Evolving Data Center Management and Virtualization Technologies Drive Cloud Computing Adoption

September 17, 2011 Off By David
Grazed from Computer Technology Review.  Author: Martin May.

Cloud computing is one of the fastest growing trends in the Information Technology world. It is a simple concept in which shared resources, including software and applications, are consumed by organizations on-demand from a cloud-based service provider. They are paid for according to consumption in much the same way as water or electricity is bought from a public utility company…

 

Cloud computing owes much of its existence to virtualization technology which is essentially a software layer that facilitates the creation of a virtual environment (rather than an actual one) in which hardware and software systems are ‘simulated’. This cuts costs in terms of capital expenditure, power utilization and space requirements for end users.

 

The concept was born back in the mainframe era of the 1970s. Then it allowed administrators to consolidate hardware and maximize costly processing power by drawing it from disparate sites, managing and using it as if it was centrally located.

 

Over time, virtualization came to encompass a broad range of resources, including application virtualization, desktop virtualization, operating system virtualization and file server virtualization.

 

According to South African-born Antonio Piraino, vice president and research director at Tier1 Research (www.t1r.com), more than 70 percent of enterprises currently use server virtualization technology to reduce costs and increase agility.

 

Server virtualization is the ‘masking’ of individual file server resources, freeing users from the physical constraints associated with their identification and location, while opening the door to resource sharing, resource optimization and easy resource expansion or reduction, as required to meet fluctuating processing demand. To the user, there is no difference between a virtual server and a physical server – except that the virtual server will not be installed at any particular site.

 

Significantly, it is possible, through virtualization, for a single physical server (running hypervisor software) to be able to contain almost any number of virtual machines, bringing a range benefits, including improved total bandwidth, to the users.

 

Other server virtualization benefits include:

  • The ability to consolidate workloads running on multiple physical servers onto a larger physical server comprising multiple independent virtual servers, with each hosting its own guest operating system and application stack; and
  • The capability to move applications from one server to another – either virtual or physical – and rapidly provision them.

 

Hypervisor software is an important element of virtualization. It allows multiple operating systems, including multiple instances of the same operating system, to share hardware resources.

 

A basic hypervisor was introduced by IBM in the ‘70s which allowed its System 370 processors to share central processing units (CPUs) and direct access storage devices and memory.

 

The concept has evolved to the point where the benefits of virtualization – chiefly its ability to drive the efficiency and effectiveness of computing and storage resources — are increasingly seen in the modern data center. However, some adaptations have been necessary because the dynamic nature of virtualization – along with the traditional organizational separation between server and network management teams – has led to management challenges.

 

In order to create an agile and automated operational model for the virtualized data center, IT administrators require a more comprehensive set of automation capabilities, coupled with increased network visibility and more granular controls.

 

Fortunately, management technologies are evolving to address these operational requirements and provide functions that ensure consistent application performance and unified management across the traditional physical and budding virtual networks.

 

Developments are driven by a new breed of IT administrators who are demanding a transparent, cross-functional service provisioning process that bridges the divides between the server, the network and the storage silos, ensuring each responsible management team has an integrated view of virtual server and network environments.

 

What’s more, because of the diversity inherent in growing networks, a variety of virtualization, storage and server platforms often needs to be supported to enable the unification of the physical and virtual network. This is also a key requirement to ensure that networks have the high availability necessary to meet the demands of mission-critical applications and the management of burgeoning business data repositories.

 

Traditionally, data center connectivity has spanned physical, virtual and storage networks using separate tools and management systems. To unify such a data center network, it is necessary to have an integrated view of the network infrastructure, servers, storage, systems and applications.

 

Coordinating the automated assignments of virtual machines (VMs) within both the virtual and physical network fabrics is a challenge that must be addressed by today’s management tools. They need to ensure proper network resources are allocated when a VM is provisioned, no matter where it is on the network.

 

This is achieved by applying individual policies to various data objects in the switching fabric, solving the challenge of virtual machine sprawl. The use of granular policies, combined with flow-based switching, helps users realize the goals of high availability and reliable delivery and supports the increasing use of bandwidth-intensive applications and the adoption of service-oriented architectures.

 

Against this backdrop, it is evident that the role of the data center is changing – and becoming increasingly important in terms of corporate efficiency and security. The dynamic allocation and location of virtual services is also driving an increase in the adoption of cloud computing technologies, enabling the mobility of virtualized servers across geographic boundaries and the provisioning of data center services across multiple data centers.

 

Cloud computing’s role in the data center should come as no surprise, mainly because it has its roots in server virtualization. Without the introduction of server virtualization, cloud computing would not have been able to gain the foothold it has.

 

Early adopters of server virtualization focused on server consolidation and benefitted from the cost savings of minimizing data center resources and increased flexibility. Today, cloud computing has taken virtualization to the next level by allowing server capacity to scale up and down dynamically and immediately on demand. In order for the cloud model, which mandates the supply of on-demand resources to users in the form of services provided by service providers, to be effective, it must make use of shared resources and applications.

 

As a result, the automation capabilities and the economies of scale offered by virtual servers are vital. The unthinkable alternative would be the manual provision of services on a per-client basis using dedicated hardware.

 

The cloud model must also cater for fluctuations in capacity and other challenges linked to dealing with an amorphous infrastructure. Only through virtualization techniques can these objectives be realized.

 

Cloud computing, data center management and virtualization technologies are committed to work hand-in-hand to deliver a common value proposition: the most cost effective means of improving agility to meet the changing business needs of large and small companies.

 

If there is an impediment to the long term success of this union, it comes in the form of security risks which will be more prevalent in poorly managed environments, where the privacy of data held on virtualized, communal systems, often in off-shore repositories, is a concern.

 

Nevertheless, the cloud and professional service providers have the potential to provide better overall security than is common to most traditional, physical computer systems. This is because security is actually enhanced when data is distributed over a wider base, multiple sites and a large number of devices.

 

More significantly, cloud computing professionals and their companies generally devote more of their resources – time, effort and money – to identify security challenges and rapidly resolve them than most end-user organizations for whom security is not a core business activity.