Next-generation data centers will be modular, cost-efficient, and green when compared to the typical current-generation data center. With the barriers to market adoption breaking down, more organizations deploying production virtualization environments, and maturation of quality solutions to manage and secure virtual infrastructure, companies are positioned to run more efficient IT operations that align with overall business objectives.
While virtualization brings great benefits and flexibility, it also brings its own set of challenges. The dynamic nature of virtualization requires a different perspective and a new way of managing infrastructure. To realize the true benefits of virtualization, administrators must address these challenges head-on.
In this article, I examine five key areas from the perspective of virtualization management:
- Change Management
Visibility and Awareness of the Virtual Infrastructure
Virtual data center administrators experience management and policy enforcement limitations due to limited visibility and control. Because you can't control what you cannot see, visibility is key in managing, monitoring, and securing a virtual infrastructure. Administrators need a visual representation of their virtual environment to understand the entire virtual network, track changes, and address virtualization challenges that have management and security implications. Next-generation data centers are more application-centric, thus requiring a different approach to visibility. Two essential views of the virtual infrastructure are necessary to understand the virtual network:
- Logical Topology. By utilizing a logical topology view that abstracts the hardware and shows the applications associated with virtual machines, how they are connected to the network, and the physical infrastructure, administrators can have a true application-centric view of the overall virtual infrastructure. The logical view is a typical topological network representation commonly used by networking professionals. Its purpose is to abstract the physical hosts that make up the virtual environment and show only the virtual network and the connectivity to the guests. Cluster network relationships are based on the clustering rules of the virtualization environment. These rules require that the network names be consistent across a cluster to ensure that virtual machines maintain their physical network connectivity during live migration. A logical view can also be used to show the relationship of guests to line-of-business applications.
For example, in an ERP application linked to web servers providing the service, virtual machine (VM) events can be tracked by their line-of-business applications, giving a clear view of what applications are impacted by VM events; see Figure 1.[Click image to view at full size]
A logical network map is used to identify the relationships between VMs, virtual networks, and line-of-business applications. Events on the virtual network will have direct impact on applications inside the virtual data center. An administrator's ability to understand those relationships is critical to quickly identifying the potential impact of virtual network events on business operations.
- Inventory Topology. An inventory view of the virtual infrastructure organizes the environment based on the physical hardware that is hosting the virtual environment. Inventory allows administrators to "drill down" into the environment using the hierarchy imposed by the chosen virtualization platform; see Figure 2.
[Click image to view at full size]
For example, VMware would start at the Virtual Center management server, then data centers, then clusters and hosts, and finally, the guest VMs. Inventory is used to easily represent how the environment is physically arranged as well as locate the specific hosts where guests are running.
The inventory mapping diagram is used to track the virtual data center inventory. Identifying and understanding the entire virtual data center components from the top level down to the virtual network configuration is essential for effectively managing and securing the virtual data center.
Server Mobility and Sprawl
Virtualization creates an invisible dynamic environment where servers and applications are easily moved around and managed by multiple cross-functional organizational departments. Organizations are still defining how to structure roles and responsibilities surrounding virtualization technology as it spans across operations, networking, server, security and compliance/auditing teams. Not only can servers, storage, and networks can be dynamically managed as a single pool of resources but servers, switches and networks can be created at the click of a mouse with applications moving seamlessly from one physical server to another. This dynamic environment can create issues not common in traditional physical environments.
- Mobility. Policy can be defined to dynamically move virtual machines based on resource utilization or for disaster recovery and business continuity. This automated mobility translates to administrators only knowing the current location of a virtual machine. A complete view of the path a virtual machine takes through the environment can help identify and control the potential issues caused by the dynamicity of the virtual infrastructure. This historical location and configuration becomes very important for both regulatory compliance and for forensic activities.
- Virtual Sprawl. The good news is that deploying servers is faster and easier with virtualization. The bad news is that this ease can lead to server sprawl, which can decrease the cost savings for the organization. To avoid sprawl, there is a need track the addition, movement, and removal of all VMs in the environment. The ability to quickly visualize, manage and report on large environments is key. In addition, the cross-functional management aspect of the virtual environment increases the number of IT staff/departments that have a hand in managing various components of the same virtual environment, leading to more configuration errors and difficulty in tracking who is doing what.