From Virtualization to the Cloud: Use IT Resources More Efficiently

Highly efficient administration and operation are key for attaining flexible IT operations in a data center, which are the core of the company's IT system. Server virtualization has been established as the preferred method for increasing productivity and flexibility in this area. A glance at the trends in business IT illustrates this. Gartner (News - Alert) market researchers report that the number of virtualized servers in data centers has noticeably increased over the last five years. Up to 50 percent of servers had been virtualized in many companies by the end of 2012. This percentage has risen to 70 percent in the last year, especially with the newly delivered systems.

Comfortable Administration Supports IT Agility

Usage and implementation of administration tools is always closely tied to the long-standing organizational structures and technological landscape. Companies risk missing out on the expected benefits of a virtualized environment, in particular the much-desired IT agility, if they fail to conduct an adequate analysis of all processes that could be affected by virtualization. The definition of the individual administration processes plays an important role in this respect. Examples include the provision of virtualized applications and servers, as well as change management. If in doubt and without specification, each application requires as many server resources as possible. Companies therefore require administration tools that show what resources the individual applications need throughout the course of the workday. When equipped with suitable administration tools, administrators are able to identify issues quickly to ensure that mission-critical applications take priority.

Virtualization Is a Good Starting Point for the Cloud

Virtualization is a central component of a successful IT strategy in any industry, regardless of the company’s size. It is often the first step toward cloud computing because IT organizations that have already virtualized their IT environment can later move to the cloud without any problems. As such, virtualization can be viewed as an excellent starting point for any cloud initiative. It is in high demand in companies, because the employees are more well versed in using the cloud in their private lives than the company.

Everyday life has practically become unimaginable without cloud services for messaging, online storage, music, and videos. Consumers use services such as Dropbox (News - Alert), Flickr, iTunes, Skype, and WhatsApp at home or on the road almost daily. These are ultimately nothing more than public cloud services. They are easily and quickly available and operate largely without problems. People desire such cloud services at work, as well. If we look at the differences between public cloud providers and corporate IT, we can see why this is not already in place. The different infrastructures of typical public cloud providers are designed for dynamic and agile operation. Everything is constantly changing. In comparison, corporate IT appears quite rigid and inflexible. However, it also has to deal with heterogeneous hardware and software solutions that were developed at different times and by different technology paradigms.

The corporate IT sector is currently starting to catch up and compete with public cloud services that are familiar to consumers. This is a must, because there is definite demand from the different industries. Not wanting to wait any longer, many are using external cloud services, bypassing the IT area altogether. IT should therefore act quickly and consider options to combine external and internal services. However, it will still require some effort to get company-owned cloud services as easily and quickly launched as the public cloud services. The appropriate administration tools play an important part in addressing this issue.

Open Hybrid Cloud Services Provide the Building Blocks for a Solution

Greater flexibility and speed are necessary to meet the growing needs of the departments. The offers from the public cloud are suitable for specific application scenarios, but they are not comprehensive enough to cover all requirements in the field of mission-critical applications. Therefore, various groups within the IT department are examining from different perspectives how the production cycle of applications can be accelerated with the help of available technologies. There are several options available for prototyping, testing of modules, and transferring finished applications to the departments, traditional virtualization, infrastructure-as-a-service (IaaS), or platform-as-a-service (PaaS) environments. The development team will prefer the PaaS approach and the infrastructure and operations team would rather opt for a virtualization or IaaS environment. There is no ‘one size fits all’ environment for all application scenarios. Generally, an IT department will opt for an open standards-based combination of the aforementioned models and use the most appropriate one for the given situation. In other words, an open hybrid cloud will be used.

It is important to note that few if any companies will completely migrate all existing, tested, and often mission-critical applications to the cloud. Every company should first carry out an inventory and identify the application scenarios that are suitable for operation in the cloud. Practical experience has shown that it is better to initially concentrate on new cloud applications.

All cloud technologies are ideally suited for certain application scenarios and not for others, and IT will therefore host various components of the application on different platforms for a department. When you start using the cloud, IT departments operate, for example, the Web front-end on a scale-out IaaS or PaaS platform, while the database system and middleware probably run on a virtualization platform, which runs in the internal data center and is nearly infinitely expandable. The horizontal scaling (scale-out) architecture will be replaced over time by a vertical scaling (scale-up) architecture, in other words by additional computers, and transferred from the internal virtualization to an IaaS platform. The IT department will migrate some application components to the PaaS platform and others to the public cloud.

However, the simultaneous operation of multiple cloud platforms is only the beginning. Without a comprehensive management environment, IT departments would be dependent on tools for each platform and would have to manage the same rules and regulations and monitor compliance to it, while running several operating systems in parallel. A unified cloud management platform links the various on-premise, PaaS, and IaaS components of an application and keeps them together, regardless of whether they are hosted on a scale-up or scale-out platform in the cloud or are operated in their own data center. Additionally, it oversees the entire lifecycle from deployment to decommissioning in a mix of private and public cloud.

OpenStack as a Cloud Operating System

Many companies currently focus on evaluating the open source cloud platform OpenStack when it comes to setting up new cloud infrastructures. This is a vendor-independent open source project that is rapidly expanding. OpenStack can essentially be described as a cloud operating system consisting of processing power, storage space, and networking capacities with which a company or service provider can build a cloud structure in a data center environment. Some cloud providers already use an OpenStack environment as a cost effective alternative to a VMware environment. Thanks to its modular design in the OpenStack core areas, cloud providers are able to adapt such an environment to various customer requirements, without performance, availability, or security suffering as a result.

About the Author: Frederik Bijlsma is Strategic Account Manager, Red Hat (News - Alert).