 | Issue: | Europe II 2010 |
| Article no.: | 12 | |
| Topic: | Virtualization 2.0: A required paradigm shift | |
| Author: | Pete Malcolm | |
| Title: | CEO | |
| Organisation: | Abiquo | |
| PDF size: | 216KB |
About author
Pete Malcolm is the CEO of Abiquo and joined the cloud management solution provider in 2009. Mr Malcolm is a serial entrepreneur and technologist and was the founder and CTO of Orchestria Corporation, Benchmark Capital Europe’s first Entrepreneur in Residence, and Senior VP of Business Management with CA Inc. Pete Malcolm is the inventor of more than 150 granted and pending software patents.
Article abstract
Virtualization has not realized its early promise. A paradigm shift is now required to get cloud strategy off the ground and to enable cloud management to be truly effective in organizations. Virtualization 2.0 provides this by: separating physical infrastructure and the virtual world; delegating authority for creating virtual machines; and allocating provisioning according to centralized policy. Virtualization 2.0 will change the way IT manages the physical and virtual worlds.
Full Article
Tantalized by visions of a cloud nirvana ahead promising improved efficiency and better machine utilization, IT organizations have spent the last 10 years virtualizing servers. In some ways, virtualization has been a victim of its own success, but it has not realized its early promise and has created more, not less, pressure on IT. The world has changed dramatically since server virtualization first emerged. The number of applications companies running today has increased at least ten-fold. However, the way IT infrastructure is organized to manage all machines – physical and virtual – has remained the same. The result has been chaos, with the IT infrastructure department struggling with the huge increase of applications because they are solely responsible for the provisioning of all physical and virtual machines. They simply are overloaded and are spending the majority of their time on the tedious task of provisioning machines, physical or virtual, while their customers are shouting at them about new applications they need now. Organizations have lost their agility due to this overload – exacerbated by the belt tightening of recent years – so when a user wants an application that requires a new machine to be provisioned, they have to get in line and endure a wait that may take weeks. Virtualization was supposed to solve this, but the provisioning of virtual machines takes this long now due to the demand and the lack of additional IT resources. It makes it impossible to do a short-term project, as the need for the project may have passed before the machine to address it can be provisioned. Complicating this further is the inefficient resource utilization created by the disconnect between IT, who must control all provisioning, and the actual application users. The IT infrastructure people responsible for provisioning the machine for a particular application have no direct knowledge of the application. They only provide the resource that is consumed by that group. If the project requiring an application ends, IT isn’t automatically notified that the machine can be decommissioned. Instead, the machine will just sit there consuming resources. It is becoming increasingly clear that a paradigm shift is needed for cloud management to be truly effective in organizations. The three cornerstones of this paradigm shift – what we are calling ‘virtualization 2.0’ – are: (1) separating physical resources from the virtualized servers and applications consuming them; (2) delegating authority for creating and managing virtual data centres to the application groups closer to the users; and (3) deploying virtual servers and applications among available physical resources according to a centralized policy from IT based on the functional, data security, compliance, cost and load balancing requirements of the organization. Separating physical infrastructure and the virtual world by a ‘resource cloud’ The current strict division between the consumers of IT resources and the owners/controllers of the physical and virtual resources does not make sense in this super-charged application environment. IT cannot expand at the rate required to control all provisioning. Plus, IT infrastructure people are the wrong ones to control this provisioning as they don’t have the direct knowledge of the applications. Virtualization 2.0 would make it possible for IT to retain the control it needs to have over physical infrastructure without having to provision all virtual machines to give users access to the physical resource. In a virtualization 2.0 world, IT’s role would be much different than it is today. IT would maintain the physical infrastructure and control the provisioning of resources to a ‘resource cloud’. Delegating authority for creating virtual machines To provide computing resources, IT would create a virtual enterprise, appoint an administrator and set limits on the resources that the enterprise can consume. IT can then, in complete safety, delegate provisioning of virtual data centres, servers, storage and networking to the virtual enterprise administrators who can manage their own environments, and even further delegate provisioning and management tasks to others they appoint. This delegation of provisioning, with IT resource limits and controls, would greatly increase the agility and efficiency of the organization. Virtual machines could be deployed much more rapidly. This would allow for temporary projects and even ‘what-if’ scenarios to be completed because the long waits of the past for provisioning would be eliminated. Because the virtual enterprises are aware that they have enforced resource limits, they would be motivated to decommission their virtual machines when application projects were completed rather than let the machines sit there unused. This way the resource would be made available for another project or, at least, the virtual enterprise could reduce the charges to their department for IT resources. Allocating provisioning according to centralized policy With virtualization 2.0, an automated provisioning system is established to increase the utilization of available resources, greatly improving efficiencies. The IT infrastructure group could determine at the CIO level how it wants to allocate and load balance both virtual and physical resources. As long as the provisioning is done according to policy, it doesn’t matter where the physical resource ends up being located. Benefits for IT and the organization For the last 30 years, the IT infrastructure group has borne full responsibility for provisioning all machines, physical and virtual. If provisioning of virtual machines instead can be delegated to application groups in a safe way, with set policies and procedures, IT can focus on service levels and capacity planning so its customers on the other side of the ‘resource cloud’ have what they need. Unshackled from their former state of disconnected drudgery of provisioning, IT can focus on what is more important to its customers and can serve them better. At the end of the day, the secret to making cloud computing truly effective has nothing to do with the use of private versus public cloud, which is much too much the focus of discussion about cloud computing today. The real key to achieving the promise of cloud computing depends on virtualization 2.0, which means changing the way IT is managing the physical and virtual worlds to be more in line with the current needs and demands of its internal customers today. From our perspective, the shift to virtualization 2.0 cannot come soon enough.
