements of a private cloud

A private cloud forms a new management layer across a number of functional elements in the data center. It sounds like a myth, but remember that migrating to a private cloud is by no means the same thing. It can be rolled out in an incremental. While many hardware and software vendors believe that businesses need a mix of new hardware racks and an industrial-scale software to build a real private cloud, businesses can often migrate to implement private cloud features by deploying to an existing virtualized infrastructure.

Calculation。 At its most basic level, private cloud computing components do work with virtualized management software: Configure, Undeploy, start, stop, pause, clone, and migrate virtual machines from central control points. The main difference is that the goal of a private cloud is to enable organizations to manage as many virtual server infrastructures as possible, which means multiple virtual machines monitor hypervisor support.

Storage。 Cross-platform Storage Management has been a dream for many years. Most private cloud platforms strive to support the storage and replication of objects, as well as block storage and snapshot management of storage systems across different data centers, although this is still a work in progress. Another important component of private cloud storage is the effective management of virtual servers and virtual disk mirroring.

Internet. Private cloud software extends to the forefront of SDN (a software-defined network). You can create virtual ports, connect virtual machines, configure virtual switches, and more. One of the important features of a private network that can be configured within a larger network infrastructure is that when you create a large private cloud, several different groups of users can access it. In a modern data center with 10G switches, converged data and storage networks, dynamic separation of computing pools and bandwidth can be said to be quite powerful.

Metering and Billing/Showback mode. When infrastructure is shared, organizations need a way to divide and measure pooled computing pools, storage and network resources, and automatically collect lines of business accounts or at least relevant presentations to stakeholders. Without this functionality, IT staff are either stuck in a complex costing spreadsheet or content with inaccurate or unfair cost allocations, which in turn will haunt them.

Self-service. Perhaps the biggest contribution of the private cloud to efficiency is that it can effectively meet the needs of the stakeholders ' self-service. For example, for developers, they can configure and use self-service portals to reconfigure their own development and test environments. The business unit even "commands" and installs the application, which is automatically configured, and does not even involve IT personnel in the process. Of course, such portals, management policies, and automated self-service configurations can help improve productivity. At the same time, the standard choice of stakeholders is bound to be limited. However, with the continuous expansion of its self-service menu, the relevant personnel do not have to spend time and resources to deal with the related requirements, configuration and other aspects of the work, which undoubtedly brings huge cost and time savings.

Automation and business processes. For self-service portals, businesses need automation, or basically a script library that performs common tasks: Server configuration, shared storage settings, network settings for a VLAN, and so on. Automated assembly tasks for business processes to predefined workflows to meet specific applications or services, which may require approval. For example, if a developer wants to test a large-scale application, it will need to use a large amount of resources and replicate the real-time database, and the software will automatically issue an approval management request.

Application management and monitoring. Private clouds can extend all the time to the management application tier. This includes deploying and configuring applications, as well as monitoring and reporting on the ability of the program to meet internal SLAs, providing administrators with the Insight templates they need to fulfill their commitments. At the state-of-the-art level, monitoring can trigger automatic scaling so that performance levels can be maintained without human intervention.

Security。 The need to isolate multiple tenants in a shared infrastructure environment, after all, security is extremely complex and critical. Identity management system, provides multi-level authorization management, according to the user group assigned permissions is a key component. It should also be extensible access control so that it can be extended to external users, such as customers and partners, as well as internal teams.

The part of the private cloud that should be implemented first in your business will depend on the workload. If you plan to host your enterprise's partners in a private cloud, the security and networking components you need must support virtual private cloud, which is likely to be the first phase of an enterprise's private cloud deployment. If managing resource allocations for enterprise development and testing is a top priority, portals that provide self-service for developers, along with automation and business processes, should be placed first.

In addition, if a fully functional private cloud environment is deployed in a company that does not have the scale, interest, or ability to leverage its self-service capabilities, it will undoubtedly be a huge waste of time and money without saying so.

Many large and medium-sized enterprises can benefit from the private cloud. What is really difficult is for each employee of the enterprise to endorse and agree to its relevant rules, rates and policies of the automated system. If you can clear the barriers in your business, your organization can virtualize everything, prioritize important workloads, and incrementally add more private cloud functionality to deal with stack problems in the way your business needs.