Virtualization technologies are becoming more advanced today and their adoption are growing. Increasingly, IT services are being delivered using virtualization technologies to derive higher cost efficiency and optimization. In the past, x86 compute hardware were designed to run a single operating system and a single application workload. As IT begins to realize that their compute assets are under-utilized and there are beginning to have so many of these under-utilized compute assets in their data center, they begin to explore using virtualization technologies to run multiple independent operating systems images (or virtual machines or VM) on each physical compute hardware. This allows sharing of resources on a single physical compute hardware with multiple environment. Not only that, you could have different virtual machines running different operating systems and multiple applications on the same physical compute hardware.
When we design an x86 virtualized architecture, it is common practice to first size up each VM environment’s resource requirements, service level requirements, criticality (mission-critical, business-critical or operational-critical), and then determine what would be the right compute hardware to place it on and which other application workloads could reside together on the same hardware.
It is also a common practice to group together several similar compute hardware as a cluster group. What this will do is to enable the VMs running in this cluster to be delivered using any of the hardware in that set of equipment within the cluster, in order to meet current demand as well as for HA (high availability) purposes. It is policy-driven. For example, we could establish a policy to shift processing from one physical hardware to another without any disruption to the service, in response to a failure or if the hardware is reaching high utilization levels during peak-demand times – all automated and automatic. This will mean even greater savings of costs as any equipment can be used at any time, without any human intervention, except to perform maintenance and replace failed equipment. On top of that, the IT Infrastructure is more resilient since any hardware is backed up by any number of similar hardware, any of which could take over a failed hardware’s workload automatically.
Service Operation
What the above means, from a Service Operation perspective, is that we will need to ensure the investment in technologies and processes to support Service Operation are adaptive rather than rigid.
When considering Service Operation, most people would think of only the day-to-day management activities and technology. However, Service Operation exists within a far greater context. As part of the Service Management Lifecycle, Service Operation is responsible for executing and performing processes that optimize the cost and quality of services. It is also responsible for enabling the business to meet its objectives as part of its role in the organization. And finally, as part of the world of technology, it is responsible for the effective functioning of components that support services.
Effective Service Operation is dependent on the ability to detect any deviations from normal operation / events and knowing the status of the infrastructure. Typically this is achieved through use of active monitoring tools that will poll key CIs (Configuration Items).
The significance of a virtualized environment leveraging on the capabilities of virtualization technologies to cluster and move workloads from one physical compute hardware to another, is changing the way we need to manage CIs, events and monitoring functions. Some examples of these changes could be in the form of:
- A shift in focus – IT organizations need to manage the virtualization technology just as if it is one of their mission-critical application services. Understanding and managing this technology is important to control the infrastructure environment’s availability, data persistence, and dependencies between the various application workloads and IT elements. It is a complex ecosystem, and without having a good grasp of the virtualization technology, this ecosystem may tail-dove into chaos.
- A different perspective of looking at Configuration Items (CIs) and the relationship between CIs to the IT services. In the past, CIs are often associated to the physical IT assets. However, in today’s context, as IT services might be running on different assets, depending on the demand, this equation is becoming more complex.
- Unification of the management systems and controls governing the Virtual Machines (VMs) and the monitoring capabilities. Instead of planting a traditional monitoring agent on the VMs, perhaps having a centralized tool to monitor both the physical hardware layer health, virtual machines health, as well as the output of the IT services instead might be a more appropriate solution.
- Establishing policy-based management, where the rules of automation and resource allocation are managed by policy.
Tags: Architecture, Consolidation, ITIL, Operation, Service Management, services, Virtualization